UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

An ecological-systems analysis of anorexia nervosa Sheppy, Margarette Isabell 1985

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

Item Metadata

Download

Media
831-UBC_1985_A2 S52.pdf [ 9.74MB ]
Metadata
JSON: 831-1.0054198.json
JSON-LD: 831-1.0054198-ld.json
RDF/XML (Pretty): 831-1.0054198-rdf.xml
RDF/JSON: 831-1.0054198-rdf.json
Turtle: 831-1.0054198-turtle.txt
N-Triples: 831-1.0054198-rdf-ntriples.txt
Original Record: 831-1.0054198-source.json
Full Text
831-1.0054198-fulltext.txt
Citation
831-1.0054198.ris

Full Text

A N E C O L O G I C A L - S Y S T E M S A N A L Y S I S OF A N O R E X I A N E R V O S A by M A R G A R E T T E ISABELL SHEPPY B.Sc, University of Alberta, 1967 M.Ed. , University of Alberta, 1971 A THESIS SUBMITTED IN P A R T I A L F U L F I L L M E N T OF T H E REQUIREMENTS F O R T H E D E G R E E O F DOCTOR O F E D U C A T I O N in T H E F A C U L T Y O F G R A D U A T E STUDIES Department of Counselling Psychology We accept this thesis as conforming to the required standard T H E UNIVERSITY O F BRITISH C O L U M B I A March 1985 ® Margarette Isabell Sheppy, 1985 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the The University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the Head of my Department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department of Counselling Psychology The University of British Columbia 2075 Wesbrook Place Vancouver, Canada V6T 1W5 Date: March 1985 ii ABSTRACT Aspects of the ecological systemic approach were used to provide a framework for the understanding of the dynamics of anorexia nervosa and were empirically tested by comparing 30 anorexics and their parents to 34 matched control subjects and their parents. The theoretical model employed was an adaptation of Conger's Ecological-Systems approach which was based on the principles of Bronfenbrennei's theory of human development The subjects were compared on selected variables arising from the individual, parent, family, and community systems using: 1) the California Psychological Inventory (CPI), 2) the Structural Analysis of Social Behavior (SASB), 3) the Family Environment Scale (FES), and 4) the Pattison Psychosocial Inventory (PPI). Statistical analysis of the difference between means was tested using Hotelling's (T2) procedure followed by a discriminant analysis. The final analysis occured using a Stepwise Discriminant procedure. The results of the stepwise analysis revealed that the Affiliation score (SASB) for the anorexics and the control subjects and the Psychopathic Deviancy score (CPI-Clinical) of the mothers of the anorexics and the controls were the variables which contributed to the discriminant analysis. With the Affiliation and the Psychopathic Deviancy scores alone, it was possible to correctiy classify 87.5% of the research subjects. Analyses also showed statistically significant results at the individual, parent, and family levels. Specifically, the anorexic daughters were found to have a negative self-worth as measured by the SASB. They were also more anxious and at odds with themselves and others (CPI). The anorexics were more depressed than the controls, had a greater tendency toward rebelliousness and hostility toward authority, and expressed their feelings in a more indirect manner. The anorexics may use ritualistic thoughts and actions in an attempt to structure their lives to overcome their feelings of losing control. They lack a sense of who they are emotionally and generally feel alienated and lonely as if on the outside looking in. The CPI-Psychopathic Deviancy scale was significantly higher for the mothers of the anorexics than for the controls. No differences were found between the two groups of iii fathers. Interactions within the family, as perceived by the anorexics, were characterized by overprotection and control by the mothers while the anorexics responded with significantly less affiliation to both their mothers and their fathers. The mothers of the anorexics also viewed their daughters as being less friendly in the relationship. There were no significant findings regarding the perceptions of the fathers of the anorexics and the controls. The families of the anorexics were less supportive, helpful and committed to each other than were the families of the control subjects as measured by the FES. iv T A B L E OF CONTENTS ABSTRACT ii LIST OF TABLES '. vi LIST OF FIGURES ix A C K N O W L E D G M E N T S x CHAPTER I - INTRODUCTION ...1 Nature of the Problem 1 Purpose of the Study 2 An Ecological-Systems Model of Anorexia Nervosa 3 Definition of Terms 9 Research Questions 10 CHAPTER II - REVIEW OF T H E LITERATURE 12 Current Theories and Research 16 Individual System 20 Parent System 29 Family System 32 Community System 45 Hypotheses 49 CHAPTER III - M E T H O D O L O G Y 54 Purpose of the Study 54 Subjects 54 Data Collection Procedure 56 Instrumentation 56 Statistical Procedures 67 CHAPTER IV - RESULTS 70 Results of the Demographic Data Analyses 70 Results of Data Analyses Related to the Hypotheses 78 V A. Analyses of the Scores in the Individual System 78 B. Analyses of the Scores in the Parent System 84 C. Analyses of the Scores in the Family System 90 D. Analyses of the Scores in the Community System 101 E Analyses of the Ecological-System 103 Summary of Results 108 CHAPTER V - DISCUSSION 110 Discussion of the Demographic Characteristics 110 Discussion of the Results for the Four Systems 114 Discussion of the Results for the Ecological-System Model 123 Limitations and Directions for Future Research 125 R E F E R E N C E NOTES 129 REFERENCES 130 Appendix A — Comparisons with Population Means 139 Appendix B - Results of A N O V A 155 Appendix C — T-Score Equivalents 157 Appendix D — Instruments 159 Appendix E — Forms 174 vi LIST OF TABLES Table 1 — An Ecological-Systems Approach to Assessing Family Relationships 5 Table 2 — Research Findings for Personality Characteristics of Anorexic Females 24 Table 3 — Research Findings for Personality Characteristics of Parents of Anorexics 30 Table 4 — Research Findings for Demographic, Interactional, and Environmental Factors of the Family of the Anorexic 39 Table 5 - Flow Chart for Statistical Procedures, Steps I to III 68 Table 6 — Means and Standard Deviations for Demographic Data for the Families of the Anorexic and Control Subjects 71 Table 7 — Stressors and Illnesses for Families of the Anorexic and Control Subjects 72 Table 8 — Family Variables: Percentages for Families of Anorexic and Control Subjects 73 Table 9 — Means and Standard Deviations for Heights, Weights, Education, and Ages of Subjects, Mothers, and Fathers 74 Table 10 — Means and Standard Deviations for the Personality Traits, Clinical Characteristics, and Self-Concept for Abstaining and Bulimic Anorexics 77 Table 11 — Means, Standard Deviations and Results of the Multivariate Analysis of Personality Traits, Clinical Characteristics, and Self-Concept for the Anorexic and Control Subjects 80 Table 12 — Means, Standard Deviations and Results of the Multivariate Analysis of Individual System Variables for Anorexic and Control Subjects 83 Table 13 — Individual System Classification by Means of the Discriminant Function 84 Table 14 — Means, Standard Deviations and Results of the Multivariate Analysis of Personality Traits, Clinical Characteristics, and Self-Concept for the Mothers of the Anorexic and Control Subjects 86 Table 15 — Means, Standard Deviations and Results of the Multivariate Analysis of Personality Traits, Clinical Characteristics, and Self-Concept for Fathers of the Anorexic and Control Subjects 89 Table 16 — Parental System Classification by Means of the Discriminant Function 90 Table 17 — Means, Standard Deviations and Results of the Multivariate Analysis of Parental Transactions with Their Daughters 92 Table 18 — Means, Standard Deviations and Results of the Multivariate Analysis of the Daughters Responses to Their Parents 95 vii Table 19 — Means, Standard Deviations and Results of the the Multivariate Analysis of the Family Environment Dimensions 97 Table 20 — Means, Standard Deviations and Results of the Multivariate Analysis of the Family System Variables 100 Table 21 — Family System Classification by Means of the Discriminant Function 101 Table 22 — Means, Standard Deviations and Results of the Multivariate Analysis of Social Network Variables for the Anorexic and Control Families 102 Table 23 — Community System Classification by Means of the Discriminant Function 103 Table 24 — Results of the Stepwise Discriminant Analysis 105 Table 25 — Integrated Systems Classification by Means of the Discriminant Function 106 Table 26 — Summary of Proportion of Correct Classification of Subjects by Means of the Discriminant Function Analysis for the Ecological-Systems Model 107 Table A - l — Means, Standard Deviations and Results of the Multivariate Analysis of the CPI and CPI-Clinical Variables for the Control Subjects and the Population Means 140 Table A - 2 — Means, Standard Deviations and Results of the Multivariate Analysis of the CPI and CPI-Clinical Variables for the Anorexic Subjects and the Population Means 141 Table A - 3 — Means, Standard Deviations and Results of the Multivariate Analysis of the CPI and CPI-Clinical Variables for the Mothers of the Control Subjects and the Population Means 144 Table A - 4 — Means, Standard Deviations and Results of the Multivariate Analysis of the CPI and CPI-Clinical Variables for the Mothers of the Anorexic Subjects and the Population Means 145 Table A - 5 — Means, Standard Deviations and Results of the Multivariate Analysis of the CPI and CPI-Clinical Variables for the Fathers of the Control Subjects and the Population Means 148 Table A - 6 — Means, Standard Deviations and Results of the Multivariate Analysis of the CPI and CPI-Clinical Variables for the Fathers of the Anorexic Subjects and the Population Means 149 Table A - 7 — Means, Standard Deviations and Results of the Multivariate Analysis of the FES Variables for the Families of the Control Subjects and the Population Means 152 Table A - 8 — Means, Standard Deviations and Results of the Multivariate Analysis of the FES Variables for the Families of the Anorexic Subjects and the Population Means 153 viii Table B - l — Results of A N O V A for the Responses of the Daughters on the Affiliation Dimension 156 Table C - l — T-Score Equivalents for the CPI-Clinical Mean Scores 158 LIST OF FIGURES ix Figure 1 — An Ecological-Systems Model for Anorexia Nervosa 6 Figure 2 — Simplified Version of the Chart of Social Behaviour 61 Figure 3 — Perceptions of the Interpersonal Transactions 94 Graph A - l — Profiles of Mean Scores for Anorexic and Control Subjects 142 Graph A - 2 — Profiles of CPI-Clinical Mean Scores for Anorexic and Control Subjects ..143 Graph A - 3 — Profiles of Mean Scores for Mothers of Anorexic and Control Subjects ....146 Graph A - 4 — Profiles of CPI-Clinical Mean Scores for Mothers of Anorexic and Control Subjects 147 Graph A - 5 — Profiles of Mean Scores for Fathers of Anorexic and Control Subjects 150 Graph A - 6 — Profiles of CPI-Clinical Mean Scores of the Father of Anorexics and Controls 151 Graph A - 7 — Profiles of FES Mean Scores for Families of Anorexics and Control Subjects 154 X ACKNOWLEDGMENTS The writer expresses her sincere appreciation to her research committee, Dr. J. D. Friesen, chairman, Dr. A. R. Hakstian, methodologist, Dr. W. A. Borgen, and Dr. R. B. Armstrong for their guidance, encouragement and understanding during the writing of this thesis. Special thanks are due Dr. Leslie Solyom who not only served as a consultant to the committee but also shared his knowledge about anorexia nervosa and treatment with the writer. Appreciation is also due to the following persons for their assistance: to Dr. D. W. Froese, Dr. G. Goertzen and the staff of the Family Practice Unit, Shaughnessy Hospital for their assistance in contacting the control subjects; to the subjects and their parents whose participation in the study was essential; to my family, fellow students, co-workers at the Vancouver General Hospital, my doctors, and friends without whose support and encouragement I would never have completed this projects. Grateful acknowledgements are also extended: to the staff in ERSC who assisted with the special computer programs; to Calvin Lai, Marsha Shroeder and Bob Prosser for their help with the statistical analysis; to Mary-Anne Lloyd for assistance with the interpretation of the CPI-Clinical scales; to Ron Davis for his helpful suggestions; to Noelle Vogel, Wes Buch and Elizabeth Fair who took time to proof read the manuscript; to Jay Handel for his time and patience in assisting me with the Textform program in the final preparation of this thesis. 1 CHAPTER I INTRODUCTION I do not remember that I did ever in all my practice see one, that was conversant with the living so much wasted with the degree of consumption (like a skeleton only clad with skin), (cited in Halmi, 1978, p. 137) This description by Richard Morton in 1689 is possibly the first recorded incident of anorexia nervosa as a distinct illness (Halmi, 1978). The condition known as anorexia is complex involving physiological, emotional, and behavioural changes within individuals who starve themselves for fear of gaining weight (Bemis, 1978; Crisp, 1980; Garfinkel & Garner, 1982). Once considered an extremely rare condition, the incidence is now reported to be on the increase (Bemis, 1978; Garfinkel & Garner, 1982; Garner, Garfinkel & Olmsted, 1983; Halmi, 1974; RichaTdson, 1980; Sours, 1969) with the possibility of it becoming an epidemic illness (Bruch, 1978). Although the etiology of anorexia has been extensively investigated, no single cause has been identified. Garfinkel and Garner (1982) postulate that there is no single pathogenesis to the illness, rather anorexia nervosa is the final common pathway of the interaction and timing of a specific combination of predisposing factors in a given individual. They state, "anorexia nervosa is a syndrome that is the product of an interplay of a number of forces" (p. 188). Nature of the Problem Anorexia nervosa is a disorder mainly of adolescence (Crisp et al., 1974, 1980), the outcome of which may be a single mild episode early in life, to a life-long persistent or recurrent disorder with a high mortality rate (Crisp, 1965, 1980). The condition has been described from a psychodynamic, family interactional, behavioural, physiological and socio-cultural perspective (Bemis, 1978; Richardson, 1980). While each of these perspectives has merit, and research in the specific areas has added considerably to an understanding of anorexia, no single approach in and of itself is adequate to explain the condition therefore, a systemic multidimensional theoretical model needed to be developed. The need for a 2 systemic approach has been stressed by Norris and Jones (1979). Similarity Yager (1982) stated that family oriented research needs to be planned hand in hand with investigations which consider specific vulnerabilities of the affected individual as this may be the best way to tease apart the various components of individual and family factors which predispose one to anorexia nervosa. In line with this thinking, a multidimensional model has been developed by Garfinkel and Garner (1982) which considers factors related to the anorexic, the family and the culture. In contrast to a systemic multidimensional model, empirical research regarding anorexia nervosa has tended to follow a traditional unidirectional approach with the study of one aspect at a time from a given theoretical position. Further, many of the studies which are multi-factoral in nature are necessarily post facto, speculative, and impressionistic, with most studies lacking normal control subjects (Bemis, 1978; Halmi, 1983). Statistical analysis has remained mainly univariate. Consequently there is a critical need for comparative empirical research which not only uses multivariate statistical analysis but which also applies a theoretical framework capable of systematizing the existing research and theoretical concepts. Purpose of the Study The purpose of this study was to apply aspects of the ecological systemic approach to the understanding of anorexia nervosa. Further, the study was designed to empirically investigate the interactive factors of the anorexic and control subjects at the individual, parent, family and cornmunity levels using multivariate statistical techniques. The conceptual framework was an adaptation of Conger's Ecological-Systems approach which was based on the ecological principles of human development put forth by Bronfenbrermer (1977, 1979). The model employed in this study went beyond the recommendations of Norris and Jones (1979) and Yager (1982) in that it included the social network of the anorexic and her family. While it has many similarities to the mulitdimensional approach of Garfinkel and 3 Garner (1982), it conceptualizes the dynamics of anorexia from a systems perspective. An Ecological-Systems Model of Anorexia Nervosa One of the first writers to develop an ecological model which places the person within his/her environment was Bronfenbrenner (1977, 1979). His human development model is based on his conviction "that further advances in the scientific understanding of the basic intrapsychic and interpersonal processes of human development require their investigation in the actual environments, both immediate and remote, in which human beings live" (p. 12). He constructed a theoretical schema which consisted of four basic systems nested one inside the other; the microsystem, the mesosystem, the exosystem, and the macrosystem. The microsystem is composed of complex dyadic interactions between the developing person and the immediate setting containing the person. The mesosystem comprised the interactions between the major microsystems at a particular point in the person's life. The exosystem is an extension of the mesosystem involving both formal and informal social structures which have an external influence on the developing person, while the macrosystem represents the cultural belief system, ideologies and mores in operation within a given society. Further, these systems are connected by interaction patterns which extend beyond the system out into the immediate setting. Within his theory, Bronfenbrenner (1977, 1979) identified other developmental features. First, the developing individual is viewed as a growing dynamic entity who has an impact on the environment Second, the environment also has an influence on the person, and therefore a process of mutual accommodation is required which results in a two-way interaction characterized by reciprocity. Further, he believed that how a person perceives the environment is more important for behaviour and development than how it exists in objective reality. Bronfenbrenner's model is extensive and accounts for a number of external influences which affect the developing person. However, he does not take into account the qualities which characterize the individual. Conger (1981) modified the approach by coupling 4 Bronfenbrenner's principles with the concerns of Nelson and Hayes (1979) that organismic measurement must go hand in hand with environmental assessment The interactive framework Conger (1981) devised was multilevel in nature, incorporating information regarding individual family members, the family system, and community influences. The interactive quality of the model was defined by the principle of reciprocity as oudined by Bronfenbrenner (1979) which not only considers the effect A has on B, but also the effect B has on A. The criteria for selecting the illustrative measures were that "they [needed] to describe the interdependencies between characteristics of individual family members, the behavioural environment within the family and points of contact between family members and the outside community" (Conger, 1981, p. 238). Illustrative measures are identified for each system level and are listed in Table 1. The model has been used for the purposes of assessment, treatment and research. 5 Table 1 An Ecological-Systems Approach to Assessing Family Relationships Level of Analysis Illustrative Measures 1. The individual family member 2. The family system 3. The community b. c. a. Social background (parents) b. Experiences in family of origin (parents) c. Emotional functioning d. Intellectual functioning e. Excessive, deficient, or inapporpriate behavioural characteristics a. Structure: (1) number of adults (2) number of children (3) ages of parents and children (4) living conditions Perceptions or attributions by family members of one another Patterns of interaction: (1) reciprocity - mutual reinforcement or mutual punishment (2) equity (3) coercion (4) competition (5) cooperation a. Social position: (1) economic status (2) educational success (3) geographic location (4) desirability of employment b. Contacts with social agencies: (1) voluntary as desired (2) coercive-economically necessary or instigated by others c. Social relationships: (1) friendship networks (2) extended family (Conger, 1981, p.203) Conger's ecological-systems model meets the theoretical and conceptual expectations of this study since it focuses on the individual within his/her environment and presents a method to analyze these elements systematically. It will therefore provide the overall framework for the present comparative investigation. The model is illustrated geographically in Figure 1. 6 Level of Analysis Variables Individual Personality Traits Clinical Characteristics Self-concept Qualities Parent Personality Traits Clinical Characteristics Self-concept Qualities Family Parent-Child Interaction Family Environment Community Social Network (a) Size (b) Quality Figure 1 An Ecological-Systems Model for Anorexia Nervosa The first level of analysis in Conger's (1981) model is- the individual family member. It can be deduced from Conger and Burgess (1980) that this implies any or all family members. In the present study, this system level has been divided into two components, the anorexic who is perceived as the individual embedded in her environment and her parents. The variables selected in this study reflect the interactive qualities within the environment and thus uphold the principles of Bronfenbrenner (1977, 1979) and adhere to the criteria of analysis set out by Conger (1981). The aspects of the ecological systems model which are applied in this study are: 1) individuals are embedded in their environment, 2) various components may be identified at each systems level and measured, and 3) a systematic framework is used for assessing anorexia nervosa. 7 Support for the Ecological-Systems Model In order to study the child within an ecological framework, Minuchin (1970) stated that three elements need to be analyzed independently: the child, the environment in which the behaviour occurs, and the linkages between the child and the environment He further expanded this notion of studying the child in his/her environmental contexts into a theoretical model which has been applied to the investigation of psychosomatic families (Minuchin, Rosman, & Baker, 1978). Powell (1979), writing from a socio-ecological approach for the investigation and treatment of the family, points out that "the socio-ecological contexts of family functioning need to be given serious consideration in the design of research and programs" (p. 1). He goes on to say that while this focus is important, "the interplay between the social environment and family childrearing processes is a relatively uncharted terrain ... [with] few useful theoretical models to guide research and program development efforts" (p. 1). The concept Powell (1979) has of social embeddedness and which he considers integral to a socio-ecological framework is that: The child is seen as embedded in a family system which in turn is enmeshed in a society. Interactions between these interconnected systems are viewed as having a critical influence on human development (p. 2). While Minuchin and Powell, along with Bronfenbrenner and Conger, stress the need for a theoretical approach which considers the individual in his/her environment, anorexia nervosa has not been conceptualized from such a perspective despite Lasegue's (1873) initial emphasis on the role of the family in anorexia nervosa. However, various systems of the ecological model have been stressed independently by writers such as Minuchin et al. (1978), Selvini (1978), Norris and Jones (1979), Yager (1982), Halmi (1978) and Garfinkel and Garner (1982). Norris and Jones (1979) along with Yager (1982) emphasized the necessity for understanding the vulnerabilities of the anorexic as she is embedded in the family context Minuchin et al. (1978) focused more specifically on the family dynamics. The socio-cultural factors are being explored by Hall and Brown (1983) and Garfinkel and Garner (1982). The above analysis of the current thought on the development of 8 symptomatology presents a strong argument for the systematization of the individual, the family and the cultural variables. However, these models do not consider influences exerted by the community or the interactions between systems. A multidimensional perspective integrating the many aspects of anorexia nervosa was developed by Garfinkel and Gamer (1982). They used the illness models of Weiner and Kubie to provide a framework for drawing together the variety of factors related to anorexia nervosa into a multidimensional perspective which constitutes a three level model: the individual, the family, and the culture. While this approach presents a multidimensional perspective of anorexia nervosa and emphasizes the interaction between the factors which predispose to, or perpetuate the condition, it is not based on the principles of systems functioning. A unidimensional outlook is maintained with each of the factors being discussed as a separate entity, such that the predisposing, precipitating, perpetuating factors, and the interplay between them, is considered highly variable for any one individual. Research into how these factors are connected and relate to the extended environment of the anorexic is not discussed. In keeping with the recommendations of Yager (1983) that further research be conducted which considers the individual and family factors simultaneously along with the ecological principles set forth by Bronfenbrenner, Powell, and Conger, the need to research anorexia nervosa from a broader ecological-systems perspective is evident While anorexia nervosa has not been researched using an ecological-systems framework before, such a model has been applied to the empirically investigation of other dysfunctional families, specifically child abuse (Belsky, 1978; Papatola, 1982; Conger & Burgess, 1980). 9 Definition of Terms Operational definitions of several variables in this study are as follows: Anorexic subjects. For the purpose of this study, anorexic subjects are defined as those females between the ages of 15 and 23 who have been diagnosed as having anorexia nervosa according to the criteria found in the Diagnostic and Statistical Manual, Vol.111 (1980). Further, as the greater percentage (90% to 95%) of anorexics are female and as all anorexic subjects in this study are female, for purposes of simplification, female pronouns will be used throughout the study when refering to the anorexic. Parents of Anorexics. Parents of anorexics, in this study, refers to the biological, adoptive or stepparents of the females diagnosed as having anorexia nervosa. Control subjects. Control subjects in this study were females who were matched according to the ages of the anorexics and the socioeconomic class of the head of the household. These females had no known psychosomatic or psychiatric illness. Parents of Controls. Parents of controls, in this study, refers to the biological, adoptive or stepparents of the matched control subjects. These parents had no known psychosomatic or psychiatric illness. Operational definitions for the dependent variables are listed below. Individual Variables Personality Traits Social Presence, Self-acceptance, Sense of Well-being, Socialization, Flexibility, and Anxiety scores as measured by the California Psychological Inventory (CPI). Clinical Characteristics Depression, Hysteria, Psychopathic Deviancy, Schizophrenia, and Social Introversion scores as measured by the subscales of the CPI and scored according to Rodgers' (1966) technique. These subscales are referred to as the CPI-Clinical measures in this study. Affiliation The weighted affiliation score as measured by the Structural Analysis of Social Behavior (SASB), intrapsychic plane. Autonomy The weighted autonomy score as measured by the SASB, intrapsychic plane. 10 Family Variables Affiliation Autonomy Family Environment Community Variables Size Quality Social Status The weighted affiliation score as measured by the SASB, interpersonal plane. The weighted autonomy score as measured by the SASB, interpersonal plane. Cohesion, Conflict, Independence, Organization and Conflict as measured by the Family Environment Scale (FES). The total number of persons listed by the family members when asked to list their important family, relatives, friends/neighbors, and social/work associates as measured by the Pattison Psychosocial Inventory (PPI). The perceived degree of reciprocated feelings and thoughts, doing things for, and emotional support between the family members and important others as measured by the PPI. Socioeconomic class as Socioeconomic Index (1976). defined by Blishen's Research Questions When the anorexic is viewed as an individual embedded in her environment, then her parents, the family, the community and other external structures directly or indirectiy influence her while she in turn has an impact on the environment A number of questions arise from such a theoretical perspective, five of which will be investigated in the present study. 1. Are there characteristic personality traits and/or a self-concept which distinquishes the interaction of the anorexic with her environment and if so, how is it different from the matched control group? 11 2. As the parents are influential in the environment of the anorexic, are there personality traits and/or self-concept qualities which characterize the parents of the anorexics, and if so, how are these different from the parents of the controls subjects? 3. As the interaction between parents and child as well as the family environment within the home are important influential factors at any given point in time, are there characteristic patterns of interactions between the anorexics and their parents as well as in the family environment, and if so, how are they different from the control families? 4. As families are influenced by the degTee and type of reciprocal support provided by important others outside the immediate family, is there a characteristic pattern of interaction within the social neworks of the anorexic families, and if so, in what ways is it differernt from the interaction patterns within the social networks of the control families? 5. Given a multi-level systems approach to anorexia nervosa, does systematizing the data from the Individual, Parent, Family and Community levels into an Ecological-Systems model discriminate to a greater degree between the anorexic and control families than any of the specific systems independently? 12 CHAPTER II REVIEW OF THE LITERATURE Research in the area of anorexia nervosa has increased dramatically over the past few years with theoreticians and researchers each pursuing their own field of interest in search of answers (Yager, 1982). This review centers on the theory and research needed to understand anorexia from the Ecological-Systems perspective outiined in Chaper I. That is, the theoretical approaches and research which relates to the aspects of the Ecological-Systems model regarding the anorexic, her parents, her family, and her social network. The review provides some general information about anorexia nervosa before presenting the literature as it relates to the Ecological-Systems approach. Diagnostic Criteria Bruch (1966) identified three basic features of primary anorexia nervosa which led the way to the development of a diagnostic model. These three areas of disordered psychological function are: * a disturbance of delusional proportion in the body image and body concept; * a disturbance in the accuracy of the perception or cognitive interpretations of stimuli arising in the body; * a paralysing sense of ineffectiveness which pervades all thinking and activities of anorexic patients. Feighner, Robins, and Guze (1972) have probably established the clearest diagnostic criteria which are the recognized international standard for primary anorexia nervosa. They are summarized as follows: * age less than 25 years. * weight loss of 25% or more of total body weight * a distorted attitude toward eating, food, or weight that overrides hunger and reason, including denial of illness, enjoyment in losing weight a desired body 13 image of extreme thinness, and unusual hoarding or handling of food. * no known medical illness that accounts for anorexia or weight loss. * no other known psychiatric illness. * at least two of the following: episodes of bulimia, lanugo, amenorrhea, periods of overactivity, bradycardia (persistent resting pulse of 60 or less), vomiting (which may be self induced). These criteria are similar to Bruch's triad but are more specific about the physical manifestations of the syndrome. The Diagnostic and Statistical Manual, Vol. Ill (1980) provides the accepted diagnostic criteria used by psychiatrists and others in Canada and the United States. Anorexia nervosa is defined in the DSM III as: * an intense fear of becoming obese, which does not diminish as weight loss progresses. * disturbance of body image, e.g. claiming to feel fat even when emaciated. * weight loss of at least 25% of original body weight or, if under 18 years of age, a combination of weight loss from original body weight plus projected weight gain expected from growth charts. * refusal to maintain body weight over a minimal normal weight for age and height * no known physical illness that would account for the weight loss (p. 69). The DMS III criteria add to Feighner et al.'s definition by including a correction for normal growth and weight loss for patients under 18 years of age and not just a simple 25% total weight loss. History Richard Morton is credited with recording the first incidence of anorexia nervosa in 1689 (Halmi, 1978). The disorder was described with the characteristic behaviours by Lasegue in 1873 and named anorexia nervosa in 1874 by Gull . Although first classified as a psychosomatic disease, investigators were misled by Simmons in 1914 into thinking 14 anorexia was due to a primary endocrine disorder. The focus shifted again to the psychological components in the 1930's when Ryle re-emphasized the psychic nature of the condition (Sours, 1969). Since that time, controversy has existed regarding the etiology of anorexia, and considerable investigation has been undertaken to determine the psychological and physiological origins. Nature of Anorexia Nervosa While the signs and symptoms of anorexia nervosa are consistent, the actual clinical course of the condition varies from one individual to another (Crisp, 1980). Some of the commonalities are presented here. The initial onset is often sudden and may be precipitated by a number of stressful events (Kalucy, Crisp & Harding, 1977; Crisp, Hsu, Harding & Hartshorn, 1980). One of the most frequent of these is the biological process of puberty. Early in the condition, there is a feeling of being fat and a desire to eliminate this feeling. The meaning of fatness to the adolescent in general stems from a number of sources, one of which is the cultural emphasis on slimness (Crisp, 1980; Garfinkel & Garner, 1982). Restriction of carbohydrates to limit caloric intake invariably becomes the method of weight control. The definite onset of anorexia occurs when this dieting is intensified, all carbohydrate is avoided, and extreme fussiness over eating occurs. Secretiveness and hostility result when the anorexic is not allowed to control the events surrounding her eating. Such behaviours result in a form of starvation unique to the anorexic and which is perhaps responsible for many of the characteristic physiological changes once attributed to the condition itself (Bemis, 1978; Crisp, 1980; Garfinkel & Garner, 1982). Amenorrhea is often one of the early symptoms and may result from a decrease in body fat (Crisp, 1980; Garfinkel & Gamer, 1982). With the increased weight loss, secret binging may begin followed by vomiting and/or the use of laxatives. This binge/vomit/purge syndrome is another method to control weight, is considered a symptom of chronicity, and is often associated with a premorbid impulsivity (Crisp, 1981). 15 Socially, it is reported that the anorexic stops associating with her friends, especially any important boyfriends (Bruch, 1973; Crisp, 1980). If still in school, she increasingly commits herself to study. Resdessness or hyperactivity is common and the anorexic frequently engages in strenuous activity. Relationships with the family are stormy. The parents resort to "two methods ... entreaties and menaces ..." to deal with the problem while the anorexic remains steadfast in her resistance (Lasegue, 1873, p. 149). Findings regarding prognosis and outcome are uncertain as a number of methodological issues exist which make comparisons between outcome studies difficult Garfinkel and Garner (1982) extensively reviewed the .major prognostic studies since 1950 in the English literature. They concluded that overall these studies indicate about 40% of all patients recover totally, 30% show improvement, and 30% either die as a result of the disorder or become chronically i l l . Another review (Steinhausen & Glanville, 1983) states the mortality rate due to suicide or complications is between 0 and 21% while the rate of chronicity ranges from less than 20 to 79%. The mortality rate of the illness as reported in the DSM III (1980) is between 15 and 20%. Weight and menses return to normal in approximately 50% of the cases with a similar number returning to normal dietary intake. Family problems persist in 40 to 50% of the cases while 25 to 45% report continued difficulty with social situations (Garfinkel & Garner, 1982). Incidence Once considered an extremely rare condition, the incidence of anorexia nervosa is reported to be on the increase over the past 20 years (Richardson, 1980; Bemis, 1978; Crisp, 1980; Sour, 1969). Exact statistics are difficult to obtain because of inconsistent diagnostic criteria and lack of reporting. Halmi (1978) believed that one in 200 females in the high-risk age group of 12 to 18 years may be affected by the disorder while others have estimated the incidence to be between .24 (Theander, 1970) and 1.6 (Kendall, Hall, Hailey & Babigian, 1973) per 100,000 population annually. Comprehensive searches of 16 hospital records have indicated an increase in the number of individuals hospitalized for anorexia nervosa over the past few years (Halmi, 1974; Kendell, Hall, Hailey & Babigian, 1973). Richardson (1980) reported an increase from 12 cases in a 10 year period, (1954 to 1964) to 35 cases in a five year period (1974 to 1979) at the Washington School of Medicine. The fact that all cases are not severe enough to enter hospital, and are treated on an outpatient basis from offices and clinics, adds to the difficulty of obtaining accurate statistical information. In Canada the incidence of hospital separations with the diagnosis of anorexia nervosa as reported by Statistics Canada (Note 1) for the five year period 1973 to 197? ranged from 2.5 to 2.7 per 100,000 population. When the high-risk age group of 10 to 20 years is considered, the incidence increases to an average of 4.9 per 100,000 population. The finding that anorexia nervosa occurs predominandy in females (Bruch, 1981; Crisp, 1980; Fromm, 1981; Garfinkel & Garner, 1982; Giannini, 1981-82) is supported by the figures from Statistics Canada. In the high-risk ages 10 to 20 years, an average of 11% were males and 89% females. The incidence of the disorder among females in the ages 10 to 20 years ranged from 8.7 to 9.6 per 100,000 female population. British Columbia has the highest incidence of anorexia nervosa (range 3.3 to 4.4 per 100,000 population) when compared with the other Canadian provinces, and this incidence was consistendy higher than the national figure. These figures relate only to hospitalized cases of anorexia nervosa and therefore give only a partial picture of the incidence of the disorder. Current Theories and Research Anorexia nervosa has been "studied comprehensively and with systematic methodologies only in the last 20 years" (Halmi, 1983). Most of the investigation has centered on the anorexic, while a paucity of empirical data and an over reliance on clinical observations and anecdotal reporting methods serve as the basis for the present 17 understanding of the personality characteristics of the parents, the parent-daughter interactions and the family features (Halmi, 1982). Researchers are now turning their attention to aspects of anorexia which extend beyond the anorexic herself. However, a number of methodological problems remain unsolved, three of which are: 1) lack of specificity in definition; 2) multiple factors of causation; and 3) the low incidence rates. Bemis (1978) states that "many of the contradictory findings of research into [anorexia nervosa] may be attributable to lack of specificity in definition" (p. 593). Controversy still exists regarding whether or not the disorder comprises a specific disease entity. Investigators such as Bruch (1973), Crisp (1965), Garfinkel and Garner (1982), King (1963), Nemiah (1950), Russell (1970), Selvini (1978), Theander (1970), and Thoma (1977) recognize the condition as a defined syndrome with characteristic signs and symptoms. Others like Sours (1969, 1974) and Kay and Leigh (1954) maintain that there is no identifiable disorder, rather anorexia is only a symptom of a variety of other organic or psychiatric illnesses. Apart from the debate about the clinical status of anorexia, difficulties exist regarding the diagnostic criteria. This difficulty was resolved in part by Feighner et al., (1972) when they defined diagnostic criteria. However, Halmi (1982) points out that "currently there are few investigators using Feighner's first criterion, age of onset prior to 25, as an exclusion criterion for the diagnoses of anorexia nervosa" (p. 248). Further, she states that "there are no data or consensus of opinion as to the degree of weight loss that is necessary for the diagnoses of anorexia nervosa" (p. 248). While considerable confusion remains regarding appropriate diagnostic criteria more agreement is becoming evident among investigators but better diagnoses could be established if the etiological factors were more clearly defined (Lester, 1981). Multiple factors of causation is the second problem. Garfinkel and Garner (1982) state that "while the exact mechanisms of etiology are not known, anorexia nervosa is a syndrome that is the product of an interplay of a number of forces" (p. 188). Because the clinical manifestations of anorexia nervosa are generally similar, investigators have erred in searching for a single pathogenesis (Garfinkel & Garner, 1982). Anorexia was first 18 conceived of as a psychosomatic entity by Lasegue (1873) and Gull (1874). However, in 1914 Simmonds introduced the possibility of there being an organic component related to the anterior pituitary (Garfinkel & Garner, 1982; Bemis, 1978; Crisp, 1980). This was the first indication that there may be a number of causal factors in anorexia nervosa. Presentiy research is being conducted which takes into account a number of different aspects of the condition (Bemis, 1978; Halmi, 1983; Garfinkel & Gamer, 1982). Bemis (1978) states that "the contributions of all areas must be incorporated into an understanding of anorexia nervosa that recognizes multiple factors of causation" (p. 611). The multidimensional approach developed by Garfinkel and Garner (1982) was designed for this purpose and integrates the existing knowledge about anorexia nervosa into a conceptual framework. Research in the treatment of anorexia also reflects the diversity of theoretical perspectives regarding the condition. Bruch (1973), Selvini (1978), Thoma (1977), Sours (1974) , Dally (1969) and others tend toward the psychodynamic viewpoint Minuchin, et al. (1978), Hall (1978, Note 2), Norris (1979), Norris and Jones (1979) and Wilson (1980) focus on the family interactional aspects. Behavioral approaches used in conjunction with medical treatment are endorsed by such researchers as Halmi, Powers, and Cunningham (1975), Azerrad and Stafford (1975), Brady and Rieger (1975), and Garfinkel, Kline, and Stancer (1975) . Drugs and other medical procedures have been used for years with varying degrees of success (Bemis, 1978; Halmi, 1983). The low incidence rate of anorexia nervosa presents the third methodological problem. The frequency is difficult to establish as only hospitalized cases are recorded by Statistics Canada. Further, the problem with definition can lead to misdiagnosed or even undiagnosed cases. Due to the nature of the illness, mild forms may be relatively common but undiagnosed (Bemis, 1978). With such a low incidence rate, research subjects are usually at varying stages of the illness with a tendency toward chronicity which may account for some of the discrepancy in the results across studies. 19 A further difficulty with research in anorexia nervosa is the retrospective nature of the research: that is, the investigation is done after the condition occurs. Until the precipitating factors of anorexia nervosa are more clearly identified, anticipatory and preventive research is difficult to conduct Rampling (1980) emphasized the difficulty with retrospective self-report data when he cited a case of an anorexic, 28 year old male, who at the onset of the illness, had the mother report a normal childhood with no eating difficulties. On further investigation, it was found from the records that at the age of three, this boy had been treated in a hospital for both eating and behavioural problems. This case highlights the unreliability of obtaining data in a retrospective manner. Despite the problems associated with research of anorexia nervosa investigators continue to expand the knowledge and the understanding of the condition. Although much of the research investigates more than one dimension of anorexia, the data are analysed using univariate statistical techniques. An example of such a study was recentiy published by Garfinkel, Garner, Rose, Darby, Brandes and O'Hanlon, (1983). These researchers have reported one of the most extensive studies yet published in the area of the anorexic and her family. However, each dimension investigated (i.e., body size, family assessment, psychological functioning) is analyzed independentiy of the other dimensions using univariate t-tests. No statistical analyses were conducted to determine the interrelationships between the variables. Further research is needed where the various dimensions of anorexia nervosa are combined into a systemic theoretical framework and analysed applying multivariate statistical techniques. While methodological problems still surround the research of anorexia nervosa, an effort to understand the dynamics of the condition is evidenced in the literature. The remainder of this review focuses on the dimensions of anorexia as they relate to the various levels of the Ecological-Systems model. Theoretical formulations are presented first followed by salient research findings for the Individual, Parent Family and Community systems. 20 Individual System The literature review in this section focuses on the personality traits, clinical characteristics, and the self-concept of the anorexic. Writings from a psychodynamic approach first appeared in the literature in the 1940's. "The classic psychoanalytic interpretation of anorexia was predicated on equating eating behaviours with sexual instincts" (Bemis, 1978, p. 600). Proponents of this theory believed that: The physical and psychological symptoms of anorexia nervosa are explained as products of oral ambivalence, with the refusal of nourishment representing a defense against oral impregnation fantasies bulimia conceptualized as a breakthrough of unconscious desire for gratification and amenorrhea both as a symbol of pregnancy ... and a denial of femininity (Bemis, 1978, p. 600). While the psychoanalytical approach provides an explanation for anorexia nervosa from a drive-dominated perspective, it emphasizes only one developmental dimension and therefore restricts the broader aspects needed to understand the condition from an integrated theory (Garfinkel & Gamer, 1982). Bruch (1970, 1973, 1977) found the psychoanalytic interpretation over-restrictive in explaining the phenomenon of anorexia nervosa and shifted her emphasis to ego-weakness and interpersonal factors. "Her basic theory is that anorexia nervosa is a struggle for self-respecting identity in the context of autonomy-inhibiting parents, it is the failures and impairments in ego-development and functioning that are of particular interest" (Small et al., 1984, p. 49). Originally, she identified a triad of ego disturbances: 1) an almost delusional disturbance of body image, 2) an arrested conceptual-perceptual development, and 3) a paralyzing sense of ineffectiveness. Later, two further features were added: namely an inability to identify hunger from other bodily needs, and a lack of identity awareness (Bruch, 1981). These two disturbances were conceived of as faulty early parent-child interactions where the child fails to develop appropriate responses to bodily sensations, such as hunger. Bruch (1973, 1978) attributed the lack of conceptual development to the continuous, inappropriate or contradictory actions usually of the mother, be they neglectful, oversolicitous, or inhibiting. For example, the mother feeds the child when she, the mother is hungry, not when the child indicates hunger. Thus, the child learns to respond 21 exclusively to the mother's needs and emotions and not to her own. This, Bruch (1970) suggests, results in a blurring and diffuseness of ego boundaries, body image, and core identity with the child becoming either overcompliant or rigidly negative. Instead of feeling in control of her own life, the child, "feels she is the property of her parents, feeling helpless under the influence of internal urges and external demands without inner guideposts" (Bruch, 1981, p. 5). Concurrent with her theory, Bruch (1973, 1978, 1981) portrays the premorbid anorexic as one whom the parents describe as a perfect girl, who is obedient, hardworking, precociously dependable, eager to please, academically above average, admired by her teachers and confiding in her parents. Prior to the onset of the illness, these girls manifest difficulties with self-assertion and decision making which reflects their lack of autonomy. They tend to be rigid in their interpretation of human relationships, and their self-concept is distorted. Both of these characteristics reflect a deficiency in their thinking ability which Bruch (1978) explains as a lack of abstract thinking in the cognitive developmental process. Socially they usually become isolated and withdrawn. For Selvini (1978), "anorexia nervosa is a distinct and well defined clinical entity" (p. 86). She proposed that the helplessness of the ego was central to the condition and turned to object relations theory to explain the dynamics of anorexia nervosa. The central feature was that the anorexic perceived her body as "a threatening force that must be held in check rather than destroyed" (p. 86). The developmental model which led to this perception started with the early mother-child relationship. The mother, the object, was typically seen as an aggressive, overprotective and unresponsive woman who interfered with the child experiencing the sensations of her body as pleasurable. The child internalized the "all-powerful indestructible, self-sufficient, growing and threatening" (Selvini, 1978, p. 87) features of the mother as she was perceived in a situation of oral helplessness. The body then became the bad object, a source of unpleasant and bad feelings. A helpless dejected feeling resulted from an unconscious awareness that the object was too strong to be destroyed. This feeling led the child to develops a life style of passive surrender during 22 the latency years, and when she entered adolescence, she already had a severely impaired perception of herself as separate from others. It was at adolescence that the body underwent rapid physical changes with new sensations and social pressures. The adolescent had to discover a new self but feared her ego was not up to the many tasks ahead. Faced with the impossible situation, the depressed ego reactivated the overwhelming feelings of helplessness established in infancy and defended itself by splitting into two parts. The body (incorporating ego) was perceived as the bad object, apart from her, overpowering and forcing a role upon her. The central ego, with superego components, accepted a desexualized, essentially powerful image which took active aggression against the body. "The ego defense which was thus built up was characterized by the rejection of the body as such and food as a bodily substance" (Selvini, 1978, p. 92). Thus, Selvini's explanation of the dynamics of anorexia nervosa was one of an intrapsychic paranoid split where the power object in the interpersonal relationship was internalized into the intrapersonal structure resulting in rigid control of the patient's body. Selvini (1978) agreed with Bruch that a "paralyzing sense of ineffectiveness" penetrated every thought and action of the anorexic patient but from her own experience added that she did not believe all anorexics were convinced their emaciated bodies were perfectly normal as did Bruch. She considered the splitting of the ego to be akin to Bruch's concepts of perceptual and conceptual disturbances (Fromm, 1981; Sours, 1969). Distortions in the body-image were due to two factors, failure to recognize signals regarding bodily needs and equating the body with the bad object Selvini's (1978) description of the anorexic is similar to that of Bruch. Characteristically, she perceived them as cold and forbidding when they initially came to the therapist They were described as having "suddenly changed from a normal peaceful girl into a hostile and solitary character" (p. 18). Selvini perceived the anorexic as friendless, having given up her old friends and refusing to make more. Anorexics tended to be irritable with their families, especially their mothers. They were often hypocritical and intolerant with their siblings while, if criticized themselves, they withdrew into their room in 23 tears. Anorexics prefered to eat alone and rarely sat at the table with others. They were stubbornly self-willed but rarely worried about how their behaviour affected others. Sours' (1969, 1974) unlike Bruch and Selvini, does not believe anorexia nervosa is a specific nosological entity but insists that it can be found in a wide variety of psychopathologies. His experience with anorexia nervosa patients led him to integrate the psychoanalytical approach with the intrapsychic developmental theory of Mahler. From developmental histories of seriously disturbed patients, he was able to trace deficiencies in the ego development of the subjects at Mahler's separation-individuation and reapproachment phases. He found separation anxiety was pronounced in the phallic-oedipal years when the child went to school. With the onset of puberty, these girls showed structural ego defects where they were unable to separate from their mothers and the ambivalent relationship forced the girls back into an infantile ego state. Here they refused to eat in an attempt to eradicate body feelings and sensations. The dominant figure throughout this process is, "a domineering and controlling motheT who attempts to attain passive submission and perfection for the child as her own fulfilment" (Sours, 1974, p. 571). Masterson (1977) emphasized the cause of the developmental arrest at Mahler's separation-individuation phase in the anorexic to be similar to that of the borderline individual: namely, the withdrawal of the mother when the child attempted to separate. That is, the mother withdrew her support when autonomous behaviours were expressed and reinforced the clinging, dependent actions. Anorexia nervosa was considered an adaptive mechanism in the existing dilemma. "The symptoms may also represent a mechanism for expressing hostility to an ambivalentiy-regarded parent" (Garfinkel & Garner, 1982, p. 180). The mother-child relationship is stressed as the crucial factor in the development of anorexia nervosa according to Bruch (1970, 1973, 1977), Selvini (1978), Sours (1969, 1974), and Masterson (1977). The core issue is the lack of ego development by the child which results in an inability to separate from the mother. With external pressures from the peer group an identity crisis develops in adolescence with the child resorting to anorexia in an attempt to control heT life and gain self autonomy. However, Kramer (1974) cautioned 24 against "an over-emphasis on pathological mothering and an inattention to ... the contributions of the child to the disturbed mother-child interaction and to [her] own pathology" (p. 577). The personality characteristics of the anorexic are described extensively in the literature. Many of these descriptions are impressionistic and based on observations in the clinical setting. Studies investigating the personality of the anorexic are presented in Table 2. The first five studies present premorbid personality characteristics, followed by descriptive Findings, comparisons with normal control subjects, comparisons with other conditions, and finally, a study reporting personality characteristics of anorexics who have regained their weight Table 2 Research Findings for Personality Characteristics of Anorexic Females Author(s) Findings Measure PREMORBID CHARACTERISTICS Dally (1969) 25% 'normal' childhood personalities Halmi (1974) Hall (1978) Norris (1979) Crisp et al. (1980) neurotic traits; 41% obsessive-compulsive; 79% depressive traits; 71% anxiety; 67% above average scholastically, 26% average, 8% below average. 14% disturbed in childhood; 42% had disturbed personality development; 44% had no evidence of maladjustment; High academic achievement 36% showed coexistence of obsessive, compliency, shyness, & dependency; 87% showed two or more of the above present; 44% self-centered & stubborn; 30% impulsive & hysteria; 18% grossly obsessive; 84% compulsive; 82% IQ greater than 110 (full scale). 80% were good compliant children; 25% very conscientious; 9% 'tomboyish'; Survey Clinical Observation Clinical Interview Clinical Criteria Retrospective Data 25 Author(s) Findings Measure 25% shy, timid, withdrawn; 40% no friends in childhood; 25% no friends in adolescence. DESCRIPTIVE STUDIES Dally (1969) Theander (1970) 47% obsessive-compulsive; 8% anxiety & 8% hysteria. depression, anxiety and obsessive traits. SSI Clinical Observation COMPARISON WITH NORMALS Theander (1970) Smart, Beumont, & George (1976) Stonehouse & Crisp (1977) Ben-Tovim, Marilov, & Crisp (1979) Gomez & Dally (1980) Small, Madero, Gross, Teagno, Leib, & Ebert (1981) Norman & Herzog (1983) Garfinkel et al. (1983) more cautious, tense, easily tired, learnable, & narrow. more introverted & neurodc; more anxious, obsessive & independent, less extroverted; Mean intelligence = 109.4. more somatic complaints, anxious, obsessive, & depressed; more introverted & neurotic. more neurodc & introverted; more hostile & intropunitive. more introverted; more obsessive & manic. T-score greater than 70 for depression, paranoia, psychasthenia, & schizophrenia; T-score 65-70 for psychopathic deviate & social introversion. T-score >70 for depression; T-score 65-70 for psychopathic deviate, hysteria, paranoia, psychasthenia, & schizophrenia. higher scores on somatization, anxiety, depression, interpersonal sensitivity, obsessive-compulsive, & total score, more depressed. Marke-Nyman Temperment Scale EPI 16PF Raven M H Q EPI EPI H D H Q EPI SSI MMPI MMPI HSCL BDI Author(s) Findings Measure COMPARISON WITH OTHER CONDITIONS Stonehouse & Crisp (1977) Gomez & Dally (1980) Strober (1980) Strober (1981) Solyom, Freeman & Miles (1983) Solyom, Thomas, Freeman & Miles (1983) less anxious, phobic, obsessive, depressed, & somatic complaints than depressed patients, more introverted than neurotics. Primary anorexics less neurotic, anxious, hysterical than secondary anorexics. compared to patients with personality disorders, anorexics are less extroverted & neurotic; less obsessive; more obsessive-compulsive & depressed, less hostile; lower scores for Dominance, Sociability, Independence, and higher scores for Socialization, Good Impression, Communality, & Femininity; significantly lower scores for Social Presence, Psychological Mindedness, & Flexibility; higher scores for Responsibility, Self-Control, Achievement via Conformance, and Intellectual Efficiency than depressed and personality disorders. significantly more emotionally controlled, moralistic, conscientious, neurotic apprehensiveness, instability of self-image, conformity to external standards, submissive, timid, restrained, and less emotional expression than antisocial and depressed adolescents. compared to obsessives, obsessive & trait scores are similar; similar ratings by self and psychiatrist on obsessional symptomatology. restricting anorexics less anxious than obsessives & agoraphobics; fewer fears than obsessives and agoraphobics; lower interference & resistance scores than obsessives. M H Q EPI SSI EPI Leyton HSCL CPI HSPQ LOI Ratings IPAT FSS LOI 27 Author(s) Findings Measure WEIGHT RESTORED ANOREXICS Pillay & Crisp lower self-esteem as indicated by lower (1977) scores for Order, Dominance, Abasement, Change, & Heterosexuality; prefer to stay in background, decreased display of emotion, conscientious, stickler for precision, effacing & feels things deeply; more hostility, moralistic, perfectionistic, fear of isolation <& anatomical destruction; greater fear of negative evaluation & social avoidance distance than normal control subjects. Note: SSI = Symptom Sign Inventory; EPI = Eysenck Personality Inventory; M H Q = Middlesex Hospital Questionnaire; 16PF=CatteU's 16 Personality Factor; Leyton = Leyton Obsessional Inventory; Raven=Raven's Standard Progressive Matrices; HSCL=Hopkin's Symptom Check List; HSPQ= High School Personality Questionnaire; MMPI = Minnesota Multiphasic Personality Inventory; EPPS=Edward's Personality Preference Schedule; HOQ = Hysteroid Obsessoid Questionnaire; FSS = Fear Survey Schedule; IPAT=IPAT Manifest Anxiety Scale; BDI = Beck Depression Inventory. It can be seen from Table 2 that considerable variability is reported in the personality features of the anorexic. This diversity supports the statements of Dally (1969), Norris (1979), and Garfinkel and Gamer (1982) that there is no one kind of personality associated with anorexia but rather a spectrum of traits. However, several characteristics are outstanding. Contrary to parental reports that anorexics are normal children, research indicates a number of them exhibited some form of emotional disturbances in childhood (Crisp et al., 1980; Dally, 1969; Hall, 1978; Norris, 1979). Some noteworthy similarities are the statistically significant findings regarding the increased introversion and neurotic scores for anorexics on the EPI when compared to normals (Ben-Tovim et al., 1979; Gomez & Dally, 1980; Smart et al., 1976; Stonehouse & Crisp, 1977). These findings are in keeping with the descriptions of the anorexic as a shy, timid, withdrawn person (Bruch, 1973; Crisp, 1980; Garfinkel & Gamer, 1982; Selvini, 1978). Findings with regard to EPPS HOQ FSS Social Questionnaire 28 obsessive-compulsive traits show that not only are these scores higher than for normal controls (Smart et al., 1976) but also the scores of the chronically ill anorexics are similar to the scores of the obsessive patients (Solyom et al., 1982) thus indicating the possible obsessive nature of the illness (Garfinkel & Garner, 1982). A number of researchers show a significant increase in the degree of depression and anxiety experienced by the anorexic when compared to normal controls (Norman & Herzog, 1983; Small et al., 1981; Smart et al., 1976; Stonehouse & Crisp, 1977; Theander, 1970). This finding is also in keeping with the observations of Bruch (1973, 1981), Crisp (1980), Dally (1969), and Selvini (1978). Other findings important for this study are the increased scores for hysteria (Dally, 1969; Norman & Herzog, 1983), somatic complaints (Stonehouse & Crisp, 1977), independence (Smart et al., 1976), cautiousness (Theander, 1970), decreased self-concept (Pillay & Crisp, 1977), psychopathic deviancy and schizophrenia (Norman & Herzog, 1983; Small et al., 1981). Research has focused mainly on the clinical characteristics (depression, introversion/extroversion, etc.) of the anorexic with little emphasis on the personality traits (self-control, self-acceptance, socialization, etc.). Two studies were found where the more traditional personality measures were administered. Smart et al. (1976) found the anorexics to be more anxious, obsessive and independent but were less extroverted when compared to normal controls using the 16PF. Strober (1980) found several differences between anorexics and subjects with depression and personality disorders using the CPI (note Table 2). However, there are no findings which indicate what differences exist between anorexics and normal control subjects using the CPI. The lack of reported research using personality measures may be due to the fact that no significant differences have been found between the anorexic and nonanorexic female or that most of the research into the personality features of the anorexic has been carried out within clinical settings where the emphasis is placed on the more clinical aspects of personality. 29 Parent System The second system under investigation in the Ecological-Systems model employed in this study are the personality characteristics of the parents. Mothers of anorexics are frequently described as overprotective, overconcerned, and dominant while the fathers are perceived as relatively weak, emotionally absent or pushed-out of the family (Bruch, 1973, 1981; Kalucy et al., 1977; Norris, 1979; Taiple, Tuomi & Aukee, 1971). Crisp (1970) and Dally (1969) do not report such characteristic behaviours in the parents and indicate there are no specific identifiable patterns related to anorexia nervosa. Bruch (1981) states, "the mothers are often women of achievement, or career women frustrated in their aspirations, who are conscientious in their concepts of motherhood. They are subservient to their husbands in many details without truly respecting them" (p. 215). Fathers in her group were extremely preoccupied with outward appearances, reinforcing physical beauty and fitness. Despite their success in their careers, they often feel second best. Expectations for achievement and proper behaviour were high. Selvini (1978) describes the parents of anorexics as having intense neurotic conflicts. They were unable to establish mature relationships with each other, displaying instead a superficial appearance of harmony while living in a constant state of tension. The mothers of the anorexic patients feigned acceptance of traditional behaviours, pretending to submit to their husband, and were devoted to their home and children. The mother usually looked after the anorexic daughter extremely well but derived very little pleasure from their mothering activities. These mothers saw themselves as Ladies Bountiful, sacrificing themselves for the good of others. The fathers consider themselves good and decent men, remaining respectful of others at all times. Dally (1969) disagreed with the accepted description of the mothers of the anorexics as the dominant parent and the fathers the weak parent He states this finding is not confirmed. In his work, no one predominant parental pattern was found, but evidence of a wide variation in the parenting relationships with both domineering (30% mothers, 20% fathers) and passive mothers and fathers. 30 Some research findings regarding the personality characteristics of the parents are presented in succinct format in Table 3 below. Table 3 Research Findings for Personality Characteristics of Parents of Anorexics Author(s) Findings Measure Kay & Leigh (1954) Dally (1969) Theander (1970) Taipale et al. (1971) Wold (1973) prevalence of neurotic illness not greater than other groups with a psychiatrically ill member; 34% psychologically disturbed; low rate of psychosis; no evidence of schizophrenia. 75% of mothers depressed if the patient young (11-14 yrs.); fathers less depressed but more so if patient young; 33% of parents psychologically disturbed; psychiatric illness present in 24% of one or both parents; 15% had psychiatric consultation prior to onset of anorexia; 40% had a psychiatric admission - no increased prevalence of schizophrenia. emotional illness in same proportion as in general population. IQ of mother higher than daughter; mother overstressed, intellectual control, superficial formal relationships, high requirements for children, low incidence of neurotic (3 out of 13) and hypochondriac (1 out of 13) symptoms in mothers. mother unable to deal with anger and attempts to please; fathers rigidly compulsive with violent tempers; parents unable to tolerate aggression toward themselves. Clinical Observation Survey Data Clinical Observation WAIS Rorschach Case Studies Crisp et al. (1974) mothers and fathers significantiy more hysterical & less somatic complaints than normals. MHQ 31 Author(s) Findings Measure Kalucy et al. (1977) Cantwell, Sturzenberger, Burroughs, Salkin & Green (1977) Winokur, March & Mendels (1977) Hall (1978) 30% of mothers suffered from migraine headaches; increased prevalence of phobic avoidance reactions in 33% of mothers & 11% of fathers; depression in 33% of mothers & 9% of fathers; obsessive-compulsive traits in fathers; 14% fathers suffered manic-depressive illness. increased family prevalence of depression; 33% diagnosed with primary affective disorder; 6% with anxiety. primary affective disorders twice as frequent in fathers of anorexics as in normal controls (16% vs 8%) and almost 4 times as frequent in mothers of anorexics as in normal controls (40% vs 12%). 10% of parents depressed. Anecdotal Clinical Interview Structured Interview Clinical Interview Note: M H Q = Middelsex Hospital Questionnaire; WAIS = Wechsler Adult Intelligence Scale It can be noted that most of these findings are observational or anecdotal in nature and therefore speculative and impressionistic. However, a low incidence of psychiatric disorders was found as noted by the lack of reporting. Affective disorders may be overrepresented in parents of anorexics (Garfinkel & Garner, 1982). In summary the mothers of the anorexic patients could be described as achievers yet self-sacrificing for the good of their families. Outwardly they submit to their husbands for appearances sake but develop a superficial relationship with them. These attitudes lead them to display overprotective, domineering behaviours with high expectations for their children (Bruch, 1978; Selvini, 1978). Often they become hysterical (Crisp et al., 1974) or depressed (Cantwell et al., 1977; Dally, 1969; Winokur et al., 1977). The fathers become preoccupied with the outward appearances of the family (Bruch, 1973). They see 32 themselves as respectable, good, decent people who may become compulsive in their attempt to maintain such an image (Kalucy et al., 1977; Selvini, 1978; Wold, 1973). Their expectations of the family members are high (Bruch, 1981). In order to maintain the facade of a respectable family, the fathers are often seen as passive and weak (Bruch, 1973, 1981; Kalucy et al., 1977; Norris, 1979; Taiple et al., 1971). Fathers are depressed but to a lesser degree than the mothers (Dally, 1969; Winokur et al., 1977). Neither of the parents are able to tolerate conflict and anger (Wold, 1973). Research into the personality traits of the parents of anorexics, like the anorexics themselves, has focused more on the clinical characteristics. While mothers are described as dominant and authoritarian and the fathers as weak and passive, no study was found which attempted to support these observations. The need exists for research which investigates the broader aspects of the parents' personality as compared to the parents of nonanorexic daughters. Family System The role of the family in anorexia nervosa has been recognized since the time of Lasegue (1873) when he paid special attention to the effects of the relationship between the anorexic and her parents. He described the situation thus: The relatives and friends begin to regard the case as desperate. It must not cause surprise to find me thus always placing in parallel the morbid condition of the hysterical subject and the preoccupations of those who surround her. These two circumstances are intimately connected, and we should acquire an erroneous idea of the disease by confining ourselves to an examination of the patient .... The moral medium amidst which the patient lives exercises an influence which it would be equally regrettable to overlook or misunderstand (Lasegue, 1873, p. 152). While the emphasis of the condition was shifted to the individual patient by Simmons in 1914, the upsurge of family therapy has refocused the attention on the role other family members have in the development and maintenance of the disorder. The family structure, interaction, and environment as they relate to anorexia nervosa have been described in the family therapy literature (Bruch, 1977, 1981; Fromm, 1981; Minuchin et al., 33 1978; Norris & Jones, 1979; Selvini, 1978; Wilson, 1980) and become the factors reviewed in this the third system of the Ecological-Systems model. Bruch (1977, 1981) points out that with the onset of anorexia, family relationships undergo marked changes. The once happy, smooth-running, quiet home deteriorates into open fighting and constant arguments. The former covert struggle for power is now overtly manifested in an exaggerated form. This struggle revolves around the adolescents' drive for individuation, and autonomy and the parents' covert need to be in control. The open expression of authority and dictatorial efforts of the parents to return the child to her former behaviours and eating patterns are the disturbed interactional patterns within the family. The parental denial of all difficulties except the weight loss at the time of presentation along with their disregard for the anorexics' needs and emotions are common issues in therapy. A discrepancy exists between the parents extreme emphasis on the family's happiness and the denial of the seriousness of the physical and emotional illness of the patient Selvini (1978) emphasized a systems approach and based her description of the family dynamics on the theoretical model designed by Haley. Many of the parents she observed appeared utterly devoted to their home and work. They observed the conventional norms and were often puritanical and bigoted. A constant state of tension was present in the home even though they did not engage in open conflict Symptoms of the parent's latent aggression is evidenced by the propensity for endless arguments about trivial matters. Selvini (1978) further observed that the mother of the anorexic was the prominent figure in the home while the father was usually emotionally absent, overshadowed and belittled either openly or secretly by his wife. When he attempted to take more control in the home, the mother enlisted the support of the children by presenting herself as an innocent victim. The mother feigned acceptance of the traditional role of the wife, pretended to submit to her husband and be a devoted mother and homemaker. The parents acted as long-suffering guardians to the family and never courted social disapproval by their actions. Inwardly, the mothers had not accepted the role of the good wife or 34 the responsive lover. Instead they secretly or openly displayed disgust with the flesh, sex, physical lust, and excrement The parents fostered assertiveness and ambition in their children but discriminated between the boys and the girls so that the girls became entrapped by the all-powerful presence of the domineering, intolerant and hypercritical mother. When the daughters reached puberty they lacked self-awareness and were incapable of coping with the pressures of adolescence. From a communciations perspective, the daughters were placed in an impossible position where each parent because of their own disillusionments, invited the daughter to ally with them against the other parent Not only were the daughters expected to form an alliance but they were also encouraged to make up for the other partner's short-comings. Thus, the daughters found themselves torn between their mother and father, "playing the role of secret husband and secret wife all at once" (Selvini, 1978, p. 211). The alliances which were formed were ever changing and unstable. Selvini (1978) explains it thus: If she attempts to engage in a real dialogue with the father, he will reject her out of fear, while her mother will reject her out of jealousy. If she gives in to her mother, she is taken over completely as if she were still a baby, and hence rejected as a person; at the same time the father will rebuff her because of infantile behaviour. If she attacks either of her parents, the other immediately rejects her for rushing to his (or her) defence. If ... she attempts to abandon the unequal struggle and tries to stand on her own feet she will ... find herself opposed by a united couple, determined to reject her bid for independence (Selvini, 1978, p. 216). Selvini (1978) states that in a system where the interactional pattern of rejection is so high, rejection of food appears to fit the same interactional style. In order to test her theory, Selvini (1978) and co-workers observed the interactions of 12 families. They found that individual family members qualified their communication in a coherent manner. That is, they were sure of what they said and their right to say it However, the messages are commonly rejected or ignored by the other family members. The parents were reluctant to accept leadership responsibilites. They had a tendency to blame the other for his/her decisions and stated they were acting only for the good of others. This behaviour was found in all family members. Coalitions within the family were the most serious and central problem as they resulted in the establisment of secret 35 family rules. Further, Selvini (1978) noted that the marital couple had a superficial relationship behind which lay unacknowledged or unresolved disillusionment A rigid symmetrical position developed in the relationship wherein both partners sought to maintain their superior moralistic position. Rather than seeking allies outside the family to strengthen their stance, the secret was kept within the family and the patient considered an arbitrator who becames the focus for establishing alliances. Fromm (1981) coded the responses of parents of anorexics using the Structural Analysis of Social Behaviour model to determine what types of responses the parents gave when presented with vignettes of typical behaviours their daughters might use when seeking individualtion or autonomy. She found that the parents of the anorexics used significantiy more attacking types of behaviours toward their daughters individuation efforts than did parents of normals. She also found that "parents of anorexics [were] more likely than parents of normals to regard their daughters as having an oppressive self-concept when she showed normal individuation efforts" (p. 119). Applying the SASB theory to her clinical observations, Fromm (1981) suggests that before the onset of anorexia there is friendly power on the part of the parents which elicits the complementary behaviour friendly submit from the child. At adolescence, when the girl begins to exert more autonomous behaviours the parents respond with hostile power because they feel the need to maintain control. This hostile power is introjected by the daughter which is then evidenced in her oppressive self-concept as well as in her hostile submissive behaviour toward her parents. These findings support Benjamin's (1979) theory when she states that "if the patient-offspring begins to show reluctance to submit as prescribed by the loving symbiosis, then parental controlling behaviours become more hostile" (p. 9). The respective complement for the child is hostile submission and the introject is oppression of self. Minuchin and co-workers (Minuchin, Baker, Rosman, Liebman, Milman & Todd, 1975; Rosman, Minuchin, Liebman & Baker, 1977; Minuchin et al., 1978) have been credited with developing the first family systems model for psychosomatic illness. They do not propose an anorexic type of family but rather a family organization which predisposes, 36 supports, and maintains somatic symptoms as an expression of family dysfunction. They suggest that three factors working together, are necessary for severe psychosomatic illness to develop: 1) a specific organ dysfunction creating physiological vulnerability, 2) the transactional patterns of enmeshment, overprotection, rigidity, and lack of conflict resolution, and 3) the role of the sick child in conflict avoidance. While these factors are present in all psychosomatic illness, the organ dysfunction in anorexia nervosa is considered the result of emotional conflict rather than the cause, and therefore anorexia is classified as a secondary psychosomatic disorder. Minuchin et al. (1978) apply these principles to the anorexic system. From their perspective the anorexic child grows up in a highly enmeshed family system where loyalty and protection take priority over self-realization and autonomy. In such an environment, the child learns to subordinate the self and her goals become those of parental love and approval. Anorexic families are typically child-oriented. These children grow up carefully protected by their parents who are hypervigilent about their psychobiological needs. The child soon learns to become overconcerned about bodily functions. An obsessive concern for perfectionism develops because of external evaluation and the great need for approval. The need for approval leads to an extreme consciousness of herself and an alertness to those around her. Socialization is within the expectations of the family, and thus the child develops a keen sense of responsibility to keep the family from embarrassment Autonomy of the child is thwarted by the intense concern and overprotection of other family members. This cloak of concern is responsible for maintaining many of the biological and psychological functions of the child in a state of dependency rather than allowing their independent development Over-involvement with the family also hinders the development of the social skills necessary for the extrafamilial world. These children are therefore socially and emotionally immature when they enter adolescence. In wanting to participate outside the family, the anorexic is hindered by a deep sense of loyalty to the family and cannot see herself as separate. The conflict between loyalty to the family and individuality is resolved by the anorexic focusing on her parents in an effort to help and change them. 37 Boundaries for the anorexic family are distinct between them and the outside world but diffuse and weak within the family. Boundaries between the nuclear family and the families of origin are unclear as one of the parents often maintain strong emotional ties with their family of origin. Wilson (1980) expands Minuchin's model. He agrees that parents are overconcerned but states that "it is the parental preoccupation with dieting and their fears of being fat which is transmitted to the daughter by the process of identification that is specifically etiologic in anorexia" (p. 343). He found that all families he observed showed perfectionism, repressed emotions, and overcontrol of the anorexic. The two features added to Minuchin's model are: 1) exhibitionistic parental sexual and toilet behaviour, and 2) the emotional selection of the child for the development of anorexia. Parental exhibitionistic behaviours and the emotional selection of the child, "usually are uncovered only by psychoanalysis" (Wilson, 1980, p. 343). The systems perspective for the understanding of anorexia is also upheld by Norris and Jones (1979). They do not entirely agree with Minuchin as they consider his model incomplete. They propose a model which encompasses the individual, the enmeshed dyad, and the family. The individual is seen as an active participating member who affects the system and not just as a passive recipient. Anorexia nervosa is a condition which "represents a compromise solution of conflicting structures and processes operating on the one hand on the child as a component of the family system and on the other hand from within the child as a subsystem herself. Both internal and external properties must co-exist for anorexia nervosa specifically to arise" (p. 109). The idiosyncratic bond, which has been established prior to the onset of the illness, between the child and one parent is viewed by Norris and Jones (1979) as forming the enmeshed dyad. Such an entwinning of two family members challenges the complex myth of the anorexic family and threatens division unless the dyad remains within the family rules. With the onset of puberty, the biological drives of the child for independence, individuality, and sexual identity come to the fore, putting stress on the enmeshed dyadic 38 relationship. The compliant, obsessional child who harbors doubts about her own effectiveness and autonomy may deny her own individuality in favor of maintaining the family myth and the dependency of the enmeshed dyad. If so, the child succumbs to the pressures, and the development of anorexia is a compromise which maintains family equilibrium. The illness provides a reason for the overconcern, maintains the family myth, and provides an excuse for the closeness and intense sharing of feelings in the enmeshed dyad. The condition allows for the asserting of independence by the refusal to eat Once the extent of the power is experienced the person becomes reluctant to give it up. Before discussing the .relevant research related to the family system, the attitudinal aspects prevalent in society which create an increased pressure on the female members of the family need to be discussed. Slimness and fitness, the Twiggy-look, is the idealized body type for females in Western and other affluent societies (Bruch, 1973; Garfinkel, 1981; Gamer & Garfinkel, 1980; Garner et al., 1983). Unlike the somewhat ample and mature Victorian and Edwardian figures favoured earlier in this century, the societal ideal for female physical attractiveness lies in a thin body shape. The "enormous emphasis fashion places on slimness" along with the persistent message from magazines, movies, and television that "one can be loved and respected only when slender" (Bruch, 1978, p. viii) is believed to have not only added to increased dieting among all females but also to the increase in anorexia nervosa (Bruch, 1978; Garner et al., 1983). This fetish-like quality to the preoccupation with thinness for women has resulted in a changed female shape as shown by Garner, Garfinkel, Schwartz and Thompson, (1980) in their study of Playboy magazine centerfolds and Miss America Pageant contestants over a 20 year period. Both these data sources indicate a shift toward a thinner ideal for women which is popularized in the high fashion trend setting magazines in our society. Along with the pressure from society to be thin is the female striving for a liberated life style. It has been suggested that just as the more voluptous female body was a symbol of fertility and mothering, an acceptance of the traditional female role, so the thin aesthetic ideal devoid of reproductive focal points may represent a restriction or 39 renunciation of the former subserviant place of women (Garner et al., 1983). "The central expression of the new, liberated woman was her thin body, which came to symbolize athleticism, nonreproductive sexuality, and a kind of androgynous independence" (Bennett & Gurin, 1982, p. 171). Thus, what may have begun as a symbol of social and sexual liberation, for some has now become an oppressive force causing them "to assume an unrealistic thin shape regardless of their biological propensities" (Garner et al., 1983). Investigation into the family system focuses on the demographic characteristics of the family, the patterns of communication, and the family environment Research findings in the three areas of the family of the anorexic are summarized in Table 4. Table 4 Research Findings for Demographic, Interactional, and Environmental Factors of the Family of the Anorexic Author(s) Findings Measure DEMOGRAPHIC DATA Kay, Schapira & Brandon (1967) Dally (1969) Theander (1970) Bruch (1973) all social classes represented; 38% of Clinical Data families experienced environmental hardships; non-serious birth complications; 50% were only children. stable homes; 18% of families with patient Clinical Data 15 years or younger experienced loss of one parent through death, separation or divorce; death or serious illness in 17% relatives; anorexia preceded by physical illness. all social classes represented; 37% firstborn, Clinical Data 32% second born, 31% third born, 15% only child; 40% of mothers less than 30 at birth of proband as compared to 55% in control group. small family size; 10% only child; Clinical Data preponderance in oldest child; parents tend to be older; stable marriages (few separations or divorces). 40 Author(s) Findings Measure Kalucy et al. (1977) Hall (1978) Norris (1979) Crisp et al. (1980) Garfinkel & Garner (1980) 50% from social class I & II; death in 29% of families; 34% patients had left home; 7% parents had separated. 70% from social class I & II; parents older at birth of child (fa.=30, mo. = 29) as compared to general population; 36% of parents had broken homes by age of 16; 6% of patients parents divorced; family size nonsignificant; 38% oldest child, 5% middle, 42% youngest. all size families; sibling position nonsignificant - roughly 6% only child, 40% firstborn, 38% youngest; girls outnumbered boys in families 3:1, nearly 1/2 siblings were female; Low incidence of divorce (4% as compared to 27-30% in the population; 70% from high income bracket 14% only child, 27% firstborn, 41% youngest; 62% from social class I & II. 59% from social class I & II; 7% only child; 17% of mothers over 35, 6% over 40 at birth of child; 34% fathers over 35 & 17% over 40 - higher than national average (social class not controlled). Clinical Data Personal Interviews Clinical Data Clinical Data Clinical Data INTERACTIONAL FACTORS Kay et al. (1967) twice as many marital difficulties as in normal controls. emphasized normality of family and parental denial of conflict Clinical Data Clinical Observation Halmi (1974) Kalucy et al. (1977) 36% reported family conflicts; 16% involved in a new relationship. 39% enmeshed mother-child relationship, 25% in father-child and 13% in both; negative feelings between father and child in 20% of families, between mother & child in 11%; marital discord in 40% of cases. Clinical Data Clinical Data 41 Author(s) Findings Measure Hall (1978) 68% unhappy marital relationships; covert Personal Interviews competitive relationship with sibling of closest age; parental expectation for patient to relate well to closest age sister. Selvini (1978) families communicate in a coherent manner; Case Studies messages sent by others commonly rejected; parents reluctant to accept leadership responsibilities; formation of covert coalitions; spirit of self-sacrifice; facade of unity in marital relationship with underlying disillusionment Norris (1979) mother initiated and controlled interaction Clinical Observation in 21 out of 28 families, father in 3 of the 28; marital conflict denied; a strange entanglement between mother & child in 64% of cases, with father in 21%. Crisp et al. (1980) minimize difficulties - present a facade of Retrospective normality; 30% of families overinvolvement Analysis in mother-child relationship, 18% father-child overinvolvement 5% with both parents; sibling rivalry in 40% of cases; 45% premorbid marital discord. Fromm (1981) parental responses to individuation behaviors SASB were more attacking than in normal controls; anorexics responded with hostile submission. ENVIRONMENTAL FACTORS Minuchin et al. avoidance of conflict; display of a variety Standardized (1978) of conflict avoidance techniques - Interview aggression, joking, rationalization; parents ruminated over wellbeing of children; triadic interaction patterns - triangulation; formation of coalitions across system boundaries. Strober (1981) families of restricting anorexics are more FES cohesive, display a greater degree of organization and report less conflict than do families of bulimic anorexics. 42 Author(s) Findings Measure Garfinkel et al. mother and child perceive more difficulty F A M (1983) on Task Accomplishment, Role Performance, Communication, & Affective Expression scales than normals, and lower Social Desirability scores. Note: SASB=Structural Analysis of Social Behavior; FAM=Family Assessment Measure; FES = Family Environment Scale. Review of the demographic characteristics of the anorexic families (note Table 4) supports Kay et al.'s (1967) findings that no cohesive group of features characterize the anorexic family. However, there are some findings which need mentioning. While all socioeconomic classes are represented there is a preponderance of cases reported from the upper middle and upper classes (Crisp et a l , 1980; Garfinkel & Garner, 1982; Hall, 1978; Kalucy et al. 1977; Norris, 1979). This finding is in keeping with the observations of Bruch (1973), Dally (1969), and Selvini (1978). The second noteworthy feature is the reported older age for the parents at the birth of the child (Bruch, 1973; Garfinkel & Garner, 1982; Hall, 1978; Theander, 1970). As noted by Garfinkel and Garner (1982) these findings may have little significance if the social class of the parents is controlled in the studies. A number of other interesting findings are also reported. The apparent stable home situation (Bruch, 1973; Dally, 1969) with a decreased incidence of separation and divorce (Hall, 1978; Norris, 1979) is noteworthy in view of the reported high rate of marital discord (Crisp et a l , 1980; Hall, 1978; Kalucy et al., 1977; Kay et al., 1967; Selvini, 1978). Such a finding would indicate these parents remain together despite an unhappy marital relationship. 43 Circumstantial events in the environment are thought to precipitate illness in some situations (Dally, 1969; Kay & Leigh, 1954; Kalucy et al., 1977; Theander, 1970). While no single precipitant of anorexia nervosa has been identified, a variety of possibilities have been suggested. Garfinkel and Garner (1982) have grouped these events into those related to separation and loss, disruption of family homeostasis, new environmental demands, direct threat of loss of self-esteem, and personal illness. Dally (1969) reported a death or serious loss for 17% of his subjects. Halmi (1974) noted family conflict (including separation) in 36% of the patients while Kay and Leigh (1954) and Theander (1970) also stressed the separation and loss theme. Events leading to a disruption of family homeostasis include such happenings as parental illness, pregnancy in a parent or friend, sibling promiscuity or a family scandal. Such events have been mentioned by Kalucy et al. (1977) and Theander (1970) as possible precipitants of anorexia nervosa. New demands or expectations which confront the individual have been considered as precipitating factors (Garfinkel & Garner, 1982). These demands often follow separation or loss but frequendy center around new relationships or the beginning of a new phase in education or career. Dally (1969) described such events in 36% of his patients and Halmi (1974) in 16%. It is not uncommon' for the condition to follow an experience where the individual has imagined personal failure (Garfinkel & Garner, 1982). This could include rejection in a relationship or lack of achievement in some athletic or artistic pursuit Dally (1969) observed this in 57% of his cases while it has also been mentioned by Theander (1970) and Halmi (1974). Finally, physical illness was found by Dally (1969) to proceed the onset of anorexia. Garfinkel and Garner (1982) point out that the preceding precipitating events are common in all forms of psychiatric illness. The difference comes in the perception of the event The anorexic "perceives personal distress in the form of 1) a threat or loss of self-control and/or 2) a threat or an actual loss of self-worth" (p. 204). The results of this threat lead her to become preoccupied with her body and ultimately to believe she will gain control and feel better should she continue to lose weight 44 Research data on family transactions are mainly clinical observations, thus, the findings are subjective in nature and open to question. Disturbed parent-child relationships are documented in the literature as a possible etiological factors in anorexia nervosa (Bruch, 1973; Selvini, 1978; Taiple et al., 1977; Hall, 1978; Crisp et al., 1980). The family presents a picture of being a normal happy family (Bruch, 1978, 1981; Crisp, 1981). However, underneath this facade of normalcy is disillusionment within the marital relationship and denial of conflict (Bruch, 1978; Selvini, 1978). The disillusionment leads to an overinvolvement or over concern on the part of one or both parents in the life of the anorexic. This involvement is described as enmeshment (Selvini, 1978; Crisp et al., 1980; Kalucy et al., 1978; Norris, 1979; Minuchin et al., 1978). Norris (1979) states this "highly idiosyncratic affiliation between one parent and the child is formed long before the onset of the illness" (p. 109). Efforts on the part of the child to break this relationship by moving toward autonomy or individuation are viewed by the parents as a threat or betrayal of themselves and they react with hostile controlling behaviours. In turn, the adolescent responds with hostile submission (Fromm, 1981). Refusal to eat becomes the anorexic's way of attempting to gain control of her own life (Bruch, 1973; Selvini, 1978; Minuchin et al., 1978; Fromm, 1981). The findings of a high incidence of marital discord and the overinvolvement of one parent with the anorexic add credence to the theoretical position where the child is perceived as a buffer in parental conflicts. However, even though the enmeshed relationship is frequentiy seen, this pattern of interaction is characteristic of other dysfunctional families and not specific to the family of the anorexic (Minuchin et al., 1978). Family members are reluctant to accept responsibility, and blame is passed from one to the other. Members invalidate their actions by explaining it was done for the good of the others (Selvini, 1978). Bruch (1978, 1980) noted family members were rigid in the way they interpreted family rules and events. They are set in their ways and closed to new approaches. Norris and Jones (1979) state the family members fit willingly into their prescribed roles. Task accomplishment, role definition, family communication, and expression of affection are reported by the anorexic and her mother as being more difficult to achieve 45 in their families when compared to normal controls (Garfinkel et al., 1983). The family is unable to deal with conflict and aggression. Peace at all cost is the rule within these families with members avoiding and denying conflict (Bruch, 1979; Minuchin et al., 1975; Norris, 1979, Norris & Jones, 1979). Should aggression arise it is watered down by denial and distortion. Expectations for achievement are high (Bruch, 1978, 1980; Norris & Jones, 1979), and members are mutually interdependent (Kalucy et al., 1977). Research to date supports the statements of Dally (1969), Kay et al. (1967) and Garfinkel and Garner (1982) that there are a variety of family interrelationships rather than a single type or constellation characteristic of the family of the anorexic. The investigation of interpersonal behaviour is not an easy task, and research measures and techniques are only now being developed. With the further development of research approaches, an anorexic type family may emerge, or the position that no real differences exist will be strengthened. The present study attempts to define some of the common interactional and environmental factors which distinguish the family of the anorexic from normal control families. Community System The fourth component of the model is the community which has been defined by Conger (1981) as, "the outside social influences that impact on patterns of family interaction" (p. 226). The community forms the larger social structure, the mesosystem, within which the family is nested (Bronfenbrenner, 1977, 1979). As elements within a system are interdependent the community has a direct or indirect influence on the family as does the family on the community (Conger, 1981). The relationship the family of the anorexic forms with the community is a relatively unexplored area. The social network and neighborhood are defined by Powell (1979) as, "layers of the social environment which stand between the family and the larger society" (p. 6). The family social network is composed of the family members' relationships with their relatives, 46 friends, neighbours, co-workers and other aquaintances. Each family member has his/her own personal network and collectively these networks make up the family social network. Interaction between the family members and the social network provides emotional and material support to the family as well as moderating life events and every day stress. Powell (1979) states, the social network plays "an indirect role ... by serving as mediating and referral systems which influence a family's use of, and relationship to, other informal and formal services" (p. 7). Two important elements within the social network are the size and nature or quality of the interactions. Size refers to the number of people listed as important in a person's life. The quality or nature of the network is rated by Pattison, Llamas and Hurd, (1979) on five variables: frequency of contact, degree of emotional intensity, positive vs. negative feelings, instrumental base, and the degree of symmetrical reciprocity. No research was found in the literature on the social network of the anorexic and her family, however, there are indications in the literature that differences may exist when compared to the nonanorexic and her family. Writers such as Crisp et al. (1980), Dally (1969), Garfinkel and Gamer (1982), and Selvini (1978) point out that the anorexic withdraws and isolates from her friends. An introverted, shy personality (Ben-Tovim et a l , 1979; Crisp, 1977; Gomez & Dally, 1980; Smart et al., 1976; Stonehouse & Crisp, 1977) does not lend itself to the establishment of social relationships. The enmeshed quality of the parent-child relationship minimizes the opportunities to develop friendships outside the family. As Norris (1979) points out, when the child reaches adolesence and begins to seek independence, the idiosyncratic bond between the parent(s) and the anorexic is threatened creating anxiety within the family. Anorexia is a compromise in this situation. Taiple et al. (1971) perceive the mothers as developing superficial relationships while Bruch (1973) observed that the fathers were overconcerned with the appearance of the family. Such behaviours do not foster meaningful interactions with others in the community. On the basis of this information, it is assumed that the social network of the anorexics and their families will be smaller and the degree of satisfaction experienced (quality) from such 47 relationships will be less than that of the control subjects and their parents. Schwartz, Barrett and Saba (1985) have observed and studied bulimic individuals and their families. They hypothesized that in part bulimia could be, "the response of a family that had become extremely isolated and intradependent ... due to being detatched from their kin network" (Schwartz et al., 1985, p. 281). Network loyalty, for these families, was seen as paramount and when the family left an established network to become "a part of bountiful, mobile, middle America" (p. 286) problems arose unless the family was able to readily develop new ways of relating to each other and the outside community. When a family was unable to make a rapid change, Schwartz et al. (1985) described the consequences thus: The family is likely to maintain its distrust of strangers and, consequentiy, become extremely isolated. The parents, not knowing how to get close to each other or not being oriented to do so, are likely to become over-involved with their children, and discourage them from leaving .... This isolation, lack of marital closeness, and child-centeredness will lead family members to feed off of each other. They will be so inter-dependent that the family will become hyper-sensitive to any threat to the marriage, or to the demeanor of the parents or to the possible departure of key members. In short, they believe that the cohesion of the family is quite delicate and they organize to avoid change (p. 286). While Schwartz et al. (1985) are describing their observations regarding bulimic individuals and their families, it is possible the same process could take place for other individuals with eating disorders such as anorexics and their families. As social network research is relatively new, it is also possible a similar phenomenon could be found in other dysfunctional families. Research into normal social networks shows they consist of 22 to 25 individuals, "drawn from a large pool of family members, relatives, friends and neighbours, and social and work associates" (Pattison et al., 1979, p. 65). This network provides the person with relatively consistent norms and social expectations, support in times of stress, positive emotional feedback, readily available instrumental assistance, a relatively conflict-free, stress reducing external environment, and reinforcement in functioning throughtout his/her life-space (Pattison et a l , 1979). The question can be asked that should such a support system be 48 decreased or diminished, as possibly in the case of the anorexic, what effects do the loss of such an important source of social feedback have on the development and/or maintenance of a symptom such as anorexia? An indication of the composition of social networks for individuals with other conditions is given by Pattison et al. (1979). In comparison to the normal social network, Pattison et al. (1979) found that the neurotic-type network consisted of fewer persons (about 15) who were mainly from the nuclear family. Emotional interactions were often weak and negative in the neurotic-type network and contacts with other individuals infrequent Such an impoverished, isolative network was stress inducing rather than stress reducing. The psychotic-type network presented a different pattern where 10 to 12 people were tightly knit together. Interpersonal relationships were negative, ambivalent and asymmetric which resulted in anxiety, stress and increased symptomatology. The individuals with a psychotic-type network had few links with the community. Tolsdorf (1976) supports Pattison et al.'s (1979) findings and reports the psychiatric patients he studied had fewer intimate relationships with their networks being dominated by family members. Psychiatric patients were impoverished in their social networks according to Silberfeld (1978). While the size of the social network was equal to that of the normal network, Silberfeld (1978) found the psychiatric subjects saw their contacts less frequentiy, spent less time with them and had fewer close relationships with kin. Miller and Ingram (1976) found that the more friends and confidents an individual had, the fewer were the psychological and physical complaints. Research to date shows that the interactions of a person with their community, as portrayed by elements within the social network, are important variables to consider when studying the individual in an ecological setting (Miller & Ingram, 1976; Pattison et al., 1979; Silberfeld, 1978; Speck & Attneave, 1973; Tolsdorf, 1976). However, further research is needed to establish the type of social network developed by the anorexics and their parents as well as determine the relationship between the networks and the other systems defined within the ecological environment 49 The preceeding review of the literature focused on the salient writings and research relevant to an investigation of the anorexic in her ecological environment While some writers (Dally, 1969; Garfinkel & Garner, 1982; Norris, 1979) indicate a diverse spectrum of personality traits exist rather than a defined syndrome for the anorexic, differences in personality, clinical charateristics and self-concept have been found between the anorexics and normal control subjects by a number of writers (Ben-Tovim et al., 1979; Garfinkel et al., 1983; Gomez & Dally, 1980; Norman & Herzog, 1983; Small et al., 1981; Smart et al., 1976; Stonehouse & Crisp, 1977; Theander, 1970). Observations regarding the personalities of the parents ' suggest they manifest more abnormal clinical characteristics than parents of nonanorexics (Bruch, 1978; Cantwell et a l , 1977; Crisp et al., 1977; Dally, 1969; Hall 1978; Kalucy et al., 1977; Kay & Leigh, 1954; Taiple et al., 1971; Theander, 1970; Winokur et al., 1977; Wold, 1973). Writers such as Bruch (1978, 1981), Crisp et al. (1980), Hall (1978), Kalucy et al. (1977), Masterson (1977), Selvini (1978) and Sours (1969, 1974) point to disturbances between the parents and their anorexic daughter especially in the relationship of the mother and the daughter. Further, writers (Crisp et al., 1980; Garfinkel et al., 1983; Hall, 1978; Halmi, 1974; Minuchin et al., 1978; Norris, 1979; Strober, 1981) have also observed dysfunctional interaction patterns within the family of the anorexic. Research into the social network of the anorexics and their families is a new area for consideration, however, Schwartz et al. (1984) suggests this aspect of the ecological environment of the anorexic is worthy of further investigation. Hypotheses The hypotheses for the present study arose out of the literature and were organized according to 1) the measures which investigate the component parts of the model; 2) the interrelationships between the component parts within each system; and 3) the systematized model which identifies the contributions of each system to the overall discriminant power of the model. 50 A. Analyses of the Component Measures Individual 1. The Personality Traits and Clinical Characteristics, as measured by the California Psychological Inventory (CPI), of the anorexic and control subjects will show a statistically significant difference. 2. The Self-concept qualities of the anorexic and control subjects will show a statistically significant difference on the dimensions of Affiliation and Autonomy as measured by the Structural Analysis of Social Behavior (SASB, intrapsychic plane). These hypotheses for the individual are consistent with the writings of Ben-Tovim et al. (1979), Garfinkel et al. (1983), Gomez & Dally (1980), Norman & Herzog (1983), Small et al. (1981), Smart et al. (1976), Stonehouse & Crisp (1977) and Theander (1970) where anorexics have been compared to normal control subjects and found to have differences in personality traits, clinical characteristics and self-concept qualities as outiined in Table 2. Mother 3. The Personality Traits and Clinical Characteristics, as measured by the CPI, of the mothers of the anorexics and control subjects will show a statistically significant difference. 4. The Self-concept qualities of the mothers of the anorexics and control subjects will show a statistically significant difference on the dimensions of Affiliation and Autonomy as measured by the SASB (intrapsychic plane). Father 5. The Personality Traits and Clinical Characteristics, as measured by the CPI, of the fathers of the anorexic and control subjects will show a statistically significant difference. 6. The Self-concept qualities of the fathers of the anorexic and control subjects will show a statistically significant difference on the dimensions of Affiliation and Autonomy as measured by SASB (intrapsychic plane). The hypotheses which address the personality traits, clinical characteristics and self-concept qualities of the mothers and father are consistent with the writings of Bruch (1978), Cantwell et al. (1977), Crisp et al. (1977), Dally (1969), Hall (1978), Kalucy et al. (1977), 51 Kay & Leigh (1954), Taiple et al. (1971), Theander (1970), Winokeur et al. (1977) and Wold (1973) as outlined in Table 3. Family 1. The perceived transactions of the parents toward their daughter (focus on other) and the responses of the daughters to their parents (focus on self) of the anorexic and control subjects will show a statistically significant difference on the dimensions of Affiliation and Autonomy as measured by the SASB (interpersonal planes). 8. Families of the anorexic and control subjects will perceive their family environments as significantiy different when measured by the Family Environment Scale (FES). Support for the hypotheses for the family are found in the writings of Bruch (1978, 1981), Crisp et al. (1977), Garfinkel et al. (1983), Hall (1978), Halmi (1974), Masterson (1977), Minuchin et al. (1978), Norris (1979), Selvini (1978), Sours (1969, 1974) and Strober (1981) where disturbed parent-daughter interactions and dysfunctional family patterns have been described as outlined in Table 4. B. Analyses for the Four Systems 9. The mean scores at the individual system level (personality traits, clinical characteristics, and self-concept qualities), together, will show a statistically significant difference between the anorexics and the control subjects. 10. The mean scores at the parent system level (mothers and fathers personality traits, clinical characteristics, and self-concept qualities), together, will show a statistically significant difference between the parents of the anorexic and the control subjects. 11. The mean scores at the family system level (parent-child interactions and family environment), together, will show a statistically significant difference between the anorexic and the control subjects. 12. The mean scores at the community system level (social network size and quality), together, will show a statistically significant difference between the anorexic and the control subjects. The hypotheses for the system level analyses find support in Garfinkel and Garner's (1982) 52 .multidimensional model of anorexia as well as in the writings of Halmi (1978), Minuchin et al (1978), Norris and Jones (1979), Selvini (1978), Wilson (1980) and Yager (1982) who each stress the need for investigation into the various systems levels except for the community. While no research was found which had analysed the dimension of the social network of the anorexic and her family, Schwartz et al. (1984) suggests this is a system which requires investigation. 53 C. Analysis for the Ecological-Systems Model 13. When all four individual systems are integrated into a comprehensive Ecological-Systems model and analysed, the four systems together will statistically differentiate to a greater degree between the anorexic and control subjects than will each of the systems separately. While no research was found which investigated the anorexic from a ecological-systemic perspective, Belsky (1978), Bronfenbrenner (1977, 1979), Conger (1981) and Powell (1979) write in favour of such a research design in the assessment of the individual in his/her environment. Similar models have been applied to the research of other dysfunctional families, specifically child abuse (Belsky, 1978; Conger & Burgess, 1980; Papatola, 1982). 54 CHAPTER in METHODOLOGY Purpose of the Study The purpose of the present study was to apply aspects of the ecological systemic approach to the understanding of anorexia nervosa. Further, the study was designed to empirically investigate the interactive factors of the anorexic and control subjects at the individual, parent family and community levels using multivariate statistical techniques. The conceptual framework was an adaptation of Conger's (1981) Ecological-Systems approach which was based on the ecological principles of human development put forth by Bronfenbrenner (1977, 1979). The empirical investigation was done in the following way. First variables within the individual, parent family and community (social network) systems were examined to determine the extent to which they differentiated between the anorexic and the control subjects and their families. Second, each of the above systems was studied to determine the degree to which its variables were able to discriminate between anorexic and control subjects and their families. Finally, all four systems were regarded together to determine if combined they would discriminate to a greater degree between the anorexic and control groups than would each system independently. Subjects The subjects were 30 anorexic females and their parents (19 fathers and 26 mothers), and 34 nonanorexic controls and their parents (24 fathers and 30 mothers) with no known psychiatric or psychosomatic illness. Of the 30 anorexics, 26 came from two-parent families, four of whom had a step-parent two were adopted, and two from the single-parent families had never known their fathers. The unequal number of daughters, mothers and fathers can be accounted for by four anorexics who did not wish to have their parents involved in the study, one father of an anorexic who refused to answer the questionnaires and six single parent families where only the mothers responded. In the 55 single parent families, four fathers were not living in the home and two were deceased. Four anorexic subjects were matched with two control subjects thus accounting for the unequal number of anorexic and control subjects (30 vs 34). The anorexic subjects were clients who had been referred from other medical practitioners to a psychiatrist over an 18 month period and who consented to participate in the study. The anorexic and control groups were matched for age and sex of the subjects and the socioeconomic status of the family as determined by Blishen's Socioeconomic Index (1976). Controls were also matched on the variables of adoption, step-parent, and single-parent status. The control subjects were receiving services from Family Practice Units and other community agencies for medical reasons. None were diagnosed as having a psychiatric conditions. The fact that the control group were also under the care of a medical practitioner controlled for the illness factor between the anorexics and the controls. While none of the control subjects were being treated for a long term illness (i.e., diabetic, cancer, physical disability), two of the parents of the control subjects had severe medical problems (i.e., leukemia, back and leg injury). The anorexic subjects included 19 abstaining and 11 previously abstaining, now bulimic, anorexics ranging in age from 15 to 23 years (M = 18.3, SD=2.29) who satisfied the DSM III (1980) criteria for anorexia nervosa. The exceptions were five subjects whose weight loss was 15% to 23% of standard body weight (Build Study, 1980; Hamill, Drizd, Johnson, Reed, Roche & Moore, 1979). However, the diagnosis of anorexia nervosa, for these subjects, was made by a psychiatrist on the basis of the other criteria specified by the DSM III. Further, Halmi (1983) points out that "there are no data or consensus of opinion as to the degree of weight loss that is necessary for the diagnosis of anorexia nervosa" (p. 248). The lowest weights for the anorexics ranged from 27 kilograms to 48 kilograms while the weights for the controls were 43 kilograms to 68 kilograms. The mean age of the anorexic at the onset of the illness was 16.2 years (SD=1.91) and the mean duration of the illness was 23.5 months (SD= 16.05). Two subjects had never menstruated. A l l others had suffered from amenorrhea for a period of 3 to 60 months (M=13.4 mons., 56 SD=12.45) prior to the presentation for treatment An equal proportion of anorexics and controls were living at home with their parents. Those not at home had contact with their parents more than once a week. Al l anorexic subjects were under the care of a medical practitioner. Data Collection Procedure All anorexic and control subjects and their parents were contacted by the researcher and the questionnaires delivered to their homes. The research procedure was explained to both groups by the researcher, and the questionnaires were left in the homes to be completed in the subjects' own time. The mothers and fathers, when available in each group, completed one California Psychological Inventory (CPI), one Structural Analysis of Social Behavior (SASB, intrapsychic), one SASB (interpersonal) for their interaction with the daughter, one Family Environment Scale (FES), and one Pattison Psychosocial Inventory (PPI). Each daughter completed one CPI, one SASB (intrapsychic), one SASB (interpersonal) for each parent one FES, and one PPI. Instrumentation The measures for this study were selected because their stated objective was to measure interactive qualities. The personality traits and clinical characteristics were measured by the use of the California Psychological Inventory (CPI); the self-concept was assessed using the Structural Analysis of Social Behavior (SASB), intrapsychic plane; the parent-daughter interactions were investigated by the use of SASB interpersonal planes; the family environment was explored with the Family Environment Scale (FES), and the social network was measured using the Pattison Psychosocial Inventory (PPI). The socioeconomic status of the families was determined using Blishen's Socioeconomic Index (1976). 57 California Psychological Inventory. The CPI was created by Gough (1975) as a means of measuring the more positive enduring aspects of the personality characteristics of the general population. The "scales are addressed to personality characteristics important for social living and social interaction" and the concepts selected "are hypothesized to be relevant to the prediction and understanding of interpersonal behaviour in any setting, culture, or circumstance" (Gough, 1975, p. 5). The CPI is a self-administering paper-and-pencil test composed of 480 statements (12 of which appear twice) which are grouped into 18 sub-scales. The raw scores, when plotted on a profde sheet, are converted to T-scores. These standard scores are based on norms obtained from over 6000 males and 7000 females. The sample, although somewhat biased in the direction of white subjects, includes populations of widely varying age, socioeconomic status, and geographical areas (Megargee, 1972). The 18 sub-scales are as follows: Dominance, Capacity for Status, Sociability, Social Presence, Self-acceptance, Sense of Well-being, Responsibility, Socialization, Self-control, Tolerance, Good Impression, Communality, Achievement via Conformance, Achievement via Independence, Intellectual Efficiency, Psychological-mindedness, Flexibility, and Femininity. Reliability of the scales was established by the test-retest method. Long-term (one year) correlation coefficients as reported by Gough (1975) are mostiy in the .60s and .70s with a median of .66. This indicates moderate stability over a period of one year. Short-term coefficients among a group of prisoners ranged from .49 to .87 with a median of .80. These correlation coefficients are relatively low for short-term testing of one to two weeks. However, coefficients reported by Hase and Goldberg (cited in Megargee, 1972) are reasonably high and range from .71 to .90 with a median of .83. Such correlation coefficients are adequate for the establishment of personality characteristics required by the present study. Validity of the CPI has been criticized, as there is considerable overlap between the scales, a condition which leads to redundancy and a high correlation between some scales. Gough's main goal in developing the battery was to predict socially relevant behaviour 58 patterns. Thus, the predictive and concurrent validity of the scales was maximized at the expense of discriminant validity (Gough, 1975). Although the validity of the CPI has been criticized by several researchers, the battery has been used extensively in the study of personality traits and has proved to be an effective tool for discriminating personality characteristics. In order to decrease the number of variables to a reasonable size for hypothesis testing in this study, the Social Presence, Self-acceptance, Sense of Well-being, Socialization, Self-concept and Flexibility scales were selected. The selection of these scales was based on the description of the anorexic as being rigid in her thinking, submissive, conforming, dependable, less spontaneous, and lacking in self-worth, poise, and self-confidence. Only one study was found, Strober (1980), in which the CPI had been used in the investigation of the personality traits of anorexics. The scales selected for the present study were all found to be below a T-score of 40, for the anorexics in Strober's study, except for Socialization (T-score = 55) and Flexibility (T-score = 43). A number of researchers have derived new scales based on the CPI. The estimation of M M P I scores from CPI data as derived by Rodgers, (1966) and the Leventhal Anxiety Scale (1968) were used in the present study. Due to the overlap of items between the CPI and the MMPI, Rodgers (1966) determined that MMPI scores could be estimated from the CPI with correlations of the corresponding sub-scales ranging from .59 to .90 (median .81) using a sample of psychiatric, neurotic, and normal subjects. "The estimated mean profiles closely duplicated the actual means in the neurotic group, slightiy overestimated the pathology in the normal sample, and slightly underestimated the pathology in the predominantiy psychotic sample" (Rodgers, 1966, p. 89). While the scales derived by Rodgers (1966) are not equivalent to the MMPI scales they provide a useful measure of psychopathology from CPI data. Raw score estimates and standard T-score equivalents (M=50, SD=10) used in this study were calculated by a computer program designed by Williams (1981). T-scores above 70 for the M M P I are interpreted as pathological (Graham, 1980). 59 Scales derived by the above method that were selected for this study were Depression, Hysteria, Psychopathic Deviancy, Psychasthenia, Schizophrenia, and Social Introversion. These scales will be referred to as the CPI-Clinical scales in this study. The selection of these scales was based on the literature (Ben-Tovim et al., 1979; Bruch, 1973, 1978; Garfinkel et al., 1983; Gomez & Dally, 1981; Norman & Herzog, 1983; Small et al., 1981; Smart et al., 1976; Stonehouse & Crisp, 1977), where the anorexic is further described as being depressed, neurotic, antisocial, anxious, obsessive-compulsive, ritualistic, withdrawn, and shy. The selected scales were found to be elevated above a MMPI T-score of 65 for abstaining anorexics (Norman & Herzog, 1983; Small et al., 1981). An anxiety scale was devised by Leventhal (1966, 1968) using twenty-four CPI items which he found to significantiy differentiate between students who had contact with the counselling center and those who did not at a given university. The test-retest reliability of the scale was .66 for males and .65 for females. Validity of the scale was initially established by the number of counselling sessions students required. It was found that students scoring one standard deviation above the mean required approximately three times the number of counselling sessions as did those who scored one standard deviation below the mean. Further research showed that a positive correlation existed between Leventhal's Anxiety Scale and certain M M P I scales (F, D, Pt, Si), the Taylor Manifest Anxiety Scale (r=.61 for males; r=.53 for females), the Welch A First Factor (r=.59 for males; r=.43 for females) and the IPTA Anxiety Scale (r=.42). Ratings by counsellors lent validity to the scale in that students scoring higher on the instrument had more severe problems and required more sessions before terminating counselling. The CPI was selected for this study because of its interactive nature, the ability to obtain a measure of psychopathology and the availability of population norms which cover a wide age range. While the scales selected for analyses from the CPI and CPI-Clinical questionnaires do not measure all the commonly identified personality traits and clinical characteristics of the anorexics outlined in the literature, they focus on some of the salient qualities helpful to an understanding of the interpersonal behaviours which were investigated 60 in this study. Structural Analysis of Social Behavior. The empirical assessment of interpersonal behaviour is not an easy task. The Structural Analysis of Social Behavior (SASB) has been chosen for this study as it was developed with the interpersonal domain in mind and, therefore, would appear appropriate for the measurement of transactions between family members. The SASB, a circumplex interpersonal model, was developed by Benjamin (1974) and is an integration and expansion of the interpersonal models of Leary (1957) and Schaefer (1965). Wiggins (1982) described the SASB as "the most detailed, clinically rich, ambitious and conceptually demanding of all contemporary models" (p. 18). Pinsoff (1981) stated that "Benjamin has developed a theoretically sophisticated, complex, dynamic and multi-purpose system for analysing a wide variety of interpersonal and intrapsychic transactions" (p. 273) and that the potential for its use in family research is just beginning to be tapped. Benjamin (1977) viewed the method as a significant contribution to the field of psychotherapy and stated that "it offers a promising beginning toward describing the process of therapeutic change through objectifying and quantifying that which has seemed vital to the clinician but elusive to the researcher" (p. 392). The questionnaire form of SASB measures the perceptions of the subjects and may or may not represent reality. "In...therapy, the reality of childhood experiences is ignored under the rationale that it is the perception and memory which affect feelings about the self and relations with significant others" (Benjamin, 1979a, p. 12). It was the perceived interactions between family members and the self-concepts of the subjects and their parents that were the focus of this study. The Chart of Social Behavior (Figure 2, simplified version) consists of three planes or surfaces which classify interpersonal transactions in terms of focus. The first plane is called Other and describes behaviours which are initiative or active in nature and are 61 E n d o r s e f reedom Manage , c o n t r o l F r e e l y come and go Y i e l d , s u b m i t , g i v e i n H a p p y - g o - 1 u c k y C o n t r o l , manage s e l f Figure 2 Simplified Version of the Chart of Social Behavior. The top two planes describe complimentary positions for two members of a dyad. The bottom plane describes the intrapsychic result when behaviors described by the top plane are turned inward on the self. (from Benjamin, L S. (1979). Structural Analysis of Differentiation Failure, Psychiatry. 42, 1-23. reprinted by permission). 62 focused toward another person. The Self plane, or second surface, describes behaviours which are reactive in nature or typical responses to the initiating behaviours of the first plane. Thus, the first two planes, Other and Self are complementary. The third surface, the Intrapsychic (Introject of Other to Self) describes behaviours which are a result of the way one has been treated by significant others and reflects the person's self-concept (Benjamin, 1974). The horizontal axis in the model represents the dimension of affiliation and contains the components of giving and responding to love and hate. The vertical axis represents the dimension of interdependence or autonomy. "Maximal interdependence is represented by the complementary poles, dominance and submission, and maximal independence is represented by the complementary poles, give autonomy - take autonomy" (Chiles et al., 1980, p. 268). When the behaviours of Other are turned inward on the Self (intrapsychic or introject) the affiliation axis ranges from love, cherish self to torture, annihilate self and the interdependence axis from happy-go-lucky to control, manage self (see Figure 2). These two axes divide each surface into four quadrants. The eight quadrants depicted on the first two (interpersonal) surfaces describe complementary relationships. That is, the third quadrant on the focus-on-other surface consists of control and annihilating attack and is thus named hostile power. The complement of such behaviour is shown on the second surface (focus-on-self), Quadrant III, hostile comply, and is made up of the elements of submission and protest This means, if one person (i.e., parent) is dominant and controlling, the person's complementary partner (i.e., child) should be submissive and give in. Benjamin (1974) contrasts opposition with complementarity in the model, in that opposition is depicted at points located 180° away from each other while complementarity is described in terms of topological^ similar locations on the first two surfaces. The Other surface represents the initiation of these types of behaviours whereas the Self surface represents the responses to the behaviours. For an explanation of the more detailed aspects of the model, the reader is referred to. Benjamin (1974, 1979b). 63 Benjamin's SASB model has been tested, refined and revised by the use of statistical analysis over the past 15 years (Benjamin, 1974, 1979b). Construct validity of the 1980 version of the model is supported by autocorrelations, factor and circumplex analyses, as well as a dimensional rating procedure. Autocorrelation is a statistical procedure whereby the relationship between the individual points of the circumplex is determined. This obtained statistic provides a measure of internal consistency of the circumplex model (Benjamin, 1974, 1979b, 1982). The correlation coefficients in which judges rated items regarding the degree of autonomy and affiliation ranged from .90 to .95. The model can be satisfactorily reconstructed by the use of factor analysis, thus validating the basic structure in terms of affiliation, autonomy and focus. Canonical correlations of dimensional ratings support the principles of complementarity and the introjecL For further information pertaining to the validity of the SASB model, the reader is referred to Benjamin (1974, 1982). The INTREX questionnaires are designed so that the subjects rate themselves or other persons on a 10-point interval scale ranging from 0 to 100. Instructions indicate a score of 0 means that the item never applies, 100 means the item always perfectly applies and 50 is the demarkation between true and false. The use of an interval scale for scoring, as opposed to a dichotomous, yes-no response, cuts down on the amount of variance attributed to the nuisance factor, acquiescence, which "reflects individual differences in the use of the response format rather than differences in the perception of self or others" (Wiggins, Steiger & Gaelick. 1981, p. 283). "The questionnaires operate on the assumption of good faith and acknowledgment of principles of psychiatric defensiveness" (Benjamin, 1977, p. 405). This means there is no conscious distortion of ratings. Computer scoring programs are required, which mathematically calculate an average or weighted affiliation and autonomy score (Benjamin, 1974, 1977). For the purpose of this study, the weighted affiliation and autonomy scores will be used to provide a measure of self-concept (intrapsychic) and to determine complimentary transactions between the parents (focus-on-other) and their daughters (focus-on-self). The 64 weighted scores "present a summary statement of the basic thrust or orientation of the above median items" (Benjamin, 1979a, p. 18) and are useful when comparing and contrasting groups of individuals for research purposes (Benjamin, 1981). Focus-on-other positive autonomy scores indicate a general tolerance for separation and autonomy while a negative score indicates an average controlling posture. Positive autonomy scores on the focus-on-self surface indicate a general taking of autonomy (differentiation), whereas a negative score indicates a general submissive posture (Chiles et al., 1980). Positive affiliation scores from the introject surface indicate friendly self-discipline while negative scores reflect hostile self-criticism and restraint Negative autonomy scores from the introject surface indicate introjected control (Benjamin, 1979b). The Family Environment Scale. The FES was developed by Moos (1974) to assess the environment of families. Moos assumed that human environments, or social climates, had unique characteristics or personalities and could be assessed in a similar way to personality qualities. The ten separate scales are: Cohesion, Expressiveness, Conflict, Independence, Achievement Orientation, Intellectual-Cultural Orientation, Active-Recreational Orientation, Moral-Religious Emphasis, Organization and Control. The instrument is a self-administered test with the responses marked either true or false. The answer sheet is hand-scored using a template. Raw scores are converted into standard score equivalents from tables provided. Profiles may be drawn for individual family members or for the family as a whole by summing and averaging each individual's subscale scores. Family scores were calculated for the present study. Reliability of the test was determined by the test-retest method. Forty-seven family members from nine families took the FES twice, with an interval of eight weeks between the testings. The correlations between the pre and post-scores for the individual sub-scales ranged from .68 for Independence to .86 for Cohesion. Internal consistencies of the sub-scales, calculated using the Kuder-Richardson Formula 20, ranged from .64 to .78. 65 Subscales significant for the present study are Cohesion and Independence, which reflect the degree of closeness or togetherness in the family, the Conflict scale, to determine the family's awareness of conflict, and Organization and Control, as a measure of family rigidity. The degree of rigidity, conflict, and closeness of family members have all been identified in the literature as characterizing the family of the anorexic (Bruch, 1973, 1978; Crisp et al., 1980; Halmi, 1974; Hall, 1978; Kalucy et al., 1977; Norris, 1979; Selvini, 1978). The Pattison Psychosocial Network Inventory. The assessment of social networks is still in the early stages of development Pattison has designed a psychosocial inventory to determine . "that part of a person's social network which theoretically comprises the existing viable social support system" (Hurd, Llamas & Pattison, 1980, p. 15). This is accomplished by having subjects list those persons who are important in their life at the present time, whether liked or disliked, from the totality of all the people they know. The selection of who is important is left entirely up to the discretion of the subject as is the definition of important Suggested categories of relationships include family members, relatives, friends, co-workers, and others. The fundamental assumption that those people listed would be the persons called upon during times of duress formed the rationale for the use of the wording in the inventory. The number of people listed forms the size of the network. Once the persons are listed the subject is then asked to rate each on a five point Likert scale for three reciprocating variables; thoughts and feelings both positive and negative in the relationship, degree of intensity of feelings, and the degree of instrumental and emotional support The reliability of the inventory has been established by the test-retest method (Hurd, Pattison & Smith, 1981). Correlation coefficients between the scales were approximately .90 (personal communication, G . Hurd, Oct 1982). 66 For the purpose of this study the size and quality of the social network for the family were investigated. The size was the average of all important people identified by the subject and her parents. The quality score is a measure of the perceived giving to and receiving from in the relationship. This measure was calculated by obtaining a difference of the means between the reciprocating variables for each person named, then averaging these differences for individual subject scores. Family scores were calculated by averaging the scores of the family members participating in the study. This is the same method used for calculating FES family scores. Socioeconomic Index Blishen (1967, 1976) has developed a socioeconomic index to determine socioeconomic class using the income level, educational status, and prestige ranking of occupations. The income value "is expressed as the percentage of males who worked in an occupation in 1970 and whose 1970 employment income was $6500 or over" (Blishen, 1976, p. 71). The education value takes provincial education differences into account and "is expressed as the percentage of males who worked in an occupation in 1970 and who had attained at least grade 12" (p. 72) or its equivalent The prestige value was determined "by assigning approximations of the Pineo-Porter prestige scores" (p. 72) to the 102 occupations. Individual index values were computed by combining the status, education and income values for each occupation into a regression formula. Social class intervals were then determined by grouping the resulting indices into six ten-digit classes as follows: 70 + 60.00 - 69.99 50.00 - 59.99 40.00 - 49.99 30.00 - 39.99 Below 30. Blishen's socioeconomic index was used in this study to determine the socioeconomic status of the anorexic and control families for the purpose of matching. 67 Statistical Procedures This study was developed on a multi-level Ecological-Systems model which includes a number of different variables grouped into four systems. In order to empirically test such a model, a series of multivariate statistical analyses were conducted. These analyses were specifically aimed at determining: 1) if each component measure throughout the model could differentiate between the anorexic and control groups; 2) if each system separately (Individual, Parent, Family, Community) could differentiate between the anorexic and control subjects; and 3) if all four systems combined would differentiate to a greater degree between the anorexic and control subjects than would each of the systems separately. The statistical procedures were grouped into three steps as outiined in Table 5. Hotelling's T 2 procedure was the statistical technique used most as it is "the uniformally most powerful test in the case of two-group ^-variate simultaneous comparisons" (Hakstian, Roed & Lind, 1979, p. 1255). Slight hetrogeneity of variance covariance matricies for the two groups was not considered a problem because the sample sizes were close to being equal (Hakstian et al., 1979). The first step was a comparison of the mean scores of the anorexic and control subjects and their parents on the CPI, CPI-Clinical, SASB (intrapsychic and interpersonal planes), FES and PPI. The measures, with their selected variables are listed in Table 5 according to system levels. Nineteen analyses were performed using Hotelling's T 2 procedure for two independent samples, followed by a discriminant analysis. The T 2 tests and the discriminant analyses were performed by means of the One-Way Multivariate Analysis of Variance program (OWMAR). Because of the large number of analyses performed, the alpha level for the overall tests and the simultaneous multiple comparisons was set at .01 to control for excessive Type I error. These analyses addressed the questions contained in the first eight hypotheses. Independent-sample T 2 tests followed by discriminant analyses were also the procedures selected for the second step, the systems level analysis. Variables which were statistically significant for each of the component measures were grouped into the Individual, Parent, Family and Community system levels (note Table 5) and analysed together to Table 5 Flow Chart for Statistical Procedures, Steps I to III System Measure V a r i a b l e Step I Step II Step I I I I n d i v i d u a l CPI Sp. Sa. Wb. So. Sc. Fx. Anx."" 1 1 C P I - C l i n i c a l D. Hy. Pd. Pt. Sc. S i . > S t a t i s t i c a l l y S t a t i s t i c a l l y SASB A f f i l i a t i o n , Autonomy Component S i g n i f i c a n t P a r e n t a l S i g n i f i c a n t Measures V a r i a b l e s Mother CPI Sp. Sa. Wb. So. Sc. Fx. Anx."" C P I - C l i n i c a l D. Hy. Pd. Pt. Sc. S i . > V a r i a b l e s Grouped SASB A f f i l i a t i o n , Autonomy _ Analyzed > A c c o r d i n g to Father CPI Sp. Sa. Wb. So. Sc. Fx. Anx." 1 Grouped C P I - C l i n i c a l D. Hy. Pd. Pt. Sc. S i . > Se p a r a t e l y E c o l o g i c a l SASB A f f i l i a t i o n , Autonomy _ 1 According to Model Family SASB A f f i l i a t i o n , Autonomy u s i n g and Analyzed Daughter Father System L e v e l s > Rates S e l f H o t e l l i n g ' s u s i n g Mother S e l f J and Analyzed Stepwise t e s t ( T 2 ) > Mother Daughter 1 L using D i s c r i m i n a n t Rates S e l f f J p l us H o t e l l i n g 1 s plus Father Daughter L Rates S e l f t e s t ( T 2 ) D i s c r i m i n a n t - D i s c r i m i n a n t FES C. Con. Ind. Org. C t l . r J plus Analys i s Community PPI S i z e , Q u a l i t y > A n a l y s i s D i s c r i m i n a n t A n a l y s i s y Note: Sp= Anx Sc = S o c i a l Presence; =Anxiety; D=Depre Schizophrenia; S i Sa=Self-Acceptance; Wb=Well-Being; S o = S o c i a l i z a t i o n ; F x = F l e x i b i l i t y ; s s i o n ; Hy=Hysteria; Pd=Psychopathic Deviate; Pt=Psychasthenia; =Social I n t r o v e r s i o n ; C=Cohesion; Con=Conflict; Ind=Independence; Org=0rganization; Ct l = C o n t r o l 69 determine which variables differentiated between the two groups at each system level. The degree to which the variables correctly classified the subjects into the anorexic and control groups was also /established. The alpha level for the simultaneous multiple comparisons was set at .05. These analyses addressed the next four hypotheses. The third step in the analysis was the combining of the systems into the larger framework, the Ecological-Systems model, as outlined in Table 5. Statistically significant variables from all four systems were analysed together using a stepwise discriminant analysis. The order of the variable inclusion was not specified. This method of discriminant analysis allowed for the variable with the greatest discriminant power to enter into the equation first followed by the next best at each successive step until all variables with discriminant power had been entered. This means that independent variables containing excess information or those with low discriminant power are not selected to enter the equation. By such a method the number of variables is reduced to include only those which have the greatest discriminant power. The alpha level for this analysis was set at .10 to ensure that all variables with discriminant power would enter the equation (i.e., to reduce possible Type II error). The stepwise discriminant analysis was calculated using the SPSS Version 9.00 Discriminant Analysis program (Nie, Hull, Jenkins, Steinbenner & Bent 1975). Finally, a discriminant analysis of the variables selected by the stepwise procedure, plus variables selected on conceptual and statistical grounds from other systems not represented in the stepwise analysis, was conducted using O W M A R , to determine to what degree each system added to the discriminant power of the model. These analyses answered the question posed by the last hypothesis. Other details of the analyses will be included with the results. The mean scores of the anorexic and control subjects and their parents on the CPI, CPI-Clinical, and FES measures were compared to the population means reported in the respective manual to determine how representative the samples were of the population at large. The statistical procedure for comparing the subjects with the population means was a one sample T 2 test 70 CHAPTER IV RESULTS The results of the analyses are divided into two sections. First the results of the demographic data comparing the anorexic and control families will be reported, followed by the results of the data analyses related to the hypotheses. Results of the Demographic Data Analyses Several analyses were conducted comparing the anorexic and control families on a number of demographic variables suggested in the literature as important characteristics which distinquish the anorexic and her family from non-anorexic families. Means, standard deviations and the results of the T 2 analyses for the variable set including socioeconomic status, number of stressors and illnesses, number of family members who dieted, and the duration of residence are reported in Table 6. The T 2 analysis revealed no statistically significant differences between the two groups (F=1.62, p>15). The nonsignificant findings for socioeconomic status indicated the matching procedures for the anorexic and control families on this variable were adequate. Calculation of the socioeconomic status of the families with the anorexic member according to Blishen's Socioeconomic Index for the experimental families revealed that 70% of the families were above an index of 50. The index 50 to 70+ represents business, managerial, and professional personnel. 71 Table 6 Means and Standard Deviations for Demographic Data for the Families of the Anorexic and Control Subjects Families of Families i of Anorexics Controls (n = 30) (n = 34) Characteristic M SD M SD Socioeconomic Status 54.60 13.02 57.00 14.22 Stressors 3.25 1.60 4.03 2.25 Illnesses 1.89 1.54 2.09 1.44 Dieters 1.50 0.63 1.21 0.97 Duration of Residence(yrs.) 8.10 5.87 11.21 7.27 Note: Stressors=geographic moves, death of family member, relative or close friend. Illnesses=chronic or wasting diseases, operations etc. suffered by a family member. Dieters=number of family members who dieted in past four years. Duration of Residence = number of years in present location. For Hotelling's T 2 analysis, F(5,54) = 1.62, p.15. None of the five simultaneous multiple comparisons were significant While there was no statistically significant difference in the number of stressors or illnesses, a further breakdown of these two variables according to Garfinkel and Garner's (1982) classification was undertaken. The number and percentages of the stress related factors for families and the total number of stressors are reported in Table 7. 72 Table 7 Stressors and Illnesses for Families of the Anorexic and Control Subjects Families of Anorexics Families of Controls Families Stressors Families Stessors (n = 30) (n= 100) (n= 34) (n = 150) Variable N % N % N % N % Separation & Loss 21 67 33 33 24 72 45 30 Death/Separation 10 19 Leaving Home(self/sib) 12 13 Geographic Move 11 13 Disruption of Family 26 82 45 45 31 93 76 52 Family Conflict 10 9 Financial/Job Insecurity 10 21 Serious Illness 20 32 Other 5 14 Environmental Demands 12 40 14 14 9 27 22 15 New Relationships 3 2 Change in Schooling 6 10 Other 5 10 Threat to Self-Esteem 4 13 4 4 0 0 0 0 Physical Illness 4 13 4 4 5 15 5 3 The calculation of percentages for the five stress related factors indicated the anorexic and control families had a similar proportion of separation/loss events, disruption of family homeostasis, environmental demands, threat to self-esteem, and personal illness of the anorexic or control subjects in the past four years or the four years prior to the onset of anorexia. It is interesting to note that while the anorexics and their families cited four incidents where self-esteem was threatened (academic, vocational or perceived social failures), the control subjects and their families reported no such events. Table 8 reports numbers and percentages of divorces, separations, family size, birth order, and the number of male and female children in the family. Percentages and number of separations and divorced/remarried are included for comparison with national 73 averages in Chapter V. Single-parent families, step-parents and adoptive parents were taken into consideration as part of the matching procedure. Table 8 Family Variables: Percentages for Families of Anorexic and Control Subjects Families of Families of Anorexics Controls (n = 30) (n = 34) Variable N % N % Separations 4 13 5 13 Divorced/Remarried 4 13 4 15 Familv Size <2 14 47 15 44 >2 16 53 19 56 Birth Order Oldest 5 17 7 30 Middle 5 17 3 9 Youngest 12 40 16 48 Only 4 14 2 6 Number of Girls 57 69 73 70 Number of Boys 26 31 31 30 Number Al l Girl Families 12 40 11 32 The findings in Table 8 show that the family size and the percentage of boys and girls are similar in the anorexic and control families. Some percentage-wise differences are noted in the birth order and the number of all girl families. The differences and similarities for the anorexic families will be compared and contrasted with the findings of the literature in Chapter V. Heights, education, and age of the subjects and heights, weights, education and age of their parents are presented in Table 9. Parental ages were calculated to reflect the age at the time of the daughter's birth. Weights of the anorexic and control subjects are not included in the analysis as the loss of weight is a required diagnostic criterion of anorexia nervosa. 74 Table 9 Means and Standard Deviations for Heights, Weights, Education, and Ages of Subjects, Mothers, and Fathers Anorexics Controls Variable M SD M SD (A) Subjects (n=30) (n= 34) Height (Cm) 164.06 5.67 165.17 7.48 Education (yrs) 12.03 2.02 12.29 1.89 Age (yrs) 18.36 2.29 18.55 2.73 (B) Mothers (n=29) (n= 32) Height (Cm) 163.51 6.48 164.54 7.09 Weight (Kg) 63.13 12.74 66.75 15.18 Education (yrs) 13.13 3.18 12.75 2.06 Age (yrs)1 28.31 5.56 29.33 5.64 (CJ Fathers (n=24) (n= 29) Height (Cm) 176.75 6.50 178.43 6.76 Weight (Kg) 79.95 9.82 82.00 8.51 Education (yrs) 14.08 4.08 13.79 3.68 Age (yrs)1 30.16 4.32 31.73 4.55 Note: Hotelling's T 2 analysis Hotelling's T 2 analysis for for subjects, mothers, F(3,60) = 0.25, F(4,56)=0.46, /?>.85; P>.15; Hotelling's T 2 analysis for fathers, F(4,48) = 1.08, p>35; none of the multiple comparisons were significant at the .05 level. 'Age at birth of daughter. No statistically significant results were found between the heights, education, and ages of the anorexic and control subjects as shown by the T 2 analysis (F=0.25, p>.B5). As age was one of the variables which was controlled for in this study, the nonsignificant finding indicates the matching procedure was adequate. The T 2 analyses also revealed that the two groups of fathers and mothers were not significandy different on height, weight, education, and their age at the time of the daughters birth (for mothers, F=0.46, p>.15; for fathers, F=1.08, p>35). Two other analyses were conducted: to determine 1) whether the control group was representative of the general population, and 2) whether a significant difference existed 75 between the anorexic abstainer and anorexic bulimic subjects on the component measures. A single sample T 2 test was used to compare the mean scores for the control subjects and their parents with the general population on the CPI, CPI-Clinical, and the FES variables. The results are found in Appendix A. An overall statistically significant difference was found with the T 2 analysis between the control subjects and the population means (F(18,16)=4.39, /?<.005) and between their mothers and the respective population means (F(18,12)=9.15, p<.0005) on the CPI scales (note Tables A - l , A-3). Although significance was found, all mean scores fell within one standard deviation of the norm population mean (note Graphs A - l , A-3) . In order to reduce the probability of Type II error in this part of the analyses, univariate t-tests, for the variables selected for this study (Social Presence, Self-acceptance, Sense of Well-being, Socialization, Self-control, Flexibility), were run and were found to be non-significant except for Social Presence (jp<.01) for the control subjects, the mean for which was above the norm group mean, and Sense of Well-being (j?<.05) for the mothers of the controls which was below the mean. The T 2 analysis for the fathers of the control subjects and the population means was non-significant (F(18,6) = 1.90, /7>.20, note Table A-5) . The results of the investigation for the CPI-Clinical scores comparing the control subjects and their parents with the appropriate population means are presented in Tables A - l , A - 3 and A - 5 . The findings of the T 2 analyses reveal statistically significant differences for the control subjects (F(13,21) = 4.85, ^<.0005), for the mothers (F(13,17) = 45.29, /K.0001) and for the fathers (F(13,ll) = 26.73, /K.0001). Al l univariate t-tests were statistically significant for the variables selected in this study (Depression, Hysteria, Psychopathic Deviancy, Psychasthenia, Schizophrenia, Social Introversion) except for the Social Introversion scores for the control subjects (pXlfJ) and for their fathers (j?>.30). However, as noted in Graphs A - 2 , A - 4 , and A - 6 , when the scores for the selected scales were converted to standard T-scores, they fell within one standard deviation of the mean except for the scores of the fathers of the control subjects on the Depression (T-score=63) and Hysteria (T-score=65) scales which were within two standard deviations of the mean. 76 As one of the criteria for selecting the control subjects was the absence of any known psychosomatic or psychiatric illness, differences on the CPI and CPI-Clinical scales were not anticipated. However, the significant differences found by these analyses suggest that the control subjects are not representative of the general population and may reflect changes which take place when individuals are faced with a physiological illness. Further, some of the elevation in the CPI-Clinical scores may be accounted for by the method used to calculate these scores. Rodgers (1966) found that the procedure used to estimate the MMPI equivalent scores from the CPI items "slightly overestimated the pathology in the normal sample" (p. 89). The T 2 analysis was also statistically significant when the families of the control subjects were compared to the population means for a four member family on the FES (F(10,24)=3.80, p<.01). Of the variables selected for this study, (Cohesion, Conflict Independence, Organization, Control) the dimensions of Cohesion and Independence were significandy higher (p<.05) on the univariate t-tests and the dimensions of Conflict and Control significandy lower (jsK.001) than the population means (note Table A-7). All these scores, however, were within one standard deviation of the mean (note Graph A-7). Again the differences may be accounted for by the referral sources of the control subjects and could reflect changes within the family environment when a family is under stressful circumstances. The significant findings on the CPI, CPI-Clinical and FES measures between the control subjects and the norm population means indicates the illness factor has possibly been accounted for by the selection method of the control group. Table 10 presents the means and standard deviations for the anorexic abstainer and anorexic bulimic subjects on the CPI, CPI-Clinical, and SASB measures. A T 2 analysis yielded no statistically significant results between the anorexic abstainers and anorexic bulimics on the CPI (F=0.79, p>.55), the CPI-Clinical (F=0.63, /?>.70), and the SASB (F=0.73, p>A5). These two groups were therefore considered equivalent and their scores combined for the remaining statistical analyses. T a b l e 10 Means and Standard Deviadons for the Personality Traits, Clinical Characteristics, and Self-Concept for Abstaining and Bulimic Anorexics Abstainers Bulimics (n = 19) (n = l l ) Variable M SD M SD CPI Social Presence 30.68 9.30 31.00 5.27 Self-Acceptance 19.05 4.08 19.52 2.79 Sense of Well-being 26.78 8.81 21.54 10.26 Socialization 31.68 7.22 31.72 3.79 Self-Control 23.15 9.73 17.72 10.72 Flexibility 8.89 3.85 8.27 4.73 Anxiety 7.89 3.51 10.36 4.82 CPI- Clinical Depression 27.15 6.75 28.63 3.38 Hysteria 26.63 4.90 28.45 4.39 Psychopathic Deviate 23.26 4.98 25.18 5.60 Psychasthenia 23.52 11.23 29.72 8.21 Schizophrenia 27.00 13.72 33.54 14.45 Social Introversion 29.84 9.79 29.72 4.31 SASB Affiliation -3.94 75.48 -34.81 66.17 Autonomy -49.36 40.28 -44.45 47.01 Note: For CPI Hotelling's T 2 analysis, F(7,22)=0.79, p>.55. For CPI-Clinical Hotelling's T 2 analysis, F(6,23) = 0.63, p>J0. For SASB Hotelling's T 2 analysis, F(2,27) = 0.73, p>A5. None of the simultaneous multiple comparison between sample means were significant at the .05 level. 78 Results of Data Analyses Related to the Hypotheses In the following section, the hypotheses are restated for clarity of presentation and to provide consistent structure. The ordering of the hypotheses and the reported results are according to the Individual, Parent, Family, Community and Ecological-System. Within each system, the results of the component measures are reported first (Step I of the statistical procedures), then the findings for the system analyses (Step II). The results of the Ecological-System (Step III) are reported after the findings of the other systems. Means and standard deviation of the statistically significant variables found at Step I of the statistical procedures (component measures), are reported again in the tables at the system level analyses. A. Analyses of the Scores in the Individual System Analyses in the Individual system investigated the personality traits, clinical characteristics and self-concept qualities of the anorexic and control subjects according to the following hypotheses: Hypothesis 1 The Personality Traits and Clinical Characteristics, as measured by the CPI, of the anorexic and control subjects will show a statistically significant difference. Hypothesis 2 The Self-concept qualities of the anorexic and control subjects will show a statistically significant difference on the dimensions of Affiliation and Autonomy as measured by the SASB. Means, standard deviations and the results of the multivariate analyses for the comparison of the anorexic and the control subjects on the CPI, CPI-Clinical, and SASB are found in Table 11. The results for the personality traits (CPI) will be discussed first 79 These will be followed by the results for the clinical characteristics (CPI-Clinical), and then those for the self-concept qualities (SASB). Comparisons of the Personality Traits An overall statistically significant difference was found between the anorexic and control subjects on the six selected variables of the CPI using Hotelling's T 2 (F=4.88, p-COOOl). The simultaneous multiple comparisons showed statistically significant between-group differences on the Sense of Well-being and the Anxiety scales at the .01 level. The mean differences of the scores on the remaining five scales, (Social Presence, Self-Acceptance, Socialization, Self-Control, Flexibility) although showing the anorexic subjects to be lower, did not reach statistical significance. The discriminant function obtained was statistically significant ( x 2 =27.39, p<.0005) resulting in 81% of the subjects being correctiy classified. The variables contributing the most to the discriminant analysis were Anxiety, Socialization, and Sense of Well-being. Comparisons of the Clinical Characteristics An overall statistically significant difference was found using T 2 analysis (F=6.77, jp<.0001) for the clinical characteristics between the anorexic and control subjects as reported in Table 11. The simultaneous multiple comparisons showed statistically significant between-group differences on the Depression, Psychopathic Deviancy, Psychasthenia and Schizophrenia scales at the .01 level. The scores on the remaining two scales, (Hysteria and Social Introversion) were also higher for the anorexic subjects, but the difference did not reach statistical significance. The discriminant function obtained was statistically significant ( x 2 =31.23, /K.0001), resulting in 87.5% of the subjects being correctly classified. The variables contributing the most to the discriminant analysis were Schizophrenia, Depression, and Psychasthenia. 80 Table 11 Means, Standard Deviations and Results of the Multivariate Analysis of Personality Traits, Clinical Characteristics, and Self-Concept for the Anorexic and Control Subjects Anorexics Controls Discriminant (n = 30) (n = 34) Function Variable M SD M SD Coefficient1 CPI Social Presence 30.79 7.96 36.94 7.40 -.135 Self-Acceptance 19.23 3.61 22.11 7.82 -.140 Sense of Well-being 3 24.86 9.55 35.00 5.97 .407 Socialization 31.69 6.11 35.55 5.99 .515 Self-Control 21.16 10.27 29.05 8.39 -.375 Flexibility 8.66 4.13 12.50 8.12 .347 Anxiety3 8.76 4.13 4.61 2.25 -.526 CPh Clinical Depression3 27.69 5.72 21.52 4.64 .520 Hysteria 27.29 4.72 23.82 3.81 .218 Psychopathic Deviate3 23.96 5.20 17.70 4.08 .246 Psychasthenia3 25.79 10.52 14.76 7.63 -.258 Schizophrenia3 29.39 14.11 14.35 8.11 .728 Social Introversion 29.79 8.12 25.08 9.37 -.151 SASB Affiliation3 -15.26 72.63 112.64 47.53 .970 Autonomy -47.56 42.13 -22.41 21.95 .241 Note: For CPI Hotelling's T 2 analysis, F(7,56) = 4.88, /K.0001; for the discriminant analysis, X 2 (7) = 27.39, /K.0005. For CPI-Clinical Hotelling's T 2 analysis, F(6,57)=6.77, /K.0001; for the discriminant analysis, x 2 (6)=31.23, /K.0001. For SASB Hotelling's T 2 analysis, F(2,61) = 36.95, ;?<.0001; for the discriminant analysis, x 2 (2) = 47.62, ^<.0001. Coefficients are standardized-normalized discriminant function coefficients. 2The higher the absolute value of the coefficient the more the variable contributes to the discriminant function. 3Simultaneous multiple comparison between sample means, p<.0\. Comparisons of the Self-Concept Qualities The T 2 analysis revealed an overall statistically significant difference (F=39.95, /K.0001) between the anorexics and controls for the SASB (intrapsychic) self-concept measure as reported in Table 11. The simultaneous multiple comparison showed a 81 statistically significant between-group difference on the dimension of Affiliation at the .01 level. While the mean score on the Autonomy dimension was considerably lower for the anorexic subjects, the difference did not reach significance. The discriminant function was statistically significant ( x 2 =47.62, /K.0001) resulting in 84% of the subjects being correctly classified. The results of these analyses support Hypothesis 1 and 2 that statistically significant differences will be found between the personality traits, clinical characteristics and self-concept qualities of the anorexic and control subjects. Statistical differences found on the CPI scores in this study suggest that females with anorexia nervosa, when compared to matched control subjects, can be described as overemphasizing their worries and personal problems, they commiserate with themselves, resent circumstances more favorable than their own and are at odds with themselves and others. The higher anxiety score suggests anorexics are more anxious than the control subjects. The test scores on the CPI-Clinical scales yeilded several significant findings, namely, the Depression, Psychopathic Deviancy, Psychasthenia and Schizophrenia scales were significantly higher for the anorexics than for the controls. The higher depression score for the anorexics suggest that they viewed themselves as more pessimistic than did the controls, they had a lower energy level and experienced a greater loss of appetite. The higher Depression score also suggests that anorexics were more despondent and displayed less affect that did the controls. The significandy higher Psychopathic Deviancy score for the anorexics suggests that relative to the controls the anorexics were more rebellious and hostile toward authority figures and more non-conforming. The significantly higher Psychasthenia score for the anorexics would indicate they turned their feelings inward to a greater degree than did the controls. That is, instead of acting out rebellious and hostile tendencies, the anorexics were more likely to feel guilty about their thought and become self-punitive in either thoughts or actions. There may also be a larger component of obsessive thought patterns or ritualistic behaviours which would give them some sense of control or a release from decision-making and/or responsibility. The higher Schizophrenia 82 score for the anorexics suggests that they felt they were in some way losing control, which fits in with the ritualistic patterns suggested by the higher Psychasthenia score. This higher score for the Schizophrenia scale also suggests that anorexics feel alienated and lonely, even though they may have many superficial relationships. If surrounded by people they may feel they are on the outside looking in. The finding on the Schizophrenia scale also suggests that the anorexics did not have a sense of personal identity, i.e., who they were and what their real thoughts, feelings and desires were for the present or the future. To summarize the significant differences on the CPI-Clinical scales between the anorexics and the control subjects, the findings suggest that the anorexics were more depressed and had a greater tendency for rebelliousness and hostile resistance toward authority figures. Rather than acting directly on their feelings, the anorexics tended to turn conflicts inward, felt guilty about their thoughts and became self-punitive. The anorexics may have use ritualistic thoughts and actions in an attempt to gain a sense of control and put order in their lives. Finally, the anorexics most likely felt alienated, had little sense of who they were emotionally and what they could realistically expect or require of themselves. The significantly lower Affiliation score for the anorexics on the SASB suggests that they directed a greater degree of hostility toward themselves, that is, they felt more guilty, ashamed, unworthy and inadequate than did the control subjects. Comparisons of the Individual System Hypothesis 9 The mean scores at the Individual System level (personality traits, clinical characteristics, and self-concept) together will show a statistically significant difference between the anorexic and control subjects. 83 In order to test this hypothesis, the variables from the component measures (CPI, CPI-Clinical, SASB) in the individual system on which significance between-group differences had been found were analysed together using a T 2 test followed by a discriminant analysis. Means, standard deviations, and the results of the multivariate analyses for the seven variables (Sense of Well-being, Anxiety, Depression, Psychopathic Deviancy, Psychasthenia, Schizophrenia and Affiliation) are reported in Table 12. Table 12 Means, Standard Deviations and Results of the Multivariate Analysis of Individual System Variables for Anorexic and Control Subjects Anorexics Controls Discriminant (n = 30) (n = 34) Function Variable M SD M SD Coefficient1 2 Sense of Well-being 3 24.86 9.55 35.00 5.97 .199 Anxiety3 8.76 4.13 4.61 2.25 .191 Depression3 27.69 5.72 21.52 4.64 -.099 Psychopathic Deviate3 23.96 5.20 17.70 4.08 .000 Psychasthenia3 25.79 10.52 14.76 7.63 -.492 Schizophrenia3 29.39 14.11 14.35 8.11 -.308 Affiliation3 -15.26 72.63 112.64 47.53 .758 Note: For Hotelling's T2 analysis, F(7,56) = 10.44, /K.0001, for the discriminant analysis, x 2 (7) = 48.04, /K.0001. 1 Coefficients are standardized-normalized discriminant function coefficients. 2The higher the absolute value of the coefficient the more the variable contributes to the discriminant function. 'Simultaneous multiple comparison between sample means, p<.05. The T 2 analysis revealed an overall statistically significant difference between the anorexic and the control subjects (F= 10.44, /K.0001) on the seven variables. All simultaneous multiple comparisons showed statistically significant between-group differences at the .05 level. The discriminant function obtained was statistically significant (x 2 =48.04, /K.0001). The variables contributing the most to the discriminant analysis according to the 84 standardized-normalized discriminant function coefficients were Affiliation, Psychasthenia and Schizophrenia. Table 13 presents the proportion of subjects correcdy classified into the anorexic or control group, (the hit rate), according to the discriminant function. T a b l e 13 Individual System Classification by Means of the Discriminant Function Predicted Group Membership Number Actual Group Anorexic Control of Cases Anorexic 26 4 30 Control 4 30 34 Note: Hit rate (proportion of correct classification) =56/64=.875. Predictor variables (7) = Sense of Well-being, Anxiety, Depression, Psychopathic Deviate, Psychasthenia, Schizophrenia, Affiliation. The findings of these analyses support Hypothesis 9, that variables within the Individual System will statistically differentiate between the anorexic and control subjects. B. Analyses of the Scores in the Parent System Hypotheses at the Parent system level investigated the personality traits, clinical characteristis and the self-concept qualities of the mothers of the anorexics and controls and the fathers of the anorexics and controls. The results of the analyses comparing the mothers will be reported first, then the findings for the fathers. 85 Hypothesis 3 The Personality Traits and Clinical Characteristics, as measured by the CPI, of the mothers of the anorexic and control subjects will show a statistically significant difference. Hypothesis 4 The Self-concept qualities of the mothers of the anorexic and control subjects will show a statistically significant difference on the dimensions of Affiliation and Autonomy as measured by the SASB. Means, standard deviations and the results of the multivariate analyses for the comparison of the mothers of the anorexic and control subjects on the CPI, CPI-Clinical, and SASB are found in Table 14. The results of these analyses will be reported in the same order as the anorexic and control subjects. Comparisons of the Personality Traits of the Mothers No overall significant difference between the mothers of the anorexic and control subjects was found on the seven personality scales of the CPI using Hotelling's T 2 (F=1.70, />>.10). The discriminant function obtained was not statistically significant (x 2 =10.98, j>>.10). Comparisons of the Clinical Characteristics of the Mothers The T 2 analysis revealed an overall significant difference for the CPI-Clinical scales (F=3.67, /7-C005). The simultaneous multiple comparisons showed a statistically significant difference on the Psychopathic Deviancy scale at the .01 level, with the mothers of the anorexic subjects having the higher mean score. While the mean scores for the other five scales (Depression, Hysteria, Psychasthenia, Schizophrenia and Social Introversion) were also higher for these mothers, the difference did not reach significance. The discriminant function was statistically significant (x 2 =18.56, /K.005) resulting in 78.5% of the mothers 86 Table 14 Means, Standard Deviations and Results of the Multivariate Analysis of Personality Traits, Clinical Characteristics, and Self-Concept for the Mothers of the Anorexic and Control Subjects Mothers of Mothers of Anorexics Controls Discriminant (n= -26) (n = 30) Function Variable M SD M SD Coefficient1 2 CPI Social Presence 32.84 5.30 33.13 5.30 -.580 Self-Acceptance 19.00 3.32 20.13 4.89 .465 Sense of Well-being 33.26 6.25 36.39 3.52 .306 Socialization 38.84 5.13 38.66 4.58 .540 Self-Control 31.23 6.46 33.76 6.52 .062 Flexibility 10.23 6.63 10.16 3.98 .023 Anxiety 5.26 2.47 4.06 1.72 -.235 CPI-Clinical Depression 25.07 5.62 23.96 4.55 -.485 Hysteria 26.19 5.23 24.63 4.62 .246 Psychopathic Deviate3 19.57 4.62 15.09 2.53 .804 Psychasthenia 15.46 7.55 11.06 6.18 .195 Schizophrenia 15.23 9.65 9.40 4.37 .014 Social Introversion 28.96 7.44 28.83 9.38 .133 SASB Affiliation 72.75 58.89 101.03 40.45 .898 Autonomy -9.08 25.47 -11.70 22.69 -.439 Note: For CPI Hotelling's T 2 analysis, F(7,48) = = 1.70, p<.!0; for the discriminant analysis, X 2 (7) = 10.98, p<.10. For CPI-Clinical Hotelling's T 2 analysis, F(6,49) = 3.67, /K.005; for the discriminant analysis, x 2 (6) = 18.56, /K.005. For SASB Hotelling's T 2 analysis, F(2,51) = 2.78, p<.05; for the discriminant analysis, X 2 (2) = 5.18, p<.05. Coefficients are standardized-normalized discriminant function coefficients. 2The higher the absolute value of the coefficient the more the variable contributes to the discriminant function. 'Simultaneous multiple comparison between sample means, ^<.01. being correctly classified according to the discriminant analysis. The variables contributing the most to the discriminant function were Psychopathic Deviancy and Depression. 87 Comparisons of the Self-Concept Qualities of the Mothers No overall statistical significance was found between the mothers on the SASB measure using the Hotelling's T 2 procedure (F=2.78, p>.05). The discriminant function obtained was not statistically significant (x2=5.18, p>.05). The results of the analysis for the CPI-Clinical scales support Hypothesis 3 in that a statistically significant difference exists between the Psychopathic Deviancy scores of the mothers of the anorexic and control subjects. No statistically signficant differences were found for the CPI and SASB mean scores. Hypothesis 5 The Personality Traits and Clinical Characteristics, as measured by the CPI, of the fathers of the anorexic and control subjects will show a statistically significant difference. Hypothesis 6 The Self-Concept Qualities of the fathers of the anorexic and control subjects will show a statistically significant difference on the dimensions of Affiliation and Autonomy as measured by the SASB. Comparisons of Personality Traits of the Fathers The means, standard deviations and the results of the multivariate analysis for the comparison of the fathers of the anorexic and control subjects on the CPI, CPI-Clinical and the SASB are reported in Table 15. No overall statistically significant difference was found at the .01 level for the personality traits (CPI) between the two groups of fathers using Hotelling's T 2 (F=1.74, p>.\0). The obtained discriminant function ( x 2 =10.94, p>10) was not significant 88 Comparisons of the Clinical Characteristics of the Fathers The P anlaysis revealed no overall significant mean difference on the CPI-Clinical scales between the fathers of the anorexics and controls (F=2.31, p>.Q5) at the .01 level. The discriminant function ( x 2 =12.08, /7>.05) was not significant Comparisons of the Self-Concept Qualities of the Fathers The results of the T 2 analysis of the self-concept qualities (SASB) between the anorexic and control fathers revealed no overall statistical significance at the .01 level (F=0.82, p>A0). The discriminant function (x 2 =l-58, p>A5) was not significant The findings of these analyses do not support Hypothesis 5 and 6 that the CPI, CPI-Clinical and SASB mean scores will differentiate between the fathers of the anorexic and control subjects. The findings of this study show that the personalities of the fathers of the anorexics (at least insofar as assessed by the CPI) are not demonstrably different than the fathers of the control subjects. Comparisons of the Parent System Hypothesis 10 The mean scores at the Parent System level (personality traits, clinical characteristics, and self-concept qualities of the mothers and fathers), together, will show a statistically significant difference between the parents of the anorexic and the control subjects. In order to test this hypothesis statistically significant variables from the component measures for both the mothers and the fathers were to be analysed together. As only one variable, the mothers' Psychopathic Deviancy mean score, was significantiy different at the first level of analysis, the component measures, further investigation was not undertaken. Using the information from the discriminant analysis of the mothers' CPI-Clinical mean scores it was possible to correctiy classify 78.5% of the mothers according to the 89 Table 15 Means, Standard Deviations and Results of the Multivariate Analysis of Personality Traits, Clinical Characteristics, and Self-Concept for Fathers of the Anorexic and Control Subjects Fathers of Fathers of Anorexics Controls Discriminant (n = 19) (n=24) Function Variable M SD M SD Coefficient1 2 CPI Social Presence 34.89 6.12 35.79 6.44 -.027 Self-Acceptance 21.73 4.03 22.29 7.83 -.106 Sense of Well-being 34.52 6.32 39.41 5.50 .778 Socialization 33.52 6.97 38.62 5.09 .336 Self-Control 28.15 9.61 35.12 8.17 .016 Flexibility 9.47 3.71 9.04 8.88 -.440 Anxiety 4.68 2.58 3.83 1.49 .271 CPI- Clinical Depression 24.26 4.31 22.16 2.54 -.492 Hysteria 24.78 3.06 24.58 4.34 .245 Psychopathic Deviate 20.10 5.32 15.70 2.25 -.746 Psychasthenia 13.52 7.29 9.00 8.29 .048 Schizophrenia 14.78 9.30 8.58 6.51 -.095 Social Introversion 26.47 7.01 26.66 8.41 .358 SASB Affiliation 76.44 55.34 83.62 50.78 .476 Autonomy -31.55 27.46 -21.83 28.06 .879 Note: For CPI Hotelling's T 2 analysis, F(7,35) = 1.74, p>.10; for the discriminant analysis, X 2 (7) = 10.94, p>.10. For CPI-Clinical Hotelling's T 2 analysis, F(6,36) = 2.31, p>.05; for the discriminant analysis, x 2 (6) = 12.08, p>.05. For SASB Hotelling's T 2 analysis, F(2,39)=0.82, />>.40; for the discriminant analysis, X 2 (2) = 1.58, p>A5. 'Coefficients are standardized-normalized discriminant function coefficients. 2The higher the absolute value of the coefficient the more the variable contributes to the discriminant function. None of the simultaneous multiple comparison between sample means were significant at the .01 level. 90 discriminant function as reported in Table 16. Table 16 Parental System Classification by Means of the Discriminant Function Predicted Group Membership Number Actual Group Anorexic Control of Cases Anorexic 21 5 26 Control 7 23 30 Note: Hit rate (proportion of correct classification) =44/56 = .785. Predictor variables(6)= Scores on Mothers' Depression, Hysteria, Psychopathic Deviate, Psychasthenia, Schizophrenia, Social Introversion scales. The results of the analysis for the CPI-Clinical scores for the two groups of mothers supports Hypothesis 10, that data from the Parent system will statistically differentiate between the parents of the anorexic and control subjects. C. Analyses of the Scores in the Family System Two component measures were used at the Family system level, the SASB (interpersonal planes) and the FES. The results of the parent-daughter transactions, as measured by the SASB, will be reported first, then the family environment as measured by the FES. 91 Hypothesis 7 The perceived transactions of the parents toward their daughter (focus on other) and the responses of the daughter to their parents (focus on self) of the anorexic and control subjects will show a statistically significant difference on the dimensions of Affiliation and Autonomy as measured by SASB (interpersonal plane). This hypothesis was investigated using a series of Hotelling's T 2 analyses followed by discriminant analyses. Four analyses were required to investigate the parental transactions and four for the responses of the daughters. The results of the parental transactions will be presented first, then the responses of the daughters. Comparisons of the Parental Transactions Means, standard deviations and the results of the multivariate analyses for the comparison of mothers' and fathers' transactions (focus-on-other) in their relationship with their daughters as perceived by the daughter, mother, and father are reported in Table 17. The T 2 analysis revealed an overall significant difference between anorexic and control subjects for the daughters' perception (Daughter-Rates-Mother) of the mothers' transactions (F= 12.26, /K.0001). The simultaneous multiple comparisons showed a statistically significant difference between-groups (/K.01) on both the Affiliation and Autonomy dimensions. The discriminant function obtained was statistically significant ( x 2 =20.28, /K.0001), resulting in 78% of the anorexic and control subjects being correctiy classified according to the discriminant analysis. Perusal of the standardized-normalized discriminant function coefficients suggested that the Affiliation dimension contributed the most to the discriminant function. No overall statistically significant differences were found by a T 2 analysis for the daughters' perception of the fathers' transactions (Daughter-Rates-Father). (F=2.20, p>.lO), for the mothers' view of the transactions (Mother-Rates-Self) (F=2.64, p>.05), or for the fathers' view (Father-Rates-Self) (F=2.57, p>.05). The discriminant analyses for these three investigations were not statistically significant (note Table 17). 92 Table 17 Means, Standard Deviations and Results of the Multivariate Analysis of Parental Transactions with Their Daughters Discriminant Anorexic Control Function Variable M SD M SD Coefficient1 2 Daughter Rates Mother (n = = 30) (n = ,34) Affiliation3 53.16 68.28 117.91 45.11 .820 Autonomy3 -29.13 53.96 14.47 33.92 .571 Father (n = = 28) (n= = 32) Affiliation 54.96 80.91 90.50 57.51 .878 Autonomy -7.42 54.83 12.18 40.49 .476 Mother Rates Self (n = = 25) (n= = 30) Affiliation 117.35 41.80 131.56 18.92 .478 Autonomy 6.48 38.22 25.09 25.75 .877 Father Rates Self (n = = 19) (n = = 24) Affiliation 94.31 45.88 120.33 33.53 .924 Autonomy 5.36 36.66 16.91 27.73 .381 Note: Daughter-Rales-Mother - For Hotelling's T 2 analysis, F(2,61) = 12.26, /K.0001; for the discriminant analysis x 2 (2) = 20.28, /><.0001. Daughter-Rates-Father - for Hotelling's T 2 analysis F(2,57) = 2.20,/>>.10, for the discriminant analysis x 2 (2) = 4.18, p>.\Q. Mother- Rates- Self - for Hotelling's T 2 analysis F(2,52) = 2.64,p>.05, for the discriminant analysis x 2 (2) = 4.93, />>.05. Father- Rates-Self - for Hotelling's T 2 analysis F(2,40) = 2.57,/>>.05, for the discriminant analysis x 2 (2)=4.72, /?>.05. 'Coefficients are standardized-normalized discriminant function coefficients. 2 The higher the absolute value of the coefficient the more the variable contributes to the discriminant function. Simultaneous multiple comparison between sample means, /K .01 . The results of the analysis for the daughters' perception of their mothers supports Hypothesis 7, that the perceived transactions of the parents of the anorexic and control subjects will differentiate between the two groups on the dimensions of Affiliation and Autonomy. However, the findings for the daughter's perception of their fathers and the parent's perception of themselves do not support this hypothesis. 93 For a clearer understanding of the data, differences in the subjects' and parents' perceptions of the transactions are presented in the upper half of Figure 3. Here, the mother-daughter and father-daughter relationships as perceived by the daughter, mother, and father are presented on the left- and right-hand sides of the figure respectively. Dark figures represent the affiliation-autonomy vector for the anorexic group and the open figures represent the affiliation-autonomy vector for the control group. Inspection of the data in graphic form shows more clearly the difference in the anorexic and control subjects view of their mothers' behaviour in the relationship. While the Autonomy dimension, as viewed by the control subjects for their mothers is a positive score (indicative of general tolerance for separation and autonomy) the Autonomy dimension as perceived by the anorexics is a negative score (indicative of a controlling posture). A difference was also found on the Affiliation dimension which indicated the anorexic subjects perceived their mothers as being significantly less friendly in the relationship. As noted in the graph, the affiliation-autonomy vector for the anorexics' perception of their mothers falls in Quadrant III of the model. In keeping with the theory of SASB, such behaviour is described as overprotective and controlling. Comparisons of the Daughters Responses Means, standard deviations and results of the multivariate analyses comparing the anorexic and control daughters' responses (focus- on- self) in the relationship with their parents as perceived by the daughter, the mother, and the father are presented in Table 18. The T2 analysis for the daughters' perception of their responses in the relationship with their mothers (Daughter-Rates-Self-re-Mother) show an overall statistically significant difference (F=14.75, /7-C0001). The simultaneous multiple comparisons showed a between-group difference for the Affiliation dimension at the .01 level. The discriminant function obtained was statistically significant (x 2 =23.66, /K.0001) resulting in 80% of the 94 Parental Behaviour Mother-Daughter Relationship Father-Daughter Relationship 30 § • § 0 H 3 + -301-20 60 100 -I 1 1 1 1 1 1 h 140 OTHER lh O ' AD •M AFFILIATION 20 60 30 2 O § 0 H 5 + -3C4-OTHER 100 AFFILIATION 140 Daughters' Responses 30 s • H 5 + -30f 60 100 H r — H 1 1 1 1 h-140 SELF A D AFFILIATION 30f o z n o 0 H 5 + -30 ., 20 60 //-. 1 1 1 -— SELF II-100 140 H 1 I—I 1 1 r AFFILIATION Anorexic=solid figure; Control = open figure D=Daughter M = Mother F=father Figure 3 Perception of Interpersonal Transactions 95 Table 18 Means, Standard Deviations and Results of the Responses to Their Parents Multivariate Analysis of the Daughters Discriminant Anorexic Control Function Variable M SD M SD Coefficient1 2 Daughter Rates Self re Mother (n= = 30) (n = 34) Affiliation3 31.56 64.18 109.85 52.74 .978 Autonomy 14.40 38.94 22.08 31.60 .208 Self re Father (n = = 28) (n = 32) Affiliation3 23.75 85.79 98.09 50.27 .954 . Autonomy 15.82 44.02 20.31 40.15 .299 Mother Rates Daughter (n = = 25) (n = 30) Affiliation3 59.31 58.05 118.50 29.15 .995 Autonomy 14.36 38.97 16.06 30.02 .092 Father Rates -Daughter (n = = 19) (n= 24) Affiliation 38.21 64.35 97.77 60.60 .912 Autonomy 15.10 30.99 21.41 29.97 .408 Note: Daughter- Rates- Self-re- Mother - For Hotelling's T 2 analysis, F(2,61) = 14.75, /K.0001; for the discriminant analysis x 2 =23.66, /K.0001. Daughter-Rales-Self-re-Father - for Hotelling's T 2 analysis F(2,57) = 9.40,/K.0005, for the discriminant analysis x 2 (2)= 15.97, /K.0005. Mother- Rales- Daughter - for Hotelling's T 2 analysis F(2,52) = 11.86,/K.0001, for the discriminant analysis x 2 (2) = 19.17, /K.0001. Father-Rales-Daughter - for Hotelling's T 2 analysis F(2,40) = 5.85,/K.005, for the discriminant analysis x 2 (2) = 10.01, / K . 0 1 . Coefficients are standardized-normalized discriminant function coefficients. :The higher the absolute value of the coefficient the more the variable contributes to the discriminant function. 'Simultaneous multiple comparison between sample means, /K .01 . subjects being correcdy classified. The standardized-normalized discriminant function coefficients indicate that the Affiliation dimension contributed the most to the discriminant analysis. An overall statistically significant difference for the daughters' perception of their responses to their fathers (Daughter-Rates-Self-re-Father) was found using T 2 analysis 96 (F=9.40, p<.0005). The simultaneous multiple comparisons showed a statistically significant between-group difference on the Affiliation dimension at the .01 level. The discriminant function obtained was statistically significant ( x 2 =15.97, /?<.0005), resulting in 77% of the subjects being correctly classified. The variable contributing the most to the discriminant analysis was Affiliation. The T 3 analysis for the mothers' perception of how their daughters responded in the relationship (Mother-Rates-Daughter) revealed an overall statistically significant difference, (F= 11.86, /7-C0001). The simultaneous multiple comparisons showed a significant between-group difference for the dimension of Affiliation at the .01 level. The discriminant function obtained was statistically significant ( x 2 =19.17, /?<.0001), resulting in 82% of the subjects being correctly classified. The variable contributing the most to the discriminant analysis was Affiliation. Although an overall statistically significant difference was found using a P analysis for the fathers' view of their daughters' response (Father-Rates-Daughter) in the relationship (F=5.85, p<.0005), none of the simultaneous multiple comparisons showed a statistically significant between-group difference at the .01 level. The discriminant function ( x 2 =10.01, p<.01) was statistically significant, resulting in 77% of the subjects being correctly classified. Although the Affiliation dimension contributed the most to the discriminant analysis a considerable amount was also added by the Autonomy dimension. It can be seen from Table 18 that the data from these analyses support Hypothesis 7, that the responses of the anorexic and control subjects in their relationship with their parents will reveal a difference on the dimensions of Affiliation. The data for the differences of the perceived responses of the anorexic and control subjects as seen by the daughters, mothers, and fathers are presented graphically in the lower half of Figure 3. Inspection of the focus-on-self surface shows that the anorexic subjects rate themselves significantly lower on the dimension of Affiliation in their relationship with both their mothers and their fathers. The mothers of the anorexics also rated them as showing significantly less friendly behaviours in the relationship. These 97 findings, while not on the hostile side of the affiliation axis, are indicative of less friendliness being shown by the anorexics in theiT relationship with their parents. Comparisons of the Family Environment H y p o t h e s i s 8 Families of the anorexic and control subjects will perceive their family environments as significantly different when measured by the FES. Means, standard deviations and the results of the multivariate analysis of the five selected dimensions of the family environment as measured by the FES (Cohesion, Conflict Independence, Organization, Control) are presented in Table 19. T a b l e 19 Means, Standard Deviations and Results of the the Multivariate Analysis of the Family Environment Dimensions Anorexic Control Discriminant (n = 30) (n = 34) Function Variable M SD M SD Coefficient1 2 Cohesion3 44.06 10.67 54.05 8.37 -.815 Conflict 47.11 11.45 40.22 9.00 -.039 Independence 41.24 15.45 51.43 8.40 -.248 Organization 48.26 10.65 51.52 9.93 -.291 Control 50.62 10.73 45.15 9.19 .432 Note: Means are reported as Standard Scores. For Hotelling's T 2 analysis, F(5,58) = 3.86, p<.005; for the discriminant analysis, X 2 (5) = 16.82, /K.005. Coefficients are standardized-normalized discriminant function coefficients. 2The higher the absolute value of the coefficient the more the variable contributes to the discriminant function. Simultaneous multiple comparison between sample means, p<.05. 98 An overall statistically significant difference was found at the .01 level between the families of the anorexics and controls on the dimensions of the FES using a T 2 analysis (F=3.86, p<.005). The simultaneous multiple comparisons showed a statistically significant between-group difference on the Cohesion scale at the .05 level. As the results of the overall T 2 analysis were significant at the .01 level but none of the variables showed a between-group difference, an exception was made to accept the .05 level for the Cohesion variable. This allowed the inclusion of data from the FES to be analysed at the Family system level. The discriminant function obtained was statistically significant ( x 2 = 16.82, /><.005), resulting in 75% of the • subjects being correctiy classified. The standardized-normalized discriminant function coefficients indicated that Cohesion and Control were the variables which contributed the most to the discriminant analysis. Hypothesis 8, that the family environment dimensions investigated in this study will differentiate between the anorexic and control subjects, was marginally supported by the findings of these analyses. Comparisons of the Family System Hypothesis 11 The mean scores at the family system level (parent-child transactions and family environment), together, will show a statistically significant difference between the families of the anorexic and control subjects. This hypothesis was tested by analysing together statistically significant variables from the parent-daughter transactions (SASB) and the family environment dimensions (FES). As there was an unequal number of respondents in the three groups (daughters, mothers, fathers) combining statistically significant variables from all three groups for the parent-daughter transactions would result in the dropping of subjects and thus a decrease in the degrees of freedom. In order to maintain the highest possible number of subjects as well as reduce 99 the parent-daughter data to a manageable form it was decided to investigate the mean scores of the daughters' responses to their parents on the dimension of affiliation to determine if a significant difference existed between the daughters' and parents' perceptions. The dimension of affiliation was the variable where statistically significant differences were found between the anorexic and control subjects (note Table 18) and therefore the appropriate dimension to make a comparison. A 2x2x2 analysis of variance (ANOVA) was chosen to compare the three independent variables, the perceptions of the anorexics and control subjects, the perceptions of the mothers and the perceptions of the fathers. The results are included in Appendix B. As no significant statistical differences were found, the daughters' mean scores were used in further analyses. Table 20 reports the means, standard deviations, and results of the multivariate analysis for the investigation of the Daughter-Rates-Mother, Daughter-Rates-Self-re-Mother, Daughter-Rates-Self-re-Father, and Cohesion variables of the Family System. 100 Table 20 Means, Standard Deviations and Results of the Multivariate Analysis of the Family System Variables Variable Anorexic (n = 28) M SD Control (n = 32) M SD Discriminant Function Coefficient1 2 Daughter Rates Mother Affiliation3 Autonomy3 53.16 -29.13 68.28 53.96 117.91 14.47 45.11 33.92 -.275 .169 Self re Mother Affiliation3 31.56 64.18 109.85 52.74 .861 Self re Father Affiliation3 23.75 85.79 98.09 50.27 .270 Cohesion3 44.06 10.67 54.05 8.37 .284 Note: For Hotelling's T 2 analysis, F(5,54) = 5.66, /K.0005, discriminant analysis x 2 (5) = 22.96, /K.0005. 'Coefficients are standardized-normalized discriminant function coefficients. 2The higher the absolute value of the coefficient the more the variable contributes to the discriminant function. Simultaneous multiple comparison between sample means, p<.Q5. The T 2 analysis revealed an overall statistically significant difference (F=5.66, /K.0005) between the anorexic and control subjects on the five variables in the Family System. Al l simultaneous multiple comparisons yielded statistically significant between-group differences at the .05 level. The obtained discriminant function was statistically significant ( x 2 =22.96, /K.0005). The standardized-normalized discriminant function coefficients indicated that the degree of Affiliation in the daughters' responses to the mother (Daughter-Rates-Self-re-Mother) was the variable which contributed the most to the discriminant analysis. By using the information from the discriminant analysis at the Family System level it was possible to correctly classify 78% of the subjects. These results are presented in Table 21. 101 Table 21 Family System Classification by Means of the Discriminant Function Predicted Group Membership Number Actual Group Anorexic Control of Cases Anorexic 20 8 28 Control 5 27 32 Note: Hit rate (proportion of correct classification) =47/60 = .783. Predictor variables (5 )= Daughter- rates- Mother- Affiliation and Autonomy, Daughter-Rates-Self-re-Mother-Affiliation, Daughter- Rates- Self- re- Father- Affiliation, Cohesion. The results of this analysis supports Hypothesis 11, that variables within the Family System would differentiate between the families of the anorexic and control subjects. D. Analyses of the Scores in the Community System As only two variables (the Size and Quality of the social network) were investigated at the Community system level and both of these came from the same component measure (PPI), only one analysis was required to investigate the data. The results for the component measure and the system level will be reported together. The hypothesis used for this investigation was: Hypothesis 1 2 The mean scores at the community system level (social network size and quality), together will show a statistically significant difference between the anorexic and control subjects. 102 Comparisons of the Community Variables Means, standard deviations and the results of the multivariate analysis comparing the Size and Quality of the social network (PPI) of the families of the anorexics and controls are reported in Table 22. Table 22 Means, Standard Deviations and Results of the Multivariate Analysis of Social Network Variables for the Anorexic and Control Families Anorexics Controls Discriminant (n=30) (n = 34) Function Variable M SD M SD Coefficient1 2 Size 12.33 4.29 15.48 6.34 .968 Quality 0.15 0.26 0.12 0.28 -.250 Note: A positive quality score suggests more giving to than receiving from For Hotelling's T 2 analysis, F(2,61) = 2.76, p>.05, for the discriminant analysis, X 2 (2) = 5.21, p>.Q5. Coefficients are standardized-normalized discriminant function coefficients. 2The higher the absolute value of the coefficient the more the variable contributes to the discriminant function. None of the simultaneous multiple comparison between sample means were significant at the .01 level. Neither the results of the T 2 analysis (F=2.76, p>.05) nor the discriminant analysis (x2=5.21, p>.05) showed a statistically significant difference between the families of the anorexics and the controls. Although the discriminant function was non-significant (x2=5.21, p>.Q5), the proportion of subjects correctiy classified according to the discriminant function was 61%. This finding is reported in Table 23. 103 T a b l e 2 3 Community System Classification by Means of the Discriminant Function Predicted Group Membership Number Actual Group Anorexic Control of Cases Anorexic 17 13 30 Control 12 22 34 Note: Hit rate Predictor (proportion variables(2) of correct classification) = = Social Network Size and 39/64=.609. Qualtiy. The results of this analysis do not support Hypothesis 12, that the Size and Quality of the social network will statistically differentiate between the anorexic and control subjects. E. Analyses of the Ecological-System It is clear from the previous analyses that the Individual, Parent and Family systems contributes significant information to the understanding of anorexia nervosa. The integration of this information (along with the variables from the Community system) into the Ecological-Systems approach was the next step undertaken to develop a greater understanding of the anorexic. The analyses for the Ecological-Systems model was the final step in the statistical procedure (Table 5) and addresses Hypothesis 13. H y p o t h e s i s 13 When all four individual systems are combined into a comprehensive Ecological-Systems model and statistically analysed, the four systems together will differentiate to a greater degree between the anorexic and control subjects than will each of the systems independently. 104 The investigation of this hypothesis required that the statistically significant variables within each system be analysed together. A stepwise discriminant analysis was the statistical procedure selected. Fourteen independent variables were entered into the equation; seven from the Individual System (Well-being, Anxiety, Depression, Psychopathic Deviancy, Psychasthenia, Schizophrenia, and Affiliation); one from the Parental System (mothers' Psychopathic Deviancy); four from the Family System (Daughter-Rates-Mother-Affiliation and Autonomy, Daughter-Rates-Self-re-Mother-Affiliation, and Cohesion); and two from the Community (Size and Quality). While Daughter-Rates-Self-re-Father-Affiliation was also significant at the system level, this variable was not included in the stepwise procedure. The decision to exclude this variable was made 1) to maximize the number of subjects in the analysis and thus conserve degrees of freedom; 2) because statistically, the variable Daughter-Rates-Self-re-Father-Affiliation, did not add to the discriminant analysis (note Table 20); and 3) conceptually on the basis of the literature which emphasizes the mother-daughter relationship over the father-daughter relationship. Variables at the community level were included, although not statistically significant at the system level, in order that all four systems would be represented in the analysis. This is in keeping with the requirements of the Hypothesis 13. The results of the stepwise discriminant analysis are reported in Table 24. 105 Table 24 Results of the Stepwise Discriminant Analysis Step Variable Entered Wilks' X at each stage at each stage Standardized Discriminant Function Coefficient Affiliation Psychopathic Deviate .475 .407 59.63 38.55 <.001 <.001 .857 -.491 Canonical Correlation .77 Wilks' X .407 47.60 P <.001 % Correct Classification 87.5 Two variables were selected by this analysis. The daughters' Affiliation mean score (Individual System) entered the equation first, followed by the mothers' Psychopathic Deviate mean score (Parent System). These findings indicated that these two variables discriminate to the greatest degree between the anorexic and control subjects. The results of the discriminant analysis showed that the two variables selected by the discriminant procedure could correcdy classify 87.5% of the subjects. These results are reported in Table 25. 106 Table 25 Integrated Systems Classification by Means of the Discriminant Function Predicted Group Membership Number Actual Group Anorexic Control of Cases Anorexic 21 5 26 Control 2 28 30 Note: Hit rate (proportion of correct classification) =49/56=.875. Predictor variables(4) = Affiliation (Individual System), Psychopathic Deviate (Parental System), Daughter Rates Self re Mother (Family System), Social Network Size (Community System). Recalling the results of the discriminant analysis from the four system levels (Tables 13, 16, 21, and 23), we note that information from the Individual System correctly classified 87.5% of the cases, the Parent System 78.5%, the Family System 78% and the Community only 61% (non-significant). The discriminant function analysis when the four systems were integrated correctly classified 87.5% of the subjects. While there is an increased rate of discrimination over that obtained for the Parent Family, and Community systems, the rate of correct classification is not increased over the Individual System. In order to determine how much each of the significant variables added to the discriminant power of the model a series of discriminant analyses were conducted. That is, to what degree did the daughters' Affiliation variable add to the discriminant power, the mothers' Psychopathic Deviancy variable, the Daughter-Rates-Self-Re-Mother, and the Size of the social network? While the last two variables were not selected by the stepwise discriminant procedure, a decision was made to include them in order to meet the requirements of Hypothesis 13. They were selected over other variables within their respective systems because 1) conceptually, they had greater support in the literature, and 2) statistically, these variables, at 107 the appropriate systems level, contributed the most to the discriminant analyses (note Tables 20 and 22). It can be noted from the Hit Rates (proportion of subjects correctiy classified) reported in Table 26 that variables from the Individual and Parent systems combined will discriminate to a slightly greater degree (.875) than will each variable individually (.839 and .833). However, the discriminant function is not increased at all with the addition of either the Family or Community variables. Table 26 Summary of Proportion of Correct Classification of Subjects by Means of the Discriminant Function Analysis for the Ecological-Systems Model Variable System Discriminant Function Coefficient Proportion of Correct Classification Affiliation Psychopathic Deviate Daughter Rates Self re Mother Size of Social Network Individual Parent Family Community .821 .513 .222 .107 .839 .833 .875 } .875 J .875 Note: Discriminant function coefficients are standardized-normalized. Hypothesis 13, an integration of the four systems will differentiate between the anorexic and control subjects to a greater degree than will each of the systems independentiy, is not supported by these findings. 108 Summary of Results The data were analysed and results presented according to each of the defined levels of the Ecological-Systems model. The Individual System representing the personality traits, clinical characteristics, and self-concept qualities of the anorexic and control subjects yielded several statistically significant findings. First the anorexic and control subjects could be differentiated from each other by the component measures. Second, when the statistically significant variables from each of the measures were analysed as a system, these variables differentiated between the anorexic and control subjects. Thus, the Ecological-Systems model discriminated significandy between the anorexic and control subjects at Step I and Step II of the data analyses procedures for the Individual system. The hit rate for the Individual System was 87.5%. Specifically, the results of these analyses indicate that the anorexics could be described, when compared to the control subjects, as being at odds with themselves and others, overemphasizing their personal problems (Sense of Well-being) and having a negative self-concept (low Affiliation score). The anorexics were more depressed, anxious and had a greater tendency toward rebelliousness and hostile resistance to authority figures. Rather than directiy act out their feelings, the anorexics turned conflicts inward, felt guilty about their thoughts and feelings and became self-punitive. They may also have experienced a sense of losing control and used ritualistic thoughts and/or actions to put more order in their lives. The anorexics most likely felt alienated and had little sense of who they were emotionally and what they could realistically expect or require from themselves. The Affilitation score was the variable which had the greatest discriminant power at the Individual System level of the Ecological-System model. The Parent System represents the personality traits, the clinical characteristics and the self-concept qualities of the mothers and fathers of the anorexic and control subject Data at this level were analyzed in the same manner as the data for the Individual System. Only one variable, the mothers' Psychopathic Deviancy mean score, yielded a significant 109 finding, with the mothers of the anorexic subjects having the higher score. No statistically significant differences existed between the fathers of the anorexic and control subjects. A 78.5% rate of correct classification was possible with the Parent System variables alone. Variables from the Family System representing parent-daughter transactions and the quality of the family environment yielded several significant findings. First statistically significant differences were found on all the component measures between the anorexic and control subjects. Also when the statistically significant variables were analysed as a system, the results indicated that the Ecological-Systems model was able to discriminate between the anorexic and control subjects. These data alone correctiy classified 78.3% of the subjects. Specifically these findings showed that the anorexics rated their mothers as being more controlling and less friendly in their interactions than did the control subjects. The mothers of the anorexics perceived their daughters as responding to them in a less affiliative manner than did the mothers of the control subjects. Also, the anorexics viewed themselves as responding to both their mothers and their fathers with less affiliation than did the control subjects. The families of the anorexics view their family as being less cohesive. This suggests that the family members are perceived as being less supportive, helpful and committed to each other than the family members in the control families. At the fourth level, the Community, the component measure yielded no statistically significant results, which suggests no difference is evidenced in the size and quality of the social networks of the anorexic and control families. Variables which entered the stepwise discriminant analysis came from two systems only, the Individual and the Parent The variables with the greatest discriminant power were the daughters' Affiliation and the mothers' Psychopathic Deviancy mean scores. Other variables from the Family System and the Community did not add to the discriminant power of the model. A hit rate of 87.5% was possible with the two variables selected by the stepwise procedure. 110 CHAPTER V DISCUSSION The purpose of this study was to apply aspects of the ecological system approach to an understanding of anorexia nervosa. Further, the study was designed to empirically investigate the interactive factors of the anorexic and control subjects at the individual, parent, family and community levels using multivariate statistical techniques. The model defined for this study, was a modification of Conger's (1981) Ecological-Systems approach which / was based on the principles of human development espoused by Bronfenbrenner (1977, 1979). The Ecological-Systems model directiy addressed four contexts in which anorexics function and are specified as the Individual, Parent, Family, and Community Systems. Discussion of the results will begin with the demographic characteristics of the families followed by the findings related to the Individual, Parent, Family and Community systems, and then the total Ecological-Systems model. Discussion of the Demographic Characteristics The family of the anorexic is considered to have specific distinguishing demographic features. The most commonly reported significant findings are the preponderance of families from the upper-middle and upper social classes (Crisp et al., 1980; Garfinkel & Gamer, 1982; Hall, 1978; Kalucy et al., 1977; Kay et al., 1967) and the increased age of the parents at the time of the child's birth (Bruch, 1973; Dally, 1969; Garfinkel & Gamer, 1982; Hall, 1978; Theander, 1970). Findings in the present study support the literature regarding social class as the majority (70%) of the anorexics came from families where the head of the household was in a professional, managerial, or business career. However, the trend toward more lower class females developing anorexia nervosa (Crisp, 1980) was also reflected in the present study as anorexics whose parents were labourers and one whose parent was on welfare were subjects in the study. I l l Garfinkel and Garner (1982) point out that the findings in the literature stating many of the parents of anorexics were older when their daughters were born (Bruch, 1973; Dally, 1969; Hall, 1978; Theander, 1970) needs further study where social class has been controlled. They state that "the relative increase in parental ages may be a phenomenon of being upper middle class and its characteristic pattern of delayed parenthood" (Garfinkel & Garner, 1982, p. 169). In the present study, the control families were matched for socioeconomic status thereby controlling for the social class variable. The findings of this study revealed the ages of the parents of the anorexics and the parents of the control subjects were nonsignificant The majority of mothers (69%) and 43%' of the fathers were below the age of 30 when their daughter was born. The average age of the mothers of the anorexics was 28 and the fathers, 30. These figures are similar to the control parents (mothers-29, fathers-32) and concur with the findings of Hall (1978) for anorexic parents. However, Garfinkel and Garner (1982) report they found the average age of the parents of the anorexics to be somewhat older (mother-30, fathers-33) and state that their findings were higher than the national average in Canada. Comparison of the parental weights revealed no statistically significant differences between the parents of the anorexics and the controls. This finding is in keeping with those of Halmi, Struss, and Golderg (1978) and would indicate, as pointed out by Halmi et al., that a genetic predisposition does not seem to exist between the anorexic and her parents regarding thinness or obesity. As investigators have searched for a cause of anorexia, it has been suggested that the size of the family and the birth order of the anorexic may be predisposing factors. Bruch (1973) stated the anorexic came from small families while Hall (1978) reported no significant difference in family size. In the present study 47% of anorexics came from families with one or two children while 44% of the control families also had the same number of children. Findings regarding birth order are varied and are closely linked to the size of the family. While Kay et al. (1967) found only children made up 50% of their subjects, the findings in the present study show 14% were only children which is 112 closer to the findings of Bruch (1973), Crisp et al. (1980), and Theander (1970). The number of firstborn (17%) is considerably less than the previous findings (Crisp et al., 1980; Hall, 1978; Norris, 1979; Theander, 1970), while the number of youngest (40%) is comparable to the other researchers. Hall (1978) points out that 5% of her subjects were middle children. The proportion of middle children (17%) in this study is more than three times the finding of Hall. The findings of this study indicate there is a tendency for the youngest child to develop anorexia nervosa. Parental age and the tendency for the youngest child to develop the condition may be connected in that parents are older when the youngest child is born. The tendency for the youngest child to develop anorexia may be more important than the age of the parents as connections could then be made between anorexia and the empty nest syndrome. Norris (1979) found girls outnumbered boys 3:1 in the families of anorexics. While there are more than twice the number of girls than boys in the anorexic families in this study (69% vs 31%) a similar number was found in the control families (70% vs 30%). Closely related to this is the proportion of families with all female children. Of the anorexics, 40% had no male siblings while 32% of the control subjects came from all female families. While no apparent differences exist between the anorexic and control families regarding the number of male (31% vs 30%) and female (69% vs 70%) children, it would appear the control group, although selected only for age, sex, and socioeconomic status, could be biased toward a preponderance of females as the male/female ratio is basically 1:1 in the general population. The rate of separation and divorce are reported to be low in the anorexic family, and this is taken as an indication of family stability (Bruch, 1973; Dally, 1969; Hall, 1978; Norris, 1979). In the present study the proportion of parents separated was 13% as was the proportion of divorced, now remarried parents. This figure is well below the rate of divorce at the national level which was reported to be 40% in 1976 (McKie, Prentice & Reed, 1983). A measure of geographic stability (duration of residence) was also investigated as it is assumed that families who move more frequendy will have smaller social networks. 113 While the number of years in the present geographical location was less for the anorexic subjects (8 yrs. vs 11 yrs.) no statistical difference was found. Findings regarding stress and illness events in families are inconsistent and little empirical research is reported. Difficulties arise in making comparisons with those of Dally (1969) and Halmi (1974) as the knowledge about which events were included was lacking. When the families of the anorexic and control subjects in this study were compared, somewhat fewer (82% vs 93%) of the anorexic families reported events which disrupt family balance and more events which require new environmental demands (40% vs 27%). The control group reported no events leading to the loss of self-esteem. The finding that more families of anorexics reported events requiring adjustment to new environmental demands could be related to the theoretical position espoused by Schwartz et al. (1984). They state that the onset of bulimia may be related to the inability of the family to adjust to a new environmental situation when they became geographically detatched from their kin-network. As bulimia is closely related to anorexia, the anorexic and her family may encounter similar types of difficulties when a new geographic location separates them from their former stable kin-relationships. The fact that no events were reported by the control group which they perceived as threatening to the self-worth lend credence to Garfinkel and Garner's (1982) view that many anorexics are sensitive "to external events for the regulation and maintenance of their sense of self-worth." (p. 205) In summary, the findings of the present study suggest that there are no demonstrable differences between the demographic variable for the families of the anorexics and the families of the controls. The one clear feature, the preponderance of upper-middle and upper social class families, which is reported consistently in the literature, is also supported in this study. There is also an indication that the youngest member of the family tends to be the one to develop anorexia more often than the oldest or middle child. 114 Discussion of the Results for the Four Systems Individual System. The Individual system was the first context of the integrated model to be investigated. This system considered the intrapsychic qualities of the anorexic female which reflect the interpersonal dimensions of the individuals' behaviour in interaction with others in the environment (Gough, 1975; Rodgers, 1966; Benjamin, 1977, 1979a). The majority of the statistically significant findings were in the more clinical aspects of the personality. These results are in keeping with other findings where anorexics have been compared with normal controls (Ben-Tovim et al., 1979; Garfinkel et a l , 1983; Gomez & Dally, 1980; Norman & Herzog, 1983; Small et al., 1981; Smart et al., 1976; Stonehouse & Crisp, 1977; Theander, 1970). The measures used in this study did not investigate neurotic, obsessive-compulsive or introverted-extroverted traits. However, some similarities are evident between the findings of other researchers and the results of this study. While the CPI-Clinical scales are not interchangable with the M M P I scales some similarities exist When the scores of the CPI-Clinical scales are converted to standard T-scores, (Table C - l , Appendix C) the elevated mean score on the Schizophrenia scale (T-score=72.9) for the anorexics supports the findings of Small et al. (1981). Other mean scores elevated between a T-score of 65 and 70 (Psychopathic Deviancy = 70, Psychasthenia = 69, Depression=66) also support the findings of Small et al. (1981) and Norman and Herzog (1983). Statistically significant differences on these scales were also found in this study between the anorexics and control subjects. Increased depression among anorexics is reported by Dally (1969), Stonehouse and Crisp (1977) and Theander (1970). A further analysis of the Psychopathic Deviancy scale, according to the Harris and Lingoes subscales, was undertaken in an earlier study of the same subjects by Sheppy and Solyom (1984). The Harris and Lingoes subscales break down the Psychopathic Deviancy scale into Family Discord, Authority Problems, Social Imperturbability, Social Alienation, and Self-alienation. The analysis showed that the anorexics were significantiy more socially and self-alienated when compared to the controls. This alienation is in keeping with the 115 elevated finding on the Schizophrenia scale which suggests anorexics feel lonely and separate from others. Other statistically significant differences between the anorexic and control subjects were on the scores for the Sense of Well-being, Anxiety, and Affiliation scales. When the anorexics scores on the Sense of Well-being scale are compared to the findings of Strober (1980) for anorexics on the CPI, the anorexic's scores in this study were lower (24.9 vs 31.1) Further, the Sense of Well-being score for the anorexics in the present study was significantly lower (24.9 vs 35.4) when compared to the population mean (p<.0001; note Table A - 2 , Appendic A). An increased anxiety score for anorexics is a common finding (Garfinkel et al., 1983; Smart et al., 1976; Stonehouse & Crisp, 1977; Theander, 1970). Selvini (1978) found anorexics displayed more hostile behaviour, while Ben-Tovim et al. (1979) confirmed hostility in their research when they compared anorexics to normal controls. Hostility was one of the findings Pillay and Crisp (1977) identified for anorexics who had recovered their weight The results of this study show that anorexics were significantly more hostile toward themselves (low Affiliation) than were the control subjects. That is, they blamed themselves more, put themselves down, neglected themselves, and were a menance to themselves. In keeping with the theory of SASB (Benjamin, 1974, 1979b), such behaviours are defined as a negative self-concept. In summary, the findings of the Individual System suggest that the anorexics, when compared to the control subjects, significantly overemphasize their worries and problems. The other personality traits as measured by the CPI were nonsignificant The significant differences for the clinical characteristics indicate the anorexics were more depressed and had a greater tendency for rebelliouness and hostile resistance toward authority figures. Rather than acting directiy on their feelings, the anorexics tended to turn conflicts inward, felt guilty about their thoughts and feelings and became self-punitive. There was a tendency for the anorexic to use ritualistic thoughts and actions in an attempt to gain a sense of control and order in their lives. They most likely felt emotionally alienated, had little sense of who they were emotionally and what they could realistically expect from 116 themselves. Further, the significantly lower Affiliation score suggests that the anorexics had a negative attitiude toward themselves. That is, they blamed themselves more, feel guilty, ashamed, unworthy and inadequate to a greater extent than did the controls. The significant findings at the Individual Systems level indicated that the variables selected for this study were able to discriminate between the anorexic and control subjects with a hit rate of 87.5%. Perusal of the standardized-normalized discriminant function coefficients showed that Affiliation, Psychasthenia and Schizophrenia were the variables which contributed the most to the discriminant analysis. Parent System. The next environmental context to be investigated was the Parent System which consisted of the personality traits, the clinical characteristics and the self-concept qualities of the parents. The paucity of significant findings between the mothers and fathers of the anorexics and the mothers and fathers of the control subjects does not support other findings in the literature (Dally, 1969; Crisp et al., 1974; Cantwell et al., 1977; Winokur et a l , 1977). However, the results of this study concur with the writings of Theander (1970) who stated that the prevalence of emotional illness in the parents of anorexics did not exceed that of the general population. A statistically significant finding between the parents of the anorexics and controls is the elevated Psychopathic Deviancy scale for the mothers of the anorexics. However, the T-score equivalent for the mothers of the anorexics (T-score=66) remains below the pathological level (T-score >70) as defined by Graham (1980). Research by Sheppy and Solyom (1984) showed that when the Psychopathic Deviancy scores of the mothers of the anorexics and controls were analyzed according to the Harris and Lingoes subscales using a P test none of the five subscales were significantiy different However, the scores of the mothers of the anorexics were more than twice that of the mothers of the controls on the Social (1.71 vs .73) and Self-alienation (1.95 vs .83) subscales. The significant finding on the Psychopathic Deviancy scale suggests the mothers of the anorexics are more rebellious 117 and hostile toward authority figures, more nonconforming and socially and self-alienated. The findings for the other personality dimensions measured by the CPi, CPI-Clinical and SASB were nonsignificant Although non-significance was found in this study between the fathers of the anorexics and the control subjects on the Depression scales, a significant difference was found between the fathers of the anorexic subjects and the population means on this scale (T-score=68, p<.0001; note Table A - 6 Appendix A) which supports the findings by Dally (1969), Cantwell et al. (1977) and Winokur et al. (1977), that depression is increased in fathers of anorexics. The nonsignificant finding between the fathers of the anorexics and the fathers of the controls could possibly be accounted for by the elevated Depression score of the fathers of the controls (T-score = 63.3) and thus depression in the parents of the anorexics may be a factor of ilness in the family rather than anorexia per se. In summary, the results of this study show that the parents of the anorexics could be discriminated from the parents of the controls on the basis of the Psychopathic Deviancy score of the mothers. The mainly nonsignificant findings between the parents emphasize Yager's (1983) point that there is a great diversity in the personality of both the mothers and fathers of the anorexics, and if common patterns are to be found they will be at more subde levels of functioning. Family System. The next environmental context to be encountered by the anorexic is her immediate family. The present study considered two aspects of the family: 1) the parent-daughter relationship, and 2) the family environment For the parental transactions, the anorexics viewed their mothers as being significantly less friendly and more controlling in the relationship than did the control subjects. However, the mothers did not perceive themselves as being overprotective and controlling. The results regarding the fathers' behaviour toward their daughters are not statistically significant from either the daughters' or the fathers' view. This means that anorexics perceive their mothers as more overprotective 118 and controlling while there is no demonstrable difference between how the parents of the anorexics and controls rate themselves when initiating transactions with their daughters. The pattern of behaviour for the anorexics in response to their parents shows that the anorexics rate themselves as being significantiy less friendly to both their mothers and their fathers than do the control subjects. The mothers of the anorexics also rate their daughters as being significantiy less friendly in their responses while the fathers perceive the same type of behaviour but not to a significant degree. Although the parents of the anorexics perceive the relationship with their daughters as being complementary, the anorexics view it as a noncomplementary relationship. In keeping with the theory of the SASB model, complementary behaviours are found in a topologically similar position on the Other and Self planes (Benjamin, 1974, 1979b). The parents rated their own behaviours (Affiliation & Autonomy) and their daughters' responses (Affiliation & Autonomy), both as positive in nature and as such they fall into Quadrant I of the model (note Figure 2 and 3). Ratings of behaviours falling in the same quadrant are considered complementary. In comparison, the anorexics rated their parents' behaviours on the dimension of Affiliation with a positive score and for Autonomy with a negative score. Such ratings are found in Quadrant IV of the model. In turn, the anorexics rated themselves in response to their parents with positive Affiliation and Autonomy scores which places the self-ratings of their own responses in Quadrant I. While the anorexics perceived their parents as more controlling, they do not submit to this perceived control (note Figure 3) and as such they perceive the relationship as noncomplementary. These perceived differences in behaviour between the anorexics and their parents could account for misunderstandings which could possibly lead to conflict in the parent-daughter relationship such as has been observed by Bruch (1973, 1978), Halmi (1974) and Selvini (1978). The observed lack of development of individuation in the anorexic due to the lack of autonomy giving by the parents has been documented in the literature (Bruch, 1973, 1978; Masterson, 1977; Selvini, 1978; Sours, 1974), as well as the conflictual mother-daughter relationship (Bruch, 1973, 1978; Selvini, 1978; Sours, 1974). The present 119 findings shed some light on these reported clinical observations. Taipale et al. (1971) depict the mothers of anorexics as being unable to tolerate independence in their children and Sours (1974) emphasizes the mothers need to have a submissive, perfect child for her own fulfillment Selvini (1978) and Bruch (1973, 1978) also describe similar relationships where the mother assumes an overprotective role which hinders the development of individuality in the anorexic. The significant findings of the anorexics' ratings of their mothers indicates they perceive their mothers as a controlling force in their life which may prevent them from becoming an autonomous self. However, the fathers, unlike the description of being weak and passive (Selvini, 1978), are also perceived by the anorexics as tending to control (Quadrant IV behaviour) and may therefore further add to the lack of the development of individuation and autonomy. The behaviours of the fathers which the anorexics perceive as controlling could be covert in nature and as such would fit the description of the weak passive father. The fact that the parents do not perceive their behaviour the same way that the anorexics do may be accounted for by Bruch's (1977, 1981) observation that parents deny all difficulties in the family except the daughters' weight loss. Such speculation of parental reporting needs to be researched further using an objective type of measure. The significant finding of the lack of affiliation (hostility) in the anorexics' responses to their parents as perceived by the anorexics and their mothers may reflect the hostile attitude the anorexics have toward themselves, as reported earlier in this paper. While anorexics are seen as hostile (Ben-Tovim et al., 1979; Pillay & Crisp, 1977; Selvini, 1978) they are rarely described as such in their relationship with their parents. Rakoff (1983) mentions that the anorexic is rebellious in her relationship with her parents but does not elaborate further. Garfinkel and Garner (1982) indicate hostility is evident in the families when they first present for treatment but they point out this is not necessarily an indication of its existence in the premorbid relationship. Fromm (1981) used the SASB model as an objective measure for coding parental responses to structured vignettes. Her results showed that parents perceived their daughters as responding with hostile submission. 120 It is in keeping with the theory of the SASB model (Benjamin, 1979a) that when the anorexic begins to exert more autonomy, the parents respond with hostile power to gain control. The response of the daughter is one of hostile submission with a resulting introject of oppression of self The anorexics' perception in the present study indicated they viewed their mothers as less friendly and more controlling and thus responded to them with less friendly behaviour. The anorexic has introjected this perceived hostile power from the mother to oppress the self (note anorexics Affiliation and Autonomy scores, Table 11). The second aspect of the family system, the family environment, showed an overall statistically significant difference between the families of the anorexics and the controls on the five dimensions investigated (Cohesion, Conflict, Independence, Organization and Control). While this finding indicated the family of the anorexic tended toward greater dysfunction, Cohesion was the only dimension which was significant on the simultaneous multiple comparisons. This finding suggests that the families of the anorexics are significantiy less committed and provide less support and help to each other. Research findings reported in the literature regarding the environments within the familiy of the anorexic (Bruch, 1973, 1978; Norris & Jones, 1979; Slevini, 1978; Wilson, 1980) are mainly observational. However, the findings of Strober (1981) for abstaining anorexics on the FES are similar to those of the anorexics in the present study. That is, when compared to the controls, the anorexics in Strober's study had lower Cohesion and Independence and higher Conflict and Control mean scores. The same directions were found for these same scales for the anorexics in the present study. The concept of enmeshment is frequentiy used to describe anorexics and their parents. This is referred to as a symbiotic bonding or pathological closeness between one or both of the parents and the anorexic. While enmeshment is not measured in this study, some of the qualities may be reflected in the trend toward less assertiveness and self-sufficiency as measured by the Independence scale. The ability to deal with conflict and aggression and thus the denial of such in families of the anorexic (Bruch, 1979; Minuchin et al., 1975; Norris, 1979; Norris & Jones, 1979) is not reflected in the findings of this study. 121 The question remains, "Are the family environment findings specific to the families of anorexics or are similar characteristics found in other distressed families?" Moos and Moos (1981) report findings which indicate other dysfunctional families, when compared to normal families, are rated as having less cohesion, organization and independence and more conflict This could be an indication that the families of the anorexics are similar to other distressed families as the families of anorexics have a significantly lower score on Cohesion and a tendency toward less Independence and more Conflict Further research is needed in order to compare anorexics and their families with families containing another type of dysfunctional problem. The significant findings at the Family systems level indicate that the variables selected for this study, at this level, were able to discriminate between the families of the anorexics and the families of the controls with a hit rate of 78%. Perusal of the standardized-normalized discriminant function coefficients indicate that the perceived degree of affiliation the daughters have in their relationship with their mothers is the variable which has the greatest discriminant power at the Family System level. Community System. The final system to be investigated in this study was the community. No significant differences were found in the study in the Size and Quality of the social network of the anorexics and their families when compared to to those of the control families. Rationale for the nonsignificant findings could be that no real differences exist between the social network of the anorexic and the control subjects and their families. However, the literature does not seem to support this position. Crisp et al. (1980), Bruch (1978), Selvini. (1978), and others describe the anorexic as isolating and withdrawing from friends, especially avoiding heterosexual contacts. These families are also noted for the importance they place on the perfect family image presented to the community (Bruch, 1973, 1978). It would appear that the withdrawal of a family member from a social network, along with the need to maintain a family facade could cause a binding together 122 of family members resulting in a smaller social network. More frequent geographic moves are also reported to be a precipitating event in the condition (Dally, 1969; Garfinkel & Garner, 1982). With the family moving more often, the size and quality of the social network would be decreased. Network loyalty is seen by Schwartz et al. (1985) to be paramount in families of bulimics and when the family leaves to become "a part of bountiful, mobile, middle America" (p. 11) problems arise unless they readily develop new ways of relating to each other and the outside environment Schwartz et al. (1985) hypothesize that the isolation and dependency on only family members for support becomes a precipitating factor in bulimia. It is possible such a phenomenon could also exist in the families of anorexics. Pattison et al. (1979) found the average size of the social network for normal subjects to be between 22 and 25 persons. The size of the social network (15 persons) for the control subjects in this study is closer to the neurotic-type network described by Pattison et al. (1979). Had the social network of the control families been closer to the average, a significant difference may have been found between them and the families of the anorexics. The smaller social network for the control families could reflect the referral source for these subjects which might suggest a small social network is characteristic of families with an ill member. A further possible reason for the lack of significant findings at this level is the type of instrumentation and/or its administration. Perhaps the instrument lacked the necessary sensitivity to detect the differences between the anorexic and control groups, it did not measure the existing differences, the administration of the measure was inappropriate, or the instrument itself does not detect the social desirability factor commonly found in the families of the anorexics (Bruch, 1973, 1978; Crisp, 1980; Garfinkel & Garner, 1982). Other researchers have found that in order to adequately explore the dimensions of the social network, a structured interview of 2 to 3 hours is required (Tolsdorf, 1976; Speck & Attneave, 1973). Such a time committment on the part of the respondents was not possible in this study. However, as the data presented in this study are a global measure 123 of the characteristics of the social network, a detailed analysis could be most revealing. Discussion of the Results for the Ecological-System Model The hypothesis, when all four systems are combined into an Ecological-Systems model and statistically analysed, the four systems together will differentiate to a greater degree than will each of the systems independently, was not supported by the statistical analysis. Rather, it was determined by the stepwise discriminant analysis that the daughters' Affiliation mean score (Individual System) and the mothers' Psychopathic Deviancy mean score (Parent System) were the variables having the greatest discriminant power. With these two variables, it was possible to correctly classify 87.5% of the subjects into either the anorexic or control group. When selected variables from the family and community systems were forced into the equation, the discriminant function was not increased. Statistically, the results at this level of the study, indicated that anorexia nervosa is a phenomenon related to the. negative self-concept (low Affiliation) of the anorexics and the psychopathic deviancy of their mothers. Further, the findings of the discriminant analysis at the separate system levels indicated that 87.5% of the subjects could be correctiy classified with the variables from the Individual System alone. The results of the step-wise discriminant analysis investigating the four systems together showed a hit rate of 87.5%. Thus the power of the Ecological-Systems model to discriminate between the anorexic and control subjects was not increased over that of the Individual System. These results present a strong argument that the personality traits, clinical characteristics and self-concept qualities of the anorexics themselves are the most powerful discriminators. Garfinkel et al. (1983) write in partial support of this finding when they state that the "illness could develop without significant demonstrable familial psychopathology" (p. 827). Henderson (1982) further supports this finding from his investigation with neurotic patients. While not negating the interactive effects of the environment he states that "attributes of the individual which are biological or constitutional are likely to be more powerful than properties of the immediate social 124 environment or other ecological variables" (p. 228). The application of a systemic multidimensional framework, as presented in this study, adds to the understanding of anorexia nervosa by identifying relationships between the levels of systems functioning. The variables which were the best discriminators came from two systems, the Individual and the Parent, which indicates information from both these systems is required in order to maximally discriminate between the anorexic and control subjects. Whether or not a relationship exists between the Affiliation score of the anorexics and the Psychopathic Deviancy score of the mothers of the anorexics is not defined. However, the literature suggests the mothers of the anorexics play a vital role in the etiology and maintanence of the condition (Bruch, 1973, 1978; Masterson, 1977; Selvini, 1978; Sours, 1974). While variables from the Family system did not increase the discriminating power of the Ecological-Systems model, analyses at the Family systems level indicated a significant lack of affiliation was perceived by both the anorexics and their mothers in their relationship. As the SASB instrument was designed to measure interpersonal qualities (Benjamin, 1974), the Affiliation score for the anorexics could reflect not only a negative self-concept but the effects of that negative self-concept in interpersonal transactions. That is, the negative self-concept of the anorexics (Individual system) could have considerable influence on their relationship with their mothers and their fathers (Family system). Also, in keeping with the theory of the SASB (Benjamin, 1974), the negative self-concept of the anorexics could be considered partially a result of the anorexics' perceptions of their parents interactions with them. That is, the overprotective and controlling behaviours of the mothers were introjected by the anorexics. Thus, a feedback loop is established whereby the anorexics negative feelings about themselves affect their behaviours in their transactions with their parents, who in turn interact to their daughters in an overprotective and controlling manner. Such behaviours on the part of the parents reinforce the inadequate and hostile feelings of the anorexics. If such a feedback loop exists between the anorexics and their parents, a link is established between the Individual system and the 125 parent-daughter transaction of the Family system in the Ecological-Systems model presented in this study. The results of this study emphasize what might be called a ripple effect. The anorexic, like a pebble dropped into a pool of water, is at the center of her environment Here, at the center, the ripples are the most intense and are therefore readily detected. As the distance from the center increases, a decrease in the force of the ripple is experienced. This analogy can be applied to the investigation of anorexia nervosa from an ecological perspective. The further the investigator moves from the anorexic, the more sensitive the measures need to be in order to pick up the existing differences. With the development and refinement of family and social network assessment tools, it may be possible to detect the more subtle differences within these areas should they exist Limitations and Directions for Future Research The limitations and directions for future research are discussed together as several of the suggestions for further research arise from the limitations of the study. The limitations as they relate to the subjects will be discussed first then the strengths and limitations of the Ecological-Systems model. The anorexic subjects were females living in Western Canada between the ages of 15 and 23 who were presently under the care of a physician. No native Indian or negroid subjects were available to participate in the study, however one anorexic was of oriental decent Therefore the results found in this study are applicable only to other anorexics who meet these criteria. To obtain greater generalizability, the study would have required anorexic subjects from other ethinic groups, age ranges and sex. Due to the difficulty in contacting anorexic subjects, they were drawn mainly from the medical practice of one psychiatrist This may have introduced unknown biases into the research sample which might not have occurred had the sample been obtained from- a number of sources. Further, the duration of illness in the subjects ranged from 4 months to 5 years. It is 126 believed (Crisp, 1980; Garfinkel and Garner, 1982) that the chronic anorexic may have different characteristics and family relationships than those who have just developed the condition. Research which compared the chronic anorexic to those with intitial onset may provide different results and thus shed light on factors which could lead to the progression of anorexia into a chronic state. Recent research comparing anorexic abstainers and anorexic bulimics indicated that differences between these two groups may exist (Garfinkel & Gamer, 1982). While no statistically significant differences were found on the CPI, CPI-Clinical and SASB measures between the anorexic abstainers and bulimcs in the present study, differences may be found with a larger sample size. Other limitations in the study relating to the subjects are the sample size and the comparison of the anorexics to only one control group. Although the statistical procedures are rigorous, differences between a number of variables indicate trends rather than significance. A larger sample size could have resulted in more significant results as the differences between groups would not have had to be as large to obtain statistical significance. Studies comparing anorexics and their family members with normal control subjects and their families are beginning to appear in the literature (Garfinkel et al. 1983). While the present study compared anorexics with a group of control subjects who attended a medical practitioner for a physical illness, the signs and symptoms of anorexia may become more clearly defined if comparative research involving more than one condition (i.e., neurotics, obsessive-compulsives, long-term illnesses) along with a normal control group was undertaken. Finally, the problem of social desirability needs to be taken into consideration as denial, and the concern with external appearances are commonly reported characteristics of the anorexics and their parents (Bruch, 1973, 1978; Crisp et al., 1980; Garfinkel et al., 1983; Minuchin et al., 1975; Norris & Jones, 1979; Selvini, 1978; Yager, 1982). Self-report questionnaires, as used in this study, are subject to falsification and more objective methods for collecting data (i.e., a structured interview) may have avoided misrepresentation of data. The second area for discussion is the strengths and limitations of the Ecological-Systems model. Research in the area of anorexia nervosa has been framed 127 within unidimensional and multidimensional approaches. The present investigation, however, systematizes theoretical aspects and former research results into a single Ecological-Systems model which shows promise and as such may provide a conceptual framework for future research. If a goal of anorexia nervosa research is to determine the specific combination of predisposing factors which produce the anorexic syndrome, further research using a methodology which systematizes these factors and which can be statistically analysed is needed. The present study points in the direction of such a methodology. Four independent systems (Individual, Parent Family Community) were investigated in the present study. Bronfenbrenner (1979) included a macrosystem in his model which accounts for the cultural beliefs, values, and mores of a given society. Variables from Bronfenbrenners' macrosystem were not included in this study. Although Garfinkel and Garner (1982) as well as Hall and Brown (1983) are investigating these factors extensively, no research was found which has attempted to combine these variables into a systems perspective. As the mesosystem plays such an important role for the anorexic, there is a definite need to look at how this system interacts with data from the Individual, Parent Family and Community systems. Although the variables in this study were selected to reflect interactive qualities within the ecological environment the number and character of the factors to be investigated within each system needs to be increased. For example, at the individual level, factors examining cognitive and perceptual processes could be added in accordance with the theory espoused by Bruch (1978, 1981). At the family level, parent-daughter interactions could be expanded to consider the present interactional components of this study in greater detail as well as investigate nonverbal communication patterns. The community system needs further researching especially into the breakdown of the size and quality according to the various component groups (i.e., family, relatives, friends etc.) within the social network. Further analyses of other variables, such as stress factors could be explored. In addition to increasing the number of variables, more sensitive instruments are needed, such as assessment tools in the family area which give reliable and valid data on 128 parent-daughter interactions. The use of audio-visual recordings with coding show promise for such analyses but are costly and time-consuming. A more comprehensive method for the assessment of the social network is needed which examines in depth factors such as the interaction between network members, the level of satisfaction, the supportive and nonsupportive qualities of the social network. Finally, while anorexia nervosa is not a new syndrome, it is reported to be on the increase (Bemis, 1978; Crisp, 1980; Richardson, 1980; Sours, 1969) with many factors regarding the condition still remaining undiscovered. The purpose of this study was not to unearth all the unknowns of anorexia. Rather, the study was designed to apply aspects of the ecological systems approach to gain a better understanding of the relationship of anorexia nervosa to the individual, the parent the family and the community systems. A further goal was to determine if a greater proportion of subjects could be correctly classified into the anorexic and control groups according to the discriminant function by combining variables from the various levels of system functioning. 129 REFERENCE NOTES Note 1: Statistics Canada. Hospital Separations for the five year period 1973 to 1977 for Anorexia Nervosa, ICDA code 306.5. Statistics Canada, Population Catalogue 91-512, occasional and the 1976 Census of Canada. Note 2: Hall, A., Leibrich, J., & Walkey, F. Preliminary results of an investigation of 204 families of schoolgirls. Unpublished manuscript. 130 REFERENCES American Psychiatric Association. (1980). Diagnostic and Statistical Manual of Mental Disorders. (3rd ed.). Washington, D C : Author. Azerrad, J. & Stafford, R. L. (1975). Restoration of eating behavior in anorexia nervosa through operant conditioning and environmental manipulation. In R. C. Katz & S. Zlutnich (eds.), Behavior therapy and health care: Principles and application. New York: Pergamon Press. Belsky, J. (1978). A. Theoretical analysis of child abuse remidiation stratagies. Journal of Clinical Child Psychology, 17-121. Bemis K. M . (1978). Current approaches to the aetiology and treatment of anorexia nervosa. Phychological Bulletin. 85, (3), 593-617. Benjamin, L. S. (1974). Structural analysis of social behavior. Psychological Review, 81, 392-425 Benjamin, L. S. (1977). Structural analysis of a family in therapy. Journal of Consulting and Clinical Psychology, 45 (3), 391-406. Benjamin, L.S. (1979a). Structural analysis of differentiation failure. Psychiatry, 42(1), 1-23. Benjamin, L. S. (1979b). A manual for using SASB questionnaires to measure correspondence among family history, self-concept and current relations with significant others. Unpublished Manuscript University of Wisconsin, Department of Psychiatry, Wisconsin Psychiatric Institute, Madison. Benjamin, L. S. (1982). Validation of structural analysis of social behavior (SASB). Unpublished manuscript University of Wisconsin, Department of Psychiatry and the Wisconsin Psychiatric Institute, Madison. Bennett W. B., & Gurin, J. (1982). The dieter's dilemma: Eating less and weighing more. New York: Basic Books. Ben-Tovim, D. I., Marilov, V., & Crisp, A. N . (1979). Personality and mental state (PSE) within anorexia nervosa. Journal of Psychosomatic Research, 23(5), 321-325. Blishen, B. (1967). A socioeconomic index for occupations in Canada. Canadian Review of Sociology and Anthropology, 4, 41-53. Blishen, R. B. & McRoberts, H . A. (1976). A revised socioeconomic index for occupations in Canada. Canadian Review of Sociology and Anthropology, 13, 71-79. Brady, J. P. & Rieger, W. (1975). Behavioral treatment of anorexia nervosa. In T. Thompson & W. S. Dodiens (eds.), Application of behavior modification. New york: Academic Press. Bronfenbrenner, U . (1977). Toward an experimental ecology of human development American Psychologist, 32, 513-531. 131 Bronfenbrenner, U . (1979). The ecology of human development, Cambridge, Mass: Harvard University Press. Bruch, H . (1966). Anorexia nervosa and its differential diagnosis. Journal of Nervous and Mental Disease, 141, 556. Bruch, H . (1970). Family background in eating disorders. In E. J. Anthony & C. Keupeni (eds.), The child in his family, (pp. 285-309). New York: John Wiley & Sons, Inc. Bruch, H . (1973). Eating disorders: Obesity, anorexia nervosa, and the person within. New York: Basic Books, Inc. Bruch, H . (1977). Psychological antecedents of anorexia nervosa. In R. A. Vigersky (ed.), Anorexia nervosa. New York: Raven Press. Bruch, H . (1978). The golden cage: The enegma of anorexia nervosa. Cambridge, Mass., Harvard University Press. Bruch, H . (1981). Developmental consideration of anorexia nervosa and obesity. Canadian Journal of Psychiatry, 26(4), 212-217. Build Study, 1979. (1980). Society of Actuaries and Association of Life Insurance Medical Directors of America. Cantwell, P. D., Sturenberger, S., Burroughs, J., Salkin, B. & Green, J. K. (1977). Anorexia nervosa: An affective disorder. Archives of General Psychiatry, 34, 1087-1093. Chiles, J. A., Stauss, F. S., & Benjamin, L. S. (1980). Marital conflict and sexual adjustment in alcoholic and non-alcoholic couples. British Journal of Psychiatry, 137, 266-273. Conger, R. D. (1981). The assessment of dysfunctional family systems. In B. B. Lakey & A. E. Kazdin (eds.), Advances in clinical child psychology (vol.4). (pp. 199-242). New York: Plenum Press. Conger, R. D., & Burgess, R. L. (1980). An ecological analysis of parental behavior. Unpublished manuscript, University of Georgia. Crisp, A. H . (1965). Clinical and therapeutic aspects of anorexia nervosa - study of 30 cases. Journal of Psychosomatic Research, 9, 67-78. Crisp, A. H . (1967). Anorexia nervosa. Hospital Medicine, 5, 713-718. Crisp, A. H . Anorexia nervosa: Let me be. (1980) New York: Grune & Stratton. Crisp, A. H . (1981). Therapeutic outcome in anorexia nervosa. Canadian Journal of Psychiatry, 26(4), 232-235. Crisp, A. H . , Harding, B. & McGuinness, B., (1974). Anorexia nervosa: Psychoneurotic characteristics of parents: Relationship to prognosis. Journal of Psychosomatic Research, 18, 167-173. 132 Crisp, A. H . , Hsu, L. K. G. , Harding, B., & Hartshorn, J., (1980) Clinical features of anorexia nervosa. Journal of Psychosomatic Research, 24(3-4), 179-191. Dally, P. J. (1969). Anorexia nervosa. New York: Grune & Stratton. Dixon, W. J. (1981) BMDP satistical software. LosAngeles: University of California Press. Feigher, J. P., Robins, E., & Guze, S. B. (1972). Diagnostic criteria for use in psychiatric research. Archives of General Psychiatry, 26, 57-63. Fromm, L. M . (1981). Parental responses to anorexia nervosa: Adolescents' attempts at individuation. Dissertation Abstracts International, 42(3), 1007A. Garfinkel, P. E. (1981). Some recent observations on the pathogenesis of anorexia nervosa. Canadian Journal of Psychiatry, 26(4), 218-223. Garfinkel, P. E., Kline, S. A., & Stancer, H . C. (1975). Treatment of anorexia nervosa using operant conditioning techniques. In R. C. Katz & S. Zlutnich (eds.), Behavior therapy and health care: Principles and application. New York: Pergamon Press. Garfinkel, P. E., & Gamer, D. M . (1982). Anorexia nervosa: A multidimensional approach. New York: Brunner/Mazel. Garfinkel, P. E., Gamer, D. M . , Rose, J., Darby, P. L., Brandes, J. S., O'Hanlon, J., & Walsh, N . (1983). A comparison of characteristics in the families of parents with anorexia nervosa and normal controls. Psychological Medicine, 13(40), 821-828. Gamer, D. M . , & Garfinkel, P. E. (1978). Sociocultural factors in anorexia nervosa. The Lancet, 2, 674. Garner, D. M„ Garfinkel, P. E., Schwartz, D., & Thompson, M . (1980). Cultural expectations of thinness in women, Psychological Report, 47, 482-491. Gamer, D. M . , Garfinkel, P. E., & Olmsted, M . P. (1983). An overview of sociocultural factors in the development of anorexia nervosa. In P. L. Darby, P. E. Garfinkel, D. M . Gamer, & D. V. Coscina. (eds.), Anorexia nervosa: Recent developments in research, (pp. 65-82). New York: Alan R. Liss, Inc. Giannini, A. J. (1981-82). Anorexia nervosa: A retrospective view. International Journal of Psychiatry in Medicine, 11(3), 199-202. Gomez, J., & Dally, P. (1980). Psychometric rating in the assessment progress in anorexia nervosa. British Journal of Psychiatry, 136, 290-296. Gough, H . G . (1968). An interpreter's syllabus for the California Psychological Inventory. In P. McReynolds (ed.), Advances in psychological assessment (Vol. 1). (pp. 53-79). Palo Alto: Science & Behavior Books, Inc. Gough, H . G . (1975). California Psychological Inventory manual, Palo Alto: Consulting Psychologists Press, Inc. Graham, J. R. (1977). The MMPI: A practical guide, New York: Oxford University Press. 133 Gull , W. W. (1874). Anorexia nervosa (Apepsia hysterica, anorexia hysterica). Transactions of the Clinical Society of London, 7, 22-29. Hakstian, A. R., Roed, J, C , & Lind, J.C. (1979). Two-sample P procedure and the assumption of homogeneous covariance matrices. Psychological Bulletin, 86(6), 1255-1263. Hall, A. (1978). Family structure and relationships of 50 female anorexic nervosa patients. Australian and New Zealand Journal of Psychiatry, 12(A), 263-269. Hall, A., & Brown, L. B. (1983). A comparison of the attitudes of young anorexic nervosa patients and non-patients with their mothers. British Journal of Medical Psychology, 56, 39-48. Halmi, K. A. (1974). Anorexia nervosa: Demographic and clinical features in 94 cases. Psychosomatic Medicine, 36(1), 18-26. Halmi, K. (1978). Anorexia nervosa: Recent investigations. Annual Review of Medicine, 29, 137-148. Halmi, K. A. (1983). The state of research in anorexia nervosa and bulimia, Psychiatric Developments, 3, 247-262. Halmi, K. A., Powers, P., & Cunningham, S. (1975). Treatment of anorexia nervosa with behavior modification. Archives of General Psychiatry, 32, 93-95. Halmi, K. A., Struss, A., & Goldberg, S. C. (1978). An investigation of weights in the parents of anorexic nervosa patients. Journal of Nervous and Mental Disease, 166, 358-361. Hamill, P. V. V., Drizd, T. A., Johnson, C. L. Reed, R. B. Roche, A. F., & Moore, W. M . (1979). Physical growth: National Center for Health Statistics percentiles, American Journal of Clinical Nutrition, 32, 607-629. Henderson, S. (1982). The significance of social relationships in the etiology of neurosis. In C. M . Parkes & J. Stevenson-Hinde (eds.), The place of attachment in human behavior, London, Tavistock Pub. Hurd, G. , Llamas, R., & Pattison, E. M . (1980). The structure and Junction of normal social networks. Unpublished manuscript Medical College of Georgia. Hurd, G. , Pattison, E. M . , & Smith, J. E. (1981). Tests, re-test reliability of social networks self reports: The Pattison Psychosocial Inventory (PPI). Paper presented to the Sun Belt Social Network Conference, Tampa, Florida. Kalucy, R. S., Crisp, A. H . , & Harding, B. (1977). A study of 56 families with anorexia nervosa. British Journal of Medical Psychology, 50, 381-395. Kay, D. W. K., & Leigh, D. (1954). The natural history, treatment and prognosis of anorexia nervosa based on a study of 38 patients. Journal of Mental Science, 100, All. Kay, D. W. K., Schapira, K., & Brandin, S. (1967). Early factors in anorexia nervosa compared with non-anorexic groups. Journal of Psychosomatic Research, 11, 133-139. 134 Kendell, R. E , Hall, D. J., Hailey, A., & Babigian, H . M . (1973). The epidemiology of anorexia nervosa. Psychological Medicine, 3, 200-203. King, A. (1963). Primary and secondary anorexia nervosa syndromes. British Journal of Psychiatry, 109, 470. Kirstein, L. (1982-82). Diagnostic issues in primary anorexia nervosa. International Journal of Psychiatry in Medicine, 77(3), 235-243. Kramer, S. (1974). A discussion of the paper by John A. Sours on "The anorexic nervosa syndrome." International Journal of Psycho-Analysis, 55, 557-559. Laseque, C. (1873). On hysterical anorexia. Translated from Archives Generates de Medicine. In M . R. Kaufman & M . Heiman (eds.), Evolution of psychosomatic concepts: Anorexia nervosa. (pp. 141-155). New York: International Universities Press, 1964. Leary, T. (1957). Interpersonal diagnosis of personality. New York: Ronald Press. Lester, E. P. (1981). Anorexia nervosa and obesity - Recent developments. Canadian Journal of Psychiatry, 26(4), 211. Leventhal, A. M . (1966). An anxiety scale for the CPI. Journal of Clinical Psychology, 22, 459-461. Leventhal, A. M . (1968). Additional technical data on the CPI Anxiety Scale, Journal of Counseling Psychology, 15(5), 479-480. McKie, D. C , Prentice, B., & Reed, P. (1983). Divorce: Law and the family in Canada. Canada. Statistics Canada, Catalogue 89-502E: Ministry of Supply and Service. Masterson, J. F. (1977). Primary anorexia nervosa in the borderline adolescent: An object relations view. In P. Hartocollis (ed.). Borderline personality disorders, (pp. 475-494). New York: International Universities Press. Megargee, E. I. (1972). The California Psychological Inventory handbook, London: Jossey-Bass. Miller, P. M . , & Ingram, J. G . (1976). Friends, confidents and symptoms. Social Psychiatry, 11, 51-58. Minuchin, S. (1970). The use of an ecological framework in the treatment of a child. In E. J. Anthony & C. Koupemik (eds.), The child in his family, (pp. 41-58). New York: John Wiley & Sons, Inc. Minuchin, S. Baker, L., Rosman, B.L., Libman, R., Milman, L., & Todd, T.C. (1975). A conceptual model of psychosomatic illness in children. Archives of General Psychiatry, 32, 1031-1038. Minuchin, S., Rosman, B., & Baker, L. (1978). Psychosomatic families: Anorexia nervosa in context. London: Harvard University Press. 135 Moos, R. H . (1974). Family environment scales manual. Palo Alto: Consulting Psychologists Press Inc. Moos, R. H . , & Moos, B. S. (1981). Family Environment Scale manual. Palo Alto: Consulting Psychologists Press. Nelson, R. O., & Hayes, S. C. (1979). The nature of behavioral assessment: A commentary. Journal of Applied Behavioral Analysis, 12, 491-500. Nemiah, J. C. (1950). Anorexia nervosa - a clinical psychiatric study. Medicine, 29, 225. Nie, N . H . , Hull, C. H. , Jenkins, J. G. , Steinbenner, K., & Bent, D. H . (1975) Statistical Package for the Social Sciences (2nd ed.). New York: McGraw-Hil l . Norman, D. K., & Herzog, D. B. (1983), Bulimia, anorexia nervosa, and anorexia nervosa with bulimia. International Journal of Eating Disorders, 2(2), 43-52. Norris, D. L. (1979). Clinical diagnostic criteria for primary anorexia nervosa. South African Medical Journal, 56(23), 987-993. Norris, D. L., & Jones, E. (1979). Anorexia nervosa: Clinical study of 10 patients and their family systems. Journal of Adolescence, 2(2), 101-113. Papatola, K. J. (1982). The effects of ontogenic, microsystem and mesosystem variables on the outcome of child abuse. Unpublished doctoral dissertation, University of British Columbia, Vancouver, B.C. Pattison, E. M . , Llamas, R., & Hurd, G . (1979). Social network mediation of anxiety. Psychiatric Annals, 9(9), 56-67. Pillay, M . , & Crisp, A. H . (1977). Some psychological charateristics of patients with anorexia nervosa whose weight has been newly restored. British Journal of Medical Psychology, 50, 375-380. Pinsof, W. M . (1981). Family therapy process research. In A. S. Gurman & D. P. Kniskern (eds.), Handbood of family therapy New York: Brunner/Mazel, Inc. Powell, D. R. (1979). Family-environmental networks and neighborhoods. Journal of Research and Development in Education, 13, 1-11. Rakoff, V. (1983). Multiple determinants of family dynamics in anorexia nervosa. In P. L. Darby, P. E. Garfinkel, D. M . Garner, & D.V. Coscins (eds.), Anorexia nervosa: Recent developments in research, (pp. 29-40). New York: Alan R. Liss, Inc. Rampling, D. (1980). Abnormal mothers in anorexia nervosa. Journal of Nervous and Mental Disease, 765(8), 501-504. Richardson, T. F. (1980). Anorexia nervosa: An overview. American Journal of Nursing, 8, 1470-1471. 136 Rodgers, D. A. (1966). Estimation of M M P I profiles from CPI data. Journal of Consulting Psychology, 30, 89. Rosman, B. L., Minuchin, S., Liebman, R., & Baker, L. (1977). Input and outcome of family therapy in anorexia nervosa. Adolescent Psychiatry (vol.5). (pp. 313-322). New York: Janson Aronson, Russell, G . F. M . (1970). Anorexia nervosa: Its identity as an illness and its treatment. In J. H . Price (ed.), Modern trends in psychological medicine, 2. London: Butterworths. Schaffer, E. S. (1965). A configurational analysis of children's reports of parent behavior. Journal of Consulting Psychology, 29, 552-557. Schwartz, R. C , Barrett, M . J., & Saba, G . (1985). Family therapy for bulimia. In D. M . Gamer & P. E. Garfinkel (eds.), Handbook of psychotherapy for anorexia nervosa and bulimia, (pp. 281-307). New York: Guilford Press. Selvini, M . (1971). Anorexia nervosa. In S. Arieti (ed.), The world biennial of psychiatry and psychotherapy. (Vol.1), (pp. 197-218). New York: Basic Books. Selvini, M . (1978). Self-starvation: From individual to family therapy in the treatment of anorexia nervosa. New York: Janson Aronson. Sheppy, M . I., & Solyom, L. (1984). Personality characteristics of anorexics and their parents compared to normal controls and their parents. Paper Presented at the First International Conference of Eating Disorders. New York. Silberfeld, M . (1978). Psychological symptoms and social supports. Social Psychiatry, 13, 11-17. Small, A. C. (1984) The contribution of psychodiagnostic test results toward understanding anorexia nervosa. International Journal of Eating Disorders, 5(2), 47-59. Small, A. C , Madero, J., Gross, H . , Teagno, L., Leib, J, & Ebert, M . (1981). A comparative analysis of primary anorexics and schizophrenics on the MMPI. Journal of Clinical Psychology, 57(4), 733-736. Smart, D. E., Beumont, P. J. V., & George, G . C. W. (1976). Some personality characteristics of patients with anorexia nervosa. British Journal of Psychiatry, 128, 57-60. Solyom, L., Freeman, R. J., & Miles, J. E. (1982). A comparative psychometric study of anorexia nervosa and obsessive neurosis. Canadian Journal of Psychiatry. 27, 282-286. Solyom, L., Thomas, C. D., Freeman, R. J., & Miles, J. E. (1983). Anorexia nevosa: Obsessive-compulsive disorder or phobia? A comparative study. In P. L. Darby, P. E. Garfinkel, D. M . Gamer, & D. V. Coscina (eds.), Anorexia nervosa: Recent developments on research, (pp. 137-147). New York: Alan R. Liss, Inc. 137 Sours, J. A. (1969). Anorexia nervosa: nosology, diagnosis, developmental patterns, and power-control dynamics. In G . Caplan & S. Lebovici (eds.), Adolescence: Psychosocial perspectives, (pp. 185-212). New York: Basic Books. Sours, J. A. (1974). The anorexic nervosa syndrome. International Journal of Psycho-Analysis, 55, 567-572. Sours, J. A. (1981). Depression and the anorexic nervosa syndrome. Psychiatric Clinics of North America, 4(1), 145-158. Speck, R. V., & Attneave, G. L. (1973). Family networks. New York: Pantheon Books. Steinhausen, H . C., & Glanville, K. (1983). Follow-up studies of anorexia nervosa: A review of research findings. Psychological Medicine, 13, 239-249. Stonehill, E., & Crisp, A. H . (1977). Psychoneurotic characteristics of patients with anorexia nervosa before and after treatment and at follow-up 4-7 years later. Journal of Psychosomatic Research, 21, 187-193. Strober, M . (1980). Personality and symptomatological features in young, nonchronic anorexic nervosa patients. Journal of Psychosomatic Research, 24, 353-359. Strober, M . (1981). A comparative analysis of personality organization in juvenile anorexia nervosa. Journal of Youth and Adolescence, 10(A), 185. Taipale, V., Tuomi, O., & Aukee, M . (1971). Anorexia nervosa: An illness of two generations? Acta Paedopsychia.try, 38, 21-25. Theander, S. (1970). Anorexia nervosa: A psychiatric investigation of 94 female patients. Acta Psychiatrica Scandinavica, (Supplement 214). Thoma, H . (1977). On the psychotherapy of patients with anorexia nervosa. Bulletin of the Menninger Clinic, 41(5), 437-452. Tolsdorf, C. C. (1976). Social networks and coping: An exploratory study. Family Process, 15(A), 407-417. Wiggins, J. S. (1982). Circumplex models of interpersonal behavior in clinical psychology. In P. C. Kendall & J. N . Butcher (eds.), Handbook on reseach methods in clinical psychology. New York: Wiley Interscience. Wiggins, J. S., Steiger, J. H . , & Gaelick, L. (1981). Evaluating circumplexity in personality data. Multivariate Behavioral Research, 16, 263-286. Williams, R.J. (1981). User's manual for CPI/TRS80 [Computer program] Odessa: Psychological Assessment Resources, Inc. Wilson, C. P. (1980). The family psychological profile of anorexia nervosa patients. Journal of the Medical Society of N. J., 77(5), 341-344. Winokur, A., March, V., & Mendels, J. (1980). Primary affective disorders in relatives of patients with anorexia nervosa, American Journal of Psychiatry, 137(6), 695-698. 138 Wold, P. (1973). Family structure in anorexia nervosa. American Journal of Psychiatry, 130, 1394. Yager, J. (1982). Family issues in the pathogenesis of anorexia nervosa. Psychosomatic Medicine, 44, 43-60. Appendix A Comparisons with Population Means 140 T a b l e A - l Means, Standard Deviations and Results of the Multivariate Analysis of CPI-Clinical Variables for the Control Subjects and the Population Means the CPI and Controls Population Variable M SD M t P 26.5 8.5 25.6 0.59 >.55 18.9 8.4 18.1 0.56 >.55 25.8 9.3 23.0 1.73 <.05 36.9 7.4 33.7 2.55 <.01 22.1 7.8 20.3 1.35 >.15 35.0 5.9 35.4 -0.39 >.65 28.8 7.0 30.3 -1.25 >.20 38.6 5.9 38.4 0.15 >.85 29.0 8.4 27.7 0.94 >.35 22.4 8.6 20.7 0.44 >.65 16.3 7.9 16.3 -0.00 >.95 26.7 6.4 25.8 0.84 >.40 27.4 7.4 25.6 1.41 >.15 20.7 8.2 17.8 1.97 <.05 38.8 7.8 36.8 1.52 X l O 12.4 9.3 9.9 1.57 X l O 12.5 8.1 9.9 1.87 <.05 24.2 8.4 20.9 2.29 <.05 CPI 1 Dominance Capacity for Status Sociability Social Presence Self-Acceptance Sense of Well-being Responsibility Socialization Self-Control Tolerance Good Impression Communality Achievement via Conformance Achievement via Independence Intellectual Efficiency Psychological Mindedness Flexibility Femininity CPI-Clinical 2 Lie Score Fake Bad K Scale Hypochondriasis Depression Hysteria Psychopathic Deviate Masculinity/ Femininity Paranoia Psychasthenia Schizophrenia Mania Social Introversion 44.6 6.7 50.0 -4.69 <.001 50.8 8.2 50.0 0.56 X 5 5 54.7 8.2 50.0 3.31 <.002 53.5 6.9 50.0 3.02 <.004 53.7 11.2 50.0 1.94 <.05 57.8 7.7 50.0 5.96 <.001 57.2 9.9 50.0 4.22 <.005 48.9 7.2 50.0 -0.82 X 4 0 55.6 6.9 50.0 4.69 <.001 53.7 10.5 50.0 2.05 <.05 55.0 11.2 50.0 2.60 <.01 52.6 8.0 50.0 1.91 <.05 46.8 11.9 50.0 -1.57 X l O Note: For CPI Hotelling's T2 analysis, F(18,16)=4.39, /K.005 For CPI-Clinical Hotelling's T 2 analysis, F(13,21)=4.85, p<.0005 •Weighted population means for the CPI were calculated from high school and college student norms (Gough, 1975) due to the age range (15-23 yrs.) of the subj sets. 2Mean scores for the CPI-Clinical are reported in T-score equivalents. Table A-2 Means, Standard Deviations and Results of the Multivariate Analysis of the CPI and CPI-Clinical Variables for the Anorexic Subjects and the Population Means Anorexics Population Variable M SD M t P CPI 1 Dominance 20.7 5.6 25.6 -4.83 <.001 Capacity for Status 14.5 4.7 18.1 -4.13 <.0005 Sociability 20.8 5.3 23.0 -2.26 <.05 Social Presence 30.8 7.9 33.7 -1.99 <.05 Self-Acceptance 19.2 3.6 20.3 -1.62 X l O Sense of Well-being 24.9 9.6 35.4 -6.04 <.0001 Responsibility 23.3 6.4 30.3 -5.97 <.0001 Socialization 31.7 6.1 38.4 -6.00 <.0001 Self-Control 21.2 10.3 27.7 -3.48 <.001 Tolerance 15.9 7.1 20.7 -4.52 <.0001 Good Impression 10.9 5.6 16.3 -5.34 <.0001 Communality 23.9 2.8 25.8 -3.65 <.001 Achievement via Conformance 21.0 5.9 25.6 -4.17 <.0005 Achievement via Independence 16.8 5.5 17.8 -1.13 >.25 Intellectual Efficiency 31.4 7.4 36.8 -3.96 <.0005 Psychological Mindedness 8.7 3.3 9.9 -1.94 <.05 Flexibility 8.7 4.1 9.9 -1.64 >.10 Femininity 23.6 3.2 20.9 4.55 <.0001 CPI-Clinical 1 Lie Score 42.7 7.3 50.0 -5.48 <.0001 Fake Bad 62.4 13.3 50.0 5.11 <.0001 K Scale 50.0 9.6 50.0 -0.00 1.00 Hypochondriasis 66.0 13.5 50.0 6.51 <.0001 Depression 65.6 11.7 50.0 7.30 <.0001 Hysteria 63.9 9.0 50.0 8.46 <.0001 Psychopathic Deviate 69.7 11.3 50.0 9.54 <.0001 Masculinity/ Femininity 49.2 7.8 50.0 -0.54 >.55 Paranoia 61.3 7.0 50.0 8.76 <.0001 Psychasthenia 68.9 14.7 50.0 7.03 <.0001 Schizophrenia 72.9 16.9 50.0 7.40 <.0001 Mania 57.2 7.9 50.0 4.93 <.0001 Social Introversion 51.7 10.7 50.0 0.85 >.40 Note: For CPI Hotelling's V analysis, F(18,12) = 6.45, /K.001 For CPI-Clinical Hotelling's T 2 analysis, F(13,17) = 9.33, ^<.0001 'Weighted population means for the CPI were calculated from high school and college student norms (Gough, 1975) due to the age range (15-23 yrs.) of the subjects. 2Mean scores for the CPI-Clinical are reported in T-score equivalents. Do Ci Sy Sp Sa Wb Ra So Sc To Gi Ca Ac Ai I* Fy F* Fa Do Ci Sy Sp Sa Wb R* So Sc To Gi Cm Ac Ai U Py Fi Fa Profiles of CPI Mean Scores for Anorexic and Control Subjects (Rcproducted from Manual for The California Psychological Inventory, by H. G . Gough, Ph.D. Copyright by Consulting Psychologists Press, Inc., Palo Alto, California, reprinted by permission) to / 144 Table A-3 Means, Standard Deviations and Results of the Multivariate Analysis of the CPI and CPI-Clinical Variables for die Mothers of the Control Subjects and the Population Means Mothers of Controls Population Variable M SD M t P CPI 1 Dominance 27.0 6.1 26.8 .21 >.80 Capacity for Status 18.8 4.3 20.1 -1.60 X l O Sociability 22.9 5.5 24.5 -1.56 X l O Social Presence 33.1 7.2 34.1 -0.74 X 4 0 Self-Acceptance 20.1 4.9 20.0 .15 X 8 5 Sense of Well-being 36.4 3.5 37.5 -1.71 <.05 Responsibility 31.0 3.8 32.1 -1.58 X l O Socialization 38.7 4.6 39.5 -1.00 X 3 0 Self-Control 33.8 6.5 32.0 1.48 X l O Tolerance 22.6 5.1 23.0 -0.47 X 6 0 Good Impression 18.8 5.9 20.0 -1.10 X 2 5 Communality 26.1 1.7 . 25.8 .88 X 3 5 Achievement via Conformance 23.0 3.9 28.2 -0.23 X 8 0 Achievement via Independence 20.9 4.6 19.0 2.28 <.01 Intellectual Efficiency 38.9 5.1 39.0 -0.14 X 8 5 Psychological Mindedness 10.8 2.7 11.0 -0.33 X 7 0 Flexibility 10.2 4.0 9.0 1.60 X l O Femininity 24.5 3.3 23.0 2.51 <.01 CPI-Clinical 2 Lie Score 49.2 7.7 50.0 -0.54 X 5 5 Fake Bad 51.9 7.3 50.0 1.42 X 1 5 K Scale 57.9 7.2 50.0 6.07 <.0001 Hypochondriasis 57.8 8.0 50.0 5.30 <.0001 Depression 58.9 8.9 50.0 5.49 <.0001 Hysteria 60.3 8.2 50.0 6.87 <.0001 Psychopathic Deviate 57.3 7.0 50.0 5.69 <.0001 Masculinity/ Femininity 46.3 4.9 50.0 -4.09 <.0005 Paranoia 57.9 4.1 50.0 10.48 <.0001 Psychasthenia 53.1 9.2 50.0 1.82 <.05 Schizophrenia 54.1 5.7 50.0 3.95 <.0005 Mania 51.9 7.9 50.0 1.34 X 1 5 Social Introversion 53.9 9.8 50.0 2.22 <.05 Note: For CPI Hotelling's T 2 analysis, F(18,12) = 9.15, /K.0005 For CPI-Clinical Hotelling's T 2 analysis, F(13,17)=45.29, ^<.0001 'Population means were obtained from the CPI manual (Gough, 1975). 2Mean scores for the CPI-Clinical are reported in T-score equivalents. 145 Table A-4 Means, Standard Deviations and Results of the Multivariate Analysis of the CPI and CPI-Clinical Variables for the Mothers of the Anorexic Subjects and the Population Means Mothers of Anorexics Population Variable M SD M t P CPI 1 Dominance 24.0 5.2 26.8 -2.40 <.05 Capacity for Status 18.6 4.9 20.1 -1.62 >.10 Sociability 21.8 4.6 24.5 -2.82 <.01 Social Presence 32.7 5.4 34.1 -1.21 X 2 0 Self-Acceptance 18.9 3.3 20.0 -1.53 >.10 Sense of Well-being 33.2 6.4 37.5 -3.45 <.005 Responsibility 29.8 4.8 32.1 -2.53 <.05 Socialization 34.9 5.2 39.5 -4.62 <.0001 Self-Control 31.6 6.4 32.0 -0.62 >.55 Tolerance 21.1 6.7 23.0 -1.44 >.15 Good Impression 17.3 5.8 20.0 -2.47 <.05 Communality 24.9 2.2 25.8 -2.06 <.05 Achievement via Conformance 25.4 4.5 28.2 -3.22 <.005 Achievement via Independence 20.5 4.9 19.0 1.58 X l O Intellectual Efficiency 36.6 _ 7.3 39.0 -1.76 X 0 5 Psychological Mindedness 11.1 3.8 11.0 0.26 X 7 5 Flexibility 10.2 4.7 9.0 1.35 X 1 5 Femininity 23.6 2.7 23.0 1.33 X 1 5 CPI-Clinical 2 Lie Score 47.1 7.7 50.0 -1.94 X 0 5 Fake Bad 55.0 9.4 50.0 2.73 <.01 K Scale 55.5 8.0 50.0 3.52 <.001 Hypochondriasis 58.2 8.1 50.0 5.17 <.0001 Depression 62.0 10.3 50.0 5.91 <.0001 Hysteria 63.2 9.4 50.0 7.18 <.0001 Psychopathic Deviate 66.3 9.6 50.0 8.65 <.0001 Masculinity/ Femininity 49.1 6.7 50.0 -0.68 X 5 0 Paranoia 60.6 6.7 50.0 8.02 <.0001 Psychasthenia 58.4 10.3 50.0 4.15 <.0001 Schizophrenia 61.3 12.0 50.0 4.80 <.0001 Mania 55.4 8.2 50.0 3.37 <.001 Social Introversion 54.2 7.7 50.0 2.75 <.01 Note: For CPI Hotelling's T 2 analysis, F(18,8) = 3.43, p<.05 For CPI-Clinical Hotelling's T 2 analysis, F(13,13) = 7.22, /K.0005 'Population means were obtained from the CPI manual (Gough, 1975). 2Mean scores for the CPI-Clinical are reported in T-score equivalents. Do Ci Sy Sp Sa Wb Ro So Sc To Gi Cm Ac Al lo Py Fa Fa 90 -eo 70 60 -o 40 30 20-10 - 55 -25 — SO 50 - U i j - 2 0 . - S S - S O - 2 9 — SO -so - 2 3 - 4 0 - 3 5 - 2 0 - 2 0 15 - 15 - 15 - 2 5 - S O - 2 0 -10 - 10 - 10- - 2 5 - 10 • IS - 5 - 2 0 FEMALE NORMS - 3 0 - 4 5 -40 •35 - 2 5 •35 - 2 5 - 20 - 2 5 -- 15 - 15 - 5 - 25 - 2 0 10 10-- O - 2 0 - 5 - 5 - 10 - 15 - 15 - 3 0 - 2 5 - 5 0 - 4 5 r9 -ss - 2 0 - 10 -20 - 2 0 - 5 - !T - 15 •15-Anorexic Control - 10 - o i i 90 ao TO 40 30 - 20 10 Do Ca Sy Sp Sa Wb Ra So Sc To Gi Can Ac Ai la Py Fa Fa Graph A - 3 Profiles of CPI Mean Scores for Mothers of Anorexic and Control Subjects (Reproducted from Manual for The California Psychological Inventory, by H . G . Gough, Ph.D. Copyright by Consulting Psychologists Press, Inc., Palo Alto, California, reprinted by permission) 7 0 Graph A - 4 Profiles of CPI-Clinical Mean Scores for Mothers of Anorexic and Control Subjects 148 Table A-5 Means, Standard Deviations and Results of the Multivariate Analysis of the CPI and CPI-Clinical Variables for the Fathers of the Control Subjects and the Population Means Fathers of Controls Population Variable M SD M CPI ' Dominance 30.6 10.7 26.8 1.75 >.05 Capacity for Status 21.1 7.2 19.4 1.17 >.25 Sociability 25.0 8.1 24.5 0.33 >.70 Social Presence 35.8 6.4 34.1 1.29 >.20 Self-Acceptance 22.3 7.8 19.2 1.93 >.05 Sense of Well-being 39.4 5.5 37.5 1.71 >.10 Responsibility 32.3 7.5 31.1 0.81 >.40 Socialization 38.6 5.1 36.7 1.82 >.05 Self-Control 35.1 8.2 31.0 2.47 <.05 Tolerance 23.4 7.7 22.9 0.33 >.70 Good Impression 22.9 8.5 20.0 1.71 >.10 Communality 27.3 6.8 25.2 1.53 X l O Achievement via Conformance 30.3 8.2 27.6 1.63 X l O Achievement via Independence 21.2 10.1 18.6 1.27 X 2 0 Intellectual Efficiency 38.8 6.5 39.3 -0.38 X 7 0 Psychological Mindedness 13.1 8.0 11.0 1.27 X 2 0 Flexibility 9.0 8.9 9.0 0.02 X 9 5 Femininity 20.4 9.2 16.2 2.25 <.05 CPI-Clinical 1 Lie Score 53.5 7.4 50.0 2.32 <.05 Fake Bad 49.8 2.9 50.0 -0.28 X 7 5 K Scale 55.8 6.1 50.0 4.71 <.0001 Hypochondriasis 62.7 7.7 50.0 8.05 <.0001 Depression 63.2 6.2 50.0 10.49 <.0001 Hysteria 64.8 7.9 50.0 9.11 <.0001 Psychopathic Deviate 57.5 6.1 50.0 6.02 <.0001 Masculinity/ Femininity 62.6 3.9 50.0 15.41 <.0001 Paranoia 55.2 6.2 50.0 4.14 <.0005 Psychasthenia 55.6 13.6 50.0 2.01 <.05 Schizophrenia 55.7 10.5 50.0 2.66 <.01 Mania 51.9 5.7 50.0 1.68 X l O Social Introversion 51.8 8.7 50.0 0.98 X 3 0 Note: For CPI Hotelling's V analysis, F(18,6)=1.90, p>.20 For CPI-Clinical Hotelling's V analysis, F(13,ll) = 26.73, Jp<.0001 'Population means were obtained from the CPI manual (Gough, 1975). JMean scores for the CPI-Clinical are reported in T-score equivalents. 149 Table A-6 Means, Standard Deviations and Results of the Multivariate Analysis of the CPI and CPI-Clinical Variables for the Fathers of the Anorexic Subjects and the Population Means Fathers of Anorexics Population Variable M SD M t P CPI 1 Dominance 28.9 7.0 26.8 1.34 >.15 Capacity for Status 18.8 3.6 19.4 -0.74 >.45 Sociability 23.3 5.5 24.5 -0.98 >.30 Social Presence 34.9 6.1 34.1 0.57 >.55 Self-Acceptance 21.8 4.0 19.2 2.74 <.01 Sense of Well-being 34.5 6.3 37.5 -2.05 <.05 Responsibility 26.7 6.5 31.1 -2.95 <.001 Socialization 33.5 6.9 36.7 -1.98 >.05 Self-Control 28.2 9.6 31.0 -1.29 >.20 Tolerance 20.3 6.2 22.9 -1.82 >.05 Good Impression 15.6 6.0 20.0 -3.19 <.005 Communality 25.5 2.3 25.2 0.63 >.50 Achievement via Conformance 25.0 5.5 27.6 -2.04 <.05 Achievement via Independence 20.2 3.7 18.6 1.82 >.05 Intellectual Efficiency 37.2 4.5 39.3 -2.08 <.05 Psychological Mindedness 11.8 2.8 11.0 1.23 >.20 Flexibility 9.5 3.7 9.0 0.56 >.55 Femininity 17.6 3.0 16.2 1.99 >.05 CPI-Clinical 2 Lie Score 48.5 7.7 50.0 -0.84 >.40 Fake Bad 55.3 8.2 50.0 2.83 <.01 K Scale 50.0 5.9 50.0 0.04 >.95 Hypochondriasis 61.6 12.0 50.0 4.19 <.0005 Depression 68.2 10.4 50.0 7.63 <.0001 Hysteria 64.9 5.6 50.0 11.67 <.0001 Psychopathic Deviate 65.4 11.6 50.0 5.77 <.0001 Masculinity/ Femininity 61.9 3.2 50.0 16.16 <.0001 Paranoia 60.3 6.3 50.0 7.18 <.0001 Psychasthenia 58.3 10.2 50.0 3.52 <.005 Schizophrenia 61.7 14.7 50.0 3.45 <.005 Mania 53.0 10.1 50.0 1.32 >.20 Social Introversion 51.5 7.3 50.0 0.91 >.30 Note: For CPI Hotelling's T 2 analysis, F(18,l) = 2.76, /?>.40 For CPI-Clinical Hotelling's T 2 analysis, F(13,l) = 45.91, /K.0001 'Population means were obtained from the CPI manual (Gough, 1975). 2Mean scores for the CPI-Clinical are reported in T-score equivalents. Pa Ca Sy S» Sa Wb to Sa Sc To Gl Ca Ac Ai U Py fx U Do Ca Sy Sa Sa Wb to So Sc To Gl C M Graph A-5 Profiles of CPI Mean Scores for Fathers of Anorexic and Control Subjects (Reproducted from Manual for The California Psychological Inventory, by H. G . Gough, Ph.D. Copyright by Consulting Psychologists Press, Palo Alto, California, reprinted by permission) Graph A-6 Profiles of CPI-Clinical Mean Scores for Fathers of Anorexic and Control Subjects 152 Table A - 7 Means, Standard Deviations and Results of the Multivariate Analysis of the FES Variables for the Families of the Control Subjects and the Population Means Families of Controls Population1 Variable M SD M t P Cohesion 7.14 1.55 6.53 2.29 <.05 Expressiveness 5.50 1.69 5.26 0.86 >.35 Conflict 2.65 1.82 3.85 -3.82 <.O01 Independence 6.82 1.02 6.46 2.07 <.05 Achievement 5.30 1.21 5.56 -1.21 >.20 Intellectual/Cultural 5.81 1.96 5.30 1.54 X l O Activity/Recreational 5.97 1.51 5.20 2.96 <.005 Moral/Religious 4.64 2.80 4.76 -0.24 X 8 0 Organization 5.54 2.05 5.35 0.56 X 5 5 Control 3.89 1.67 4.99 -3.80 <.001 Note: For FES Hotelling's T 2 analysis, F(10,24) = 3.80, p<.01 'Population means are for a 4 member family (Moos & Moos, 1981, FES Manual). 153 Table A - 8 Means, Standard Deviations and Results of the Multivariate Analysis of the FES Variables for the Families of the Anorexic Subjects and 1 the Population Means Families of Anorexics Population1 Variable M SD M t P Cohesion 5.34 2.03 6.53 -3.23 <.005 Expressiveness 4.18 1.94 5.26 -3.09 <.005 Conflict 3.97 2.30 3.85 0.30 >.75 Independence 5.58 1.84 6.46 -2.65 <.01 Achievement 5.16 1.64 5.56 -1.35 >.15 Intellectual/Cultural 5.16 2.12 5.30 -0.35 >.70 Activity/Recreational 5.16 1.95 5.20 -0.10 >.90 Moral/Religious 4.20 2.04 4.76 -1.51 >.10 Organization 4.96 2.18 5.35 -1.04 >.30 Control 4.80 2.00 4.99 -0.51 >.60 Note: For FES Hotelling's T 'Population means are analysis, F(10,21) = 2.71, for a 4 member family p<.05 (Moos & Moos, 1981, FES Manual). 154 1 0 0 9 0 8 0 7 0 £ 6 0 o u 1/1 "2 5 0 I 4 0 » — A 3 0 2 0 1 0 S/S Scale Anorexic -i : Control 1 0 0 - 9 0 - 8 0 7 0 6 0 £ 3 C so a n O 4 0 3 3 0 2 0 1 0 Con Ind AO ICO ARO MRE Org C t l Graph A-7 Profiles of FES Mean Scores for Families of Anorexic and Control Subjects (Scales developed by R. H . Moos and associates.© Copyright, 1975 by Consulting Psychologists Press, Inc., Palo Alto, California, reprinted by permission) Appendix B Results of ANOVA 156 T a b l e B - l Results of A N O V A for the Responses of the Daughters on the Affiliation Dimension Source of Variance df MS F P Between Subjects (A) (Exp. vs Con.) 1 187,434.13 23.31 <.001 Within Subjects Rater (B) 1 3,147.10 1.26 >.25 Target (C) 1 4,953.61 2.06 >.15 Interaction AxB 1 4.47 .00 >.95 A x C 1 1299.90 .54 >.45 B x C 1 2300.05 2.72 >.10 A x B x C 1 48.88 .06 >.80 Error Terms for A Main Effects 41 8041.78 for B & AB 41 2501.76 for C & A C 41 2408.49 for BC & ABC 41 846.44 Note: Rater=Daughter vs Parent; Target=Mother vs Father Appendix C T-Score Equivalents Table C - l T - S c o r e E q u i v a l e n t s f o r the C P I - C l i n i c a l M e a n Scores Anorexics Controls V a r i a b l e . R a w S c o r e T - S c o r e R a w S c o r e T - S c o r e (A) Daughter D e p r e s s i o n 27.69 65.56 21.52 53.70 H y s t e r i a 27.29 63.93 28.82 57.85 P s y c h o p a t h i c D e v i a t e 23.96 69.69 17.70 57.20 P s y c h a s t h e n i a 25.79 68.93 14.76 53.70 S c h i z o p h r e n i a 29.39 72.93 14.35 55.00 S o c i a l I n t r o v e r s i o n 29.79 51.66 35.08 46.79 ) Mothers D e p r e s s i o n 25.07 62.00 23.96 58.96 H y s t e r i a 26.19 63.19 24.63 60.03 P s y c h o p a t h i c D e v i a t e 19.57 66.26 15.09 57.29 P s y c h a s t h e n i a 14.46 58.38 12.06 53.06 S c h i z o p h r e n i a 15.23 61.34 9.40 54.09 S o c i a l I n t r o v e r s i o n 28.96 54.19 28.83 53.93 ) Fathers D e p r e s s i o n 24.26 68.21 22.16 63.25 H y s t e r i a 27.78 64.94 24.58 64.75 P s y c h o p a t h i c D e v i a t e 20.10 65.42 15.70 57.54 P s y c h a s t h e n i a 13.52 58.26 9.00 55.58 S c h i z o p h r e n i a 14.78 61.68 8.58 55.70 S o c i a l I n t r o v e r s i o n 26.47 51.52 26.66 51.75 Appendix D Instruments 160 INTRODUCTION TO THE QUESTIONNAIRES T h i s study is designed to e v a l u a t e f a m i l y r e l a t i o n s h i p s w i t h i n the f a m i l y and in the community. The q u e s t i o n n a i r e s I am a s k i n g you to answer w i l l g i v e me i n f o r m a t i o n about how your f a m i l y members r e l a t e to each o ther and the support system you have in the community. T h i s i n f o r m a t i o n w i l l help answer some important q u e s t i o n s about the nature of a n o r e x i a n e r v o s a . The i n f o r m a t i o n about your f a m i l y w i l l be t r e a t e d as anonymous, w i l l be kept e n t i r e l y c o n f i d e n t i a l , used f o r r e s e a r c h purposes o n l y , and w i l l be d e s t r o y e d at the end of i t s u s e f u l -n e s s . P a r t i c i p a t i o n in the p r o j e c t is v o l u n t a r y and withdrawal or r e f u s a l to answer any q u e s t i o n s w i l l in no way a f f e c t the treatment you r e c e i v e from your a t t e n d i n g p h y s i c i a n . Should you complete the q u e s t i o n n a i r e s i t w i l l be assumed that you have done so w i t h f u l l c o n s e n t . P a r t i c i p a t i o n in the p r o j e c t r e q u i r e s two or t h r e e hours of your time f o r answering q u e s t i o n n a i r e s . T h i s can be done in your own home d u r i n g a one week per i o d . Each p a r t i c i p a n t needs to complete 1 - C a l i f o r n i a Psychology Inventory (CPI) 1 - Family Environment S c a l e 1 - INTREX Q u e s t i o n n a i r e - Form A 1 - P a t t i s o n P s y c h o - S o c i a l Inventory F u r t h e r , each parent needs to complete 1 - INTREX Q u e s t i o n n a i r e - Form C f o r t h e i r daughter and the daughter needs to complete 1 - INTREX Q u e s t i o n n a i r e f o r EACH parent - Form B f o r her f a t h e r - Form C f o r her mother B o o k l e t s and q u e s t i o n n a i r e s can be shared when o n l y one copy is i n c l u d e d . PLEASE RETURN ALL QUESTION BOOKLETS AND ANSWER SHEETS. 161 STANDARD INTERPERSON ISTORY. (MALE SIGNIFICANI OTHER) D I R E C T I O N S for I N T R E X Q U E S T I O N N A I R E S This package of questionnaires asks you to rate ways you feel about yourself and some significant others. You are asked to rate each question on a scale of 0 (never, not at all) to 100 (always, perfectly). The analysis of the questionnaires organizes your answers in ways which can help you and your health care provider understand why you feel the way you do, and can even make some suggestions for what to do about it. People who go ahead and answer honestly, avoiding any temptation to "wh i tewash" or "paint a rosy picture" usually are very pleased with the results. If your health care provider agrees, the results can be made available to you to help you with your understanding of yourself. If you are in psychotherapy, it can help with therapy planning and can be used to measure progress. In most cases , this procedure, if used correctly, can improve the efficiency of psychotherapy. Sometimes it is helpful to use the INTREX questionnaires again in different " p h a s e s " of psychotherapy. Please answer the questions for how you really think or feel. Your initial reaction to each question will most often be your best answer. If a question offends you, score it zero or leave it blank. There are no "r ight" or " w r o n g " answers. It's your view which is important—not what is necessari ly " t rue , " " f a l s e " or what someone else might think you should say. If there are questions about the past (e.g. when you were age 5-10) try to recall your home and the people to get back into the " m o o d " of things as they were. Put your answers on the proper answer sheets making sure you match the letter of the questionnaire with the letler of the answer sheet. When you've finished, give the package of questionnaires and the answer sheets back to the person who gave you these materials. The results will be returned as quickly as possible to your health care provider. •Cr -Cr.. -d- -Cr PLEASE DO NOT WRITE ON THIS P A C K A G E A N D BE S U R E TO START WITH THE BACK OF THIS PAGE. INTREX FORM A Please use the answer sheet marked "A" and indicate how well each question describes yourself. Use the scale which appears at the top of the answer sheet. 1. I neglect myself, don't try to develop my own potential skills, ways of being. 2. I examine, analyze myself sensibly, carefully, realistically. 3. I leave myself to daydream and fantasize instead of actually doing what would be good for me. 4. I just let important choices, thoughts, issues, options slip by me without paying much attention. 5. Knowing both my faults and my strong points, I comfortably let myself be as I am. 6. I let myself feel glad about and pleased with myself just as I am. 7. I accuse and blame myself, make myself feel bad, guilty, ashamed, unworthy. 8. I practice, work on developing worthwhile skills, ways of being. 9. I love, cherish, adore, leel really good—maybe even sexy—about myself. 10. I naturally and easily nurture, care for, resiore, heal myself as needed. 11. I harshly reject, dismiss myself as worthless. 12. I let unwarranted ideas I have about myself go unchallenged. I don't bother to know myself. 13. I like myself very much and welcome and enjoy opportunities to be with myself. 14. I am very careful to restrain myself, to hold back. 15. I control, manage myself according to goals I've set for mysell. 16. I torture, kill, annihilate myself just because I'm basically so bad. 17. I drain, overburden, and deplete myself greatly. 18. I gently and warmly "pat myself on the back" just because I feel very good about the way I am. 19. I keep an eye on myself to be sure I'm doing what I should be doing. 20. I try very hard to make myself as ideal as I can. 21. I understand and accept myself and let myself go by what lies deep within. 22. I let my own sickness and injury go unattended even when it means harming myself greatly. 23. I put a lot of energy into making sure I conform to standards, am proper. 24. I vengefully punish mysell. I "take it out on myself." 25. I make myself do and be things which I know are not right for me. I fool myself. 26. I am happy-go-lucky, content with "here today, gone tomorrow." 27. I reliably protect myself, look after my own interests. 28. I drift with the moment, have no particular internal direction, standards. 29. I put a lot of energy into anticipating and finding everything I need for myself. 30. In a free and easy way, I let myself do what comes naturally, and everything goes well enough to suit me. 31. I understand and feel good about myself. I'm relaxed, solid, "together", completely okay " a s is ." 32. I feel free to let my basic nature unfold as it will. 33. I am reckless, carelessly end up in self-destructive situations. 34. I am always open to and "up for" situations which will be very pleasant and good for me. 35. I am very unsure of myself because I tell mysell I do things all wrong. I feel others can do better. 36. I approach myself with a negative, destructive attitude; I am my own worst enemy. ' © 1 9 8 0 , INTREX Interpersonal Institute, Inc. 163 INTREX FORM B* P l e a s e u s e t h e a n s w e r s h e e t m a r k e d "B" a n d i n d i c a t e h o w w e l l e a c h q u e s t i o n d e s c r i b e s YOUR SIGNIFICANT OTHER PERSON  U s e t h e s c a l e w h i c h a p p e a r s at t h e t o p of t h e a n s w e r s h e e t . 1. Constructively, sensibly, persuasively analyzes situations involving me. 2. Has his own separate identity, internal standards. 3. Insists*! follow,his norms and rules so that I do things "properly." 4. Puts me down, tells me I do things all wrong, tells me his ways are superior. 5. Learns from me, takes advice from me. 6. Just does things my way without much feeling of his own, is apathetic. 7. Angrily leaves me out, absolutely refuses to have anything to do with me. 8. Warmly, comfortably accepts help, caregiving when I offer it. 9. Does his own thing by doing the exact opposite of what I want. 10. Is straightforward. Clearly expresses his positions so I can give them due consideration. 11. Enthusiastically shows, shares himself or " th ing" with me. 12. Tortures, murders,-annihilates me no matter what I do just because "I'm me. " 13. Does strange, irrelevant, unrelated things with what I say or do; goes on his "own trip." 14. Ecstatically, joyfully, exuberantly, lovingly responds to me sexually. 15. Warmly, cheerfully invites me to be in touch with him as often as I want. 16. Warmly, happily keeps in contact with me. 17. Freely comes and goes as he pleases. 18. Out of great love for me, he tenderly, lovingly touches me sexually if I seem receptive. 19. Stimulates and teaches me, shows me how to understand, do. 20. Accuses and blames me; tries to get me to admit I am wrong. 21. Enthusiastically, very lovingly shows me how glad he is to see me just as I am. 22. Looks to me, depends on me to take care of everything for him. 23. Harshly punishes me, takes revenge, makes me sulfer greatly. 24. Understands me well, shows empathy and warmth even if I don't see things as he does. 25. Is trusting. Asks for what he wants and counts on me to be kind and considerate. 26. Willingly accepts, yields to my reasonable suggestions, ideas. 27. Screams, agonizes, protests desperately that I am destroying, killing him. 28. Gently strokes me verbally and/or physically; he lovingly gives me pleasure with "no strings attached." 29. Intrudes, blocks, restricts me. 30. Even though very suspicious and distrustful of me, he goes along with my arguments, ideas. 31. Obeys my preferred rules, standards, routines. 32. Rips me off, gouges me, grabs all he can from me. 33. Pleasing me is so important that he checks with me on every little thing. 34. Is obviously terrified, very fearful of me; is extremely wary. 35. Misleads, deceives, deludes and diverts me. 36. In a very grouchy, surly manner, he goes along with my needs and wants. — 37. Provides for, nurtures, takes care of me. 38. Lets me speak freely and can be trusted to negotiate fairly even if we disagree. 39. Ignores me, just doesn't notice me at all. 40. Uncaringly lets me go, do what I want. 41. Snarls angrily, hatefully refuses my caregiving, my offers to assist. 42. Filled with rage and/or fear, he does what he can to escape, flee, or hide from me. •©1980 , INTREX Interpersonal Institute, Inc. 164 43. Believing it's tor my own good, he checks on me and reminds me ol what I should do. 44. Gives me his "b less ing" and leaves me free to develop my own separate identity. 45. Forgets me, just doesn't remember our agreements, plans. 46. Gives in and does things the way I want, but sulks quietly with resentment and anger. 47. Yields, submits, gives in to me. 48. Approaches me very menacingly; hurts me very badly if he gets a chance. 49. Manages, controls me, takes charge of everything. 50. Leaves me to do things on my own because he believes I'm competent. 51 Expresses his thoughts in a clear and friendly manner so I have every opportunity to understand him well. 52. Feels, thinks, becomes what he thinks I want. 53. Leaves me to starve, to get what I vitally need all on my own. 54. Actively listens, accepts and affirms me as a person even if our views disagree. 55. Angrily detaches from me, doesn't ask for anything; weeps alone about me. 56. Pays close attention in order to anticipate all my needs; takes care of absolutely everything for me. 57 Whines, protests, tries to explain, justify, account for himself. 58. Asserts, holds his own without needing external support. 59. Avoids me by being busy and alone with his "own thing." 60. Warmly shows how much he likes and appreciates me just exactly as I am. 61. Walls himself off from me, doesn't hear, doesn't react. 62. Relaxes, enjoys, really lets go with me. Feels wonderful about being with me. 63. Believing it's for my own good, he tells me exactly what to do, be, think. 64. Buries his rage and resentment and scurries to appease me to avoid my disapproval. 65. Approaches me with unwarranted, even crazy ideas about me, and doesn't notice how or if I respond. 66. Goes his own separate way. 67. Looks after my interests, takes steps lo protect me, actively backs me up. 68. Freely and openly discloses his innermost self so I can truly know "who he is . " 69. Is joyful and exuberant and expects to have wonderful fun with me. 70. Just when he is needed most, he abandons me, leaves me " in the lurch." 71. Neglects me, my interests, needs. 72. Leaves me free to do and be whatever I want. 165 For questions #73-144, change from rating him to rating yourself in this relationship. Continue using the same scale at the top of answer sheet " B " . 73. I constructively, sensibly, persuasively analyze situations involving him. 74. I have my own separate identity, internal standards. 75. I insist he follow my norms and rules so that he does things "properly." 76. I put him down, tell him he does things all wrong, tell him my ways are superior. 77. I learn from him, take advice from him. 78. I just do things his way without much feeling of my own; I am apathetic. 79. I angrily leave him out, absolutely refuse to have anything to do with him. 80. I warmly, comfortably accept help, caregiving when he offers it. 81. I do my own thing by doing the exact opposite of what he wants. 82. I am straightforward. I clearly express my positions so he can give them due consideration. 83. I enthusiastically show, share myself or " th ing" with him. 84. I torture, murder, annihilate him no matter what he does just because he is who he is. 85. I do strange, irrelevant, unrelated things with whal he says or does; I go on my "own trip." 86. I ecstatically, joyfully, exuberantly, lovingly respond to him sexually. 87. I warmly, cheerfully invite him to be in touch with me as often as he wants. 88. I warmly, happily keep in contact with him. 89. I freely come and go as I please. 90. Out of great love for him, I tenderly, lovingly touch him sexually if he seems receptive. 91. I stimulate and teach him, show him how to understand, do. 92. I accuse and blame him; try to get him to admit he is wrong. 93. I enthusiastically, very lovingly show him how glad I am to see him just as he is. 94. I look to him, depend on him to take care of everything for me. 95. I harshly punish him, take revenge, make him suffer greatly. 96. I understand him well, show empathy and warmth even if he doesn't see things as I do. 97. I am trusting. I ask for what I want and count on him to be kind and considerate. 98. I willingly accept, yield to his reasonable suggestions, ideas. 99. I scream, agonize, protest desperately that he is destroying, killing me. 100. I gently stroke him verbally and/or physically; I lovingly give him pleasure with "no strings attached." 101. I intrude, block, restrict him. 102. Even though very suspicious and distrustful of him, I go along with his arguments, ideas. 103. I obey his preferred rules, standards, routines. 104. I rip him off, gouge him, grab all I can from him. 105. Pleasing him is so important that I check with him on every little thing. 106. I am obviously terrified, very fearful of him; I am extremely wary. 107. I mislead, deceive, delude and divert him. 108. In a very grouchy, surly manner, I go along with his needs and wants. 109. I provide for, nurture, take care of him. 110. I let him speak freely and can be trusted to negotiate fairly even if we disagree. 111. I ignore him, just don't notice him at all. 112. I uncaringly let him go, do what he wants. 113. I snarl angrily, hatefully refuse his caregiving, his offers to assist. 114. Filled with rage and/or fear, I do what I can to escape, flee, or hide from him. 115. Believing it's for his own good, I check on him and remind him of what he should do. 116. I give him my "b less ing" and leave him free to develop his own separale identity. 117. I forget him. just don't remember our agreements, plans. 118. I give in and do things the way he wants, but sulk quietly with resentment and anger. 119. I yield, submit, give in to.him. 120. I approach him very menacingly; I hurt him very badly if I get a chance. 121. I manage, control him, take charge of everything. 122. I leave him to do things on his own because I believe he is competent. 123. I express my thoughts in a clear and friendly manner so he has every opportunity to understand me well. 124. I feel, think, become what I think he wants. 125. I leave him to starve, to get what he vitally needs all on his own. 126. I actively listen, accept and affirm him as a person even if our views disagree. 127 I angrily detach from him, don't ask lor anything; I weep alone about him. 128. I.pay close attention in order to anticipate all his needs; I take care of absolutely everything for him. 129. I whine, protest, try to explain, justify, account for myself. 130. I assert, hold my own without needing external support. 131. I avoid him by being busy and alone with my "own thing." 132. I warmly show how much I like and appreciate him just exactly as he is. 133. I wall myself off from him, don't hear, don't react. 134. I relax, enjoy, really let go with him. I feel wonderful about being with him. 135. Believing it's for his own good, I tell him exactly what to. do, be, think. 136. I bury my rage and resentment and scurry to appease him to avoid his disapproval. 137. I approach him with unwarranted, even crazy ideas about him; I don't notice how or if he responds. 138. I go my own separale way. 139 I look after his interests, take steps to protect him, actively back him up 140. I freely and openly disclose my innermost sell so he truly can know "who I am . " 141. I am joyful and exuberant and expect to have wonderful fun with him. 142 Just when I am needed most, I abandon him, leave him " in the lurch." 143. I neglect him, his interests, needs. 144. I leave him free to do and be whatever he wants. 167 INTREX FORM C* Please use the answer sheet marked "C" and indicate how well each question descri Use the scale which appears at the top of the answer sheet. 1. Constructively, sensibly, persuasively analyzes situations involving me. 2. Has her own separate identity, internal standards. 3. Insists I lollow her norms and rules so that I do things "properly." 4. Puis me down, tells me I do things all wrong, tells me her ways are superior. 5 Learns from me, takes advice from me 6. Just does things my way without much feeling of her own, is apathetic. 7. Angrily leaves me out, absolutely refuses to have anything to do with me. 8. Warmly, comfortably accepts help, caregiving when I offer it. 9. Does her own thing by doing the exact opposite of what I want. 10. Is straightforward. Clearly expresses her positions so I can give them due consideration. 11. Enthusiastically shows, shares herself or " th ing" with me. 12. Tortures, murders, annihilates me no matter what I do just because "I 'm me. " 13. Does strange, irrelevant, unrelated things with what I say or do; goes on her "own trip." 14. Ecstatically, joyfully, exuberantly, lovingly responds to me sexually. 15. Warmly, cheerfully invites me to be in touch with her as often as I want. 16. Warmly, happily keeps in contact with me. 17. Freely comes and goes as she pleases. 18. Out of great love for me, she tenderly, lovingly touches me sexually if I seem receptive. 19. Sl imulaies and teaches me, shows me how to understand, do. 20. Accuses and blames me: tries to get me to admit I am wrong. 21. Enthusiastically, very lovingly shows me how glad she is to see me just as I am. 22. Looks to me, depends on me to take care of everything for her. 23. Harshly punishes me, takes revenge, makes me suffer greatly. 24. Understands me well, shows empathy and warmth even if I don't see things as she does. 25. Is trusting. Asks for what she wants and counts on me to be kind and considerate. 26. Willingly accepts, yields to my reasonable suggestions, ideas. 27. Screams, agonizes, protests desperately that I am destroying, killing her. 28. Gently strokes me verbally and/or physically; she lovingly gives me pleasure with "no strings attached." 29. Intrudes, blocks, restricts me. 30. Even though very suspicious and distrustful of me, she goes along with my arguments, ideas. 31. Obeys my preferred rules, standards, routines. 32. Rips me off, gouges me, graDs all she can from me. 33. Pleasing me is so important that she checks with me on every little thing. 34. Is obviously terrified, very tearful of me, is extremely wary. 35. Misleads, deceives, deludes and diverts me. 36. In a very grouchy, surly manner, she goes along with my needs and wants. 37. Provides for, nurtures, takes care ol me. 38 Lets me speak freely and can be trusted to negotiate fairly even if we disagree. 39. Ignores me, just doesn't notice me at all. 40. Uncaringly lets me go. do what I want. 41. Snarls angrily, hatefully refuses my caregiving, my offers to assist. 42. Filled with rage and/or fear, she does what she can to escape, flee or hide from me. •©1980 , INTREX Interpersonal Institute, Inc 43 Believing it's for my own good, she checks on me and reminds me of what I should do. 44. Gives me her "blessing'' and leaves me free to develop my own separate identity. 45. Forgets me, just doesn't remember our agreements, plans. 46. Gives in and does things the way I want, but sulks quietly with resentment and anger. 47. Yields, submits, gives in to me. 48. Approaches me very menacingly; hurts me very badly if she gets a chance. 49 Manages, controls me, takes charge of everything. 50. Leaves me to do things on my own because she believes I am competent. 51. Expresses her thoughts in a clear and friendly manner so I have every opportunity to understand her well. 52. Feels, thinks, becomes what she thinks I want. 53 Leaves me to starve, to get what I vitally need all on my own. 54. Actively listens, accepts and affirms me as a person even if our views disagree. 55. Angrily detaches from me, doesn't ask for anything; weeps alone about me. 56. Pays close attention in order to anticipate all my needs; takes care of absolutely everything for me. 57. Whines, protests, tries to explain, justify, account for herself. 58. Asserts, holds her own without needing external support. 59. Avoids me by being busy and alone with her "own thing." 60. Warmly shows how much she likes and appreciates me just exactly as I am. 61. Walls herself off from me, doesn't hear, doesn't react. 62. Relaxes, enjoys, really lets go with me. Feels wonderful about being with me. 63. Believing it's for my own good, she tells me exactly what to do, be, think. 64. Buries her rage and resentment and scurries to appease me to avoid my disapproval. 65. Approaches me with unwarranted, even crazy ideas about me, and doesn't notice how or if I respond. 66. Goes her own separate way. 67. Looks atler my interests, takes steps to protect me, actively backs me up. 68. Freely and openly discloses her innermost self so I can truly know "who she is." 69. Is joyful and exuberant and expects to have wonderful fun with me. 70. Just when she is needed most, she abandons me, leaves me "in the lurch." 71. Neglects me, my interests, needs. 72. Leaves me free to do and be whatever I want. 169 For questions #73-144, change Irom rating her to rating yourself in this relationship. Continue using the same scale at the top of answer sheet "C . " 73. I constructively, sensibly, persuasively analyze situations involving her. 74. I have my own separate identity, internal standards. 75. I insist she follow my norms and rules so that she does things "properly." 76 I put her down, tell her she does things all wrong, tell her my ways are superior. 77. I learn from her, take advice from her. 7B. I just do things her way without much feeling of my own; I am apathetic. 79. I angrily leave her out, absolutely refuse to have anything to do with her. 80. I warmly, comfortably accept help, caregiving when she offers it. 81 I do my own thing by doing the exact opposite of what she wants. 82. I am straightforward. I clearly express my positions so she can give them due consideration. 83. I enthusiastically show, share myself or "thing" with her. 84. I torture, murder, annihilate her no matter what she does just because she is who she is. 85. I do strange, irrelevant, unrelated things with what she says or does; I go on my "own trip." 86. I ecstatically, joyfully, exuberantly, lovingly respond to her sexually. 87. I warmly, cheerfully invite her to be in touch with me as oflen as she wants. 88: I warmly, happily keep in contact with her. 89. I freely come and go as I please. 90. Out of great love for her, I tenderly, lovingly touch her sexually if she seems receptive. 91. I stimulate and teach her. show her how to understand, do. 92. I accuse and blame her; try to gel her to admit she is wrong. 93. I enthusiastically, very lovingly show her how glad I am to see her just as she is. 94. I look to her, depend on her to take care of everything for me. 95. I harshly punish her, take revenge, make her suffer greatly. 96. I understand her well, show empathy and warmth even if she doesn't see things as I do. 97. I am trusting. I ask tor what I want and count on her to be kind and considerate. 98. I willingly accept, yield to her reasonable suggestions, ideas. 99. I scream, agonize, protest desperately that she is destroying, killing me. 100. I gently stroke her verbally and/or physically; I lovingly give her pleasure with "no strings attached." 101. I intrude, block, restrict her. 102. Even though very suspicious and distrustful of her, I go along with her arguments, ideas. 103. I obey her preferred rules, standards, routines. 104. I rip her off, gouge her. grab all I can from her. 105. Pleasing her is so important that I check with her on every little thing. 106. I am obviously terrified, very fearful of her; I am extremely wary. 107. I mislead, deceive, delude and divert her. 108. In a very grouchy, surly manner, I go along with her needs and wants. 109. I provide for, nurture, take care of her. 110. I let her speak freely and can be trusted to negotiate fairly even if we disaoree 111. I ignore her, just don't notice her at all. 112. I uncaringly let her go, do what she wants. 113. I snarl angrily, hatefully refuse her caregiving, her offers to assist. 114. Filled with rage and/or fear, i do what I can to escape, flee, or hide from her. 170 1 1 5 Believing it's for her own good, I check on her and remind her of what she should do. 116 l give her my "blessing" and leave her free to develop her own separate identity. 117. I forget her, just don't remember our agreements, plans. 118. I give in and do things the way she wants, but sulk quietly with resentment and anger. 119. I yield, submit, give in to her. 1 2 0 . I approach her very menacingly, I hurt her very badly if I get a chance. 1 2 1 . I manage, control her, take charge of everything. 1 2 2 . I leave her to do things on her own because I believe she is competent. 1 2 3 . I express my thoughts in a clear and friendly manner so she has every opportunity to understand me well. 1 2 4 . I feel, think, become what I think she wants. 1 2 5 . I leave her to starve, to get what she vitally needs all on her own. 1 2 6 . I actively listen, accept and affirm her as a person even if our views disagree. 127. I angrily detach from her, don't ask for anything; I weep alone about her. 1 2 8 I pay close attention in order to anticipate all her needs; I take care of absolutely everything lor her. 1 2 9 . I whine, protest, try to explain, justify, account for myself. 1 3 0 . I assert, hold my own without needing external support. 1 3 1 . I avoid her by being busy and alone with my "own thing." 1 3 2 . I warmly show how much I like and appreciate her just exactly as she is. 1 3 3 . I wall myself off from her, don't hear, don't react. 1 3 4 . I relax, enjoy, really let go with her. I feel wonderful aboul being with her. 1 3 5 . Believing it's for her own good. I tell her exactly what to do, be, think. 1 3 6 . I bury my rage and resentment and scurry to appease her to avoid her disapproval. 1 3 7 . I approach her with unwarranted, even crazy ideas about her; I don't notice how or if she responds. 1 3 8 . I go my own separale way. 1 3 9 . I look after her interests, take steps to protect her, actively back her up. 1 4 0 . I freely and openly disclose my innermost self so she truly can know "who I am." 1 4 1 . I am joyful and exuberant and expect to have wonderful fun with her. 1 4 2 . Just when I am needed most, I abandon her, leave her "in the lurch." 1 4 3 . I neglect her, her interests, needs. 1 4 4 . I leave her free to do and be whatever she wants. 171 Pattison Psycho-Social Network Inventory INSTRUCTIONS (Part 1) L i s t by f i r s t name and l a s t i n i t i a l a l l persons who are- IMPORTANT i n your l i f e at this moment, whether you l i k e them or not. These persons may be, for example family members, r e l a t i v e s , f r i e n d s , neighbors, workmates, clergy, bosses, r e c r e a t i o n a l associates, etc. Use your own d e f i n i t i o n of who i s important and use the following to guide yourself. A f t e r l i s t i n g each person, f i l l i n each person's sex, how long you've known each other, t h e i r age, and the r e l a t i o n s h i p to you. SOCIAL NETWORK INVENTORY WORK SHEET Name and Last I n i t i a l Sex Years Known Age Relationship A B C D E. F G Bob A. M 12 33 husband 5 ' 5 4 5 5 5 5 Johnny A. M 3 3 son 5 5 4 5 3 4 5 Nancy B. F 6 29 g i r l f r i e n d 5 4 3 5 5 4 5 Susan C. F 1 30 f r i e n d 4 4 3 4 4 3 4 Marge D. F 30 59 mother 3 3 3 4 3 3 4 Carol E. F 4 35 neighbor 4 4 3 4 3 3 4 Mark A. M 27 27 step-brother 2 1 3 3 2 2 3 P h i l 0. M 1 40 supervisor 5 3 3 4 3 2 4 INSTRUCTIONS (Part 2) Refer to the scales marked A through G below. These scales correspond to the A through G that head the columns on the Inventory Work Sheet. Each scale has 5 possible choices. S t a r t i n g with scale A, take each person l i s t e d i n turn and sel e c t the choice that best describes your r e l a t i o n s h i p with that person. Mark your choice (number 1 through 5) i n the appropriate box a f t e r the name or i n i t i a l s of each person l i s t e d . A f t er you have completed A repeat the process with scale B and so on u n t i l a l l persons l i s t e d have been assessed on a l l scales. Above i s a sample i l l u s t r a t i n g what this might look l i k e when completed. 172 A. Indicate how often you have CONTACT (face-to-face, by phone or by l e t t e r ) with this person. 5. Usually d a i l y 4. Usually at least once a week 3. Usually at least once a month 2. Usually at least once every 6 months 1. Usually at least once a year B. Indicate the kind of FEELINGS and THOUGHTS you have toward this person: 5. Mostly very strong, p o s i t i v e feelings & thoughts 4. Mostly moderate, p o s i t i v e feelings & thoughts 3. About equally mixed po s i t i v e and negative 2. Mostly moderate, negative feelings & thoughts 1. Mostly very strong, negative feelings & thoughts C. Indicate the degree to which this person may help you by DOING THINGS when you may need i t , such as a s s i s t i n g on the job, helping with house-hold tasks, providing personal or family care, or even lending money: 5. Very frequently .4. Often 3. On some occasions 2. Rarely 1. Not at a l l D. Indicate the degree to which this person may help you by providing EMOTIONAL SUPPORT when you may need i t : 5. Very frequently 4. Often 3. On some occasions 2. Rarely 1. Not at a l l E. Indicate the kind of FEELINGS and THOUGHTS this person has toward you: 5. Mostly very strong, positive f e l l i n g s & thoughts 4. Mostly moderate, po s i t i v e feelings & thoughts 3. About equally mixed positive and negative 2. Mostly moderate, negative feelings a thoughts 1. Mostly very strong, negative feelings & thoughts Indicate the degree to which you may help this person by DOING THINGS when they may need i t , such as a s s i s t i n g on the job, helping with house-hold tasks, providing personal or family care, or even lending money: 5. Very frequently 4. Often 3. On some occasions 2. Rarely 1. Not at a l l Indicate the degree to which you may help this person by providing EMOTIONAL SUPPORT when they may need i t : 5. Very frequently 4. Often 3. On some occasions 2. Rarely 1. Not at a l l 174 Appendix £ Forms 176 • 1 GENERAL INFORMATION SHEET (Please answer a l l questions) Family Name Family Case #| | | |S.E.s| \ \ Address ' Phone Father's Name Age| | |present Ht. Occupation Education YRS | | I I 2 Mother's Name Age! I JPresent Ht. I I I Occupation Education YRS I 3| Daughter's Name Aqe| | IPresent Ht. I I I Wt. I 1 I I Occupation Education YRS | | | Indicate the number of c h i l d r e n in the family - Boys m G i r l s CD In which p o s i t i o n is the daughter who is taking part in the study? 1st. • 2 n d . n 3rd. • 4th. • 5th. [~~| Has any of the above family members ever dieted or used a weight control method? YES NO [ | If so, Who When Reason What was the lowest weight reached during the weight reduction program? L i s t any s t r e s s f u l s i t u a t i o n s which have taken place in the family during the past 4 years. These can include family moves from one geographic lo c a t i o n to another; marriages, deaths, separation or divorce of parents; brother or s i s t e r leaving home; f i n a n c i a l c r i s i s . L i s t any major i l l n e s s e s any of the above family members have had during the past four years for which they have sought medical advice. 177 SUBJECTS CONSENT FORM We agree to participate in a research project about family relat ion-ships and anorexia nervosa. We understand that participation in the study is voluntary, that we are free to withdraw at any time or refuse to answer any question, and that our involvement will in no way affect the treatment received from our attending physician. We understand we wil l be required to answer questionnaires which wil l take each of us approximately two to three hours of our time. We do this with the understanding that the information wil l be kept conf ident ia l , used for research purposes only, and destroyed at the end of its usefulness. F a t h e r Mother Daughter Date PARTICIPANTS CONSENT FORM We w i l l i n g l y g i v e o u r c o n s e n t t o h a v e a summary o f t h e i n f o r m a t i o n o b t a i n e d f r o m t h e q u e s t i o n n a i r e s on f a m i l y r e l a t i o n s h i p s f o r w a r d e d t o t h e a t t e n d i n g p h y s i c i a n . We u n d e r s t a n d t h i s i n f o r m a t i o n w i l l be h e l d i n s t r i c t c o n f i d e n c e by t h e p h y s i c i a n a n d u s e d f o r t h e p u r p o s e s o f f u r t h e r a s s e s s m e n t and t r e a t m e n t . F a t h e r M o t h e r D a u g h t e r O a t e 

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:
https://iiif.library.ubc.ca/presentation/dsp.831.1-0054198/manifest

Comment

Related Items