THESIS ‘ LIBRARY Michigan State University fi This is to certify that the dissertation entitled THE EFFECT OF PSYCHOTHERAPY ON THE SCHIZOPHRENIC' 5 HUMAN PERCEPT: A STUDY OF OBJECT REPRESENTATIONS IN SCHIZO - PHRENIA presented by DAVID LEE GREEN has been accepted towards fulfillment of the requirements for g’\ ./5’,.:.; P711/ degree 1“ I. 5/) C [/25 /,_ C/L/ Major professor Date / I /5"/7"4 MS U is an Affirmative Action/Equal Opportunity Institution 0-12771 MSU LIBRARIES W \. RETURNING MATERIALS: Place in book drop to remove this checkout from your record. flfi§§_will be charged if book is returned after the date stamped below. THE EFFECT OF PSYCHOTHERAPY ON THE SCHIZOPHRENIC'S HUMAN PERCEPT: A STUDY OF OBJECT REPRESENTATIONS IN SCHIZOPHRENIA BY David Lee Green A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Psychology 1984 ABSTRACT THE EFFECT OF PSYCHOTHERAPY ON THE SCHIZOPHRENIC'S HUMAN PERCEPT: A STUDY OF OBJECT REPRESENTATIONS IN SCHIZOPHRENIA BY David Lee Green Thirty three schizophrenic patients randomly assigned to three treatments, psychotherapy only, psychotherapy plus medication, and medication only, were assessed before treatment, at six months, at 12 months, and at 20 months as part of the Michigan State Psychotherapy Project. Research data were scored for level of object representation on the Krohn and Blatt scales. The Krohn and Blatt object representation scales were so highly correlated (r = .90) that it was concluded that they measure the same construct. It was hypothesized that the level of object representation should improve with psychotherapy. Of the three groups, only the group receiving both psychotherapy and medication showed an improvement in the articulation and organization of realistic object representations. Both this group and the medication-only group showed a decrease in the articulation and organization of unrealistic object representations. The differences between the three groups, however, fell short of statistical significance. Patients of experienced therapists showed more organization and articulation of realistic object representation than patients of inexperienced therapists. However, the object representation scales did not correlate with clinical status interview ratings of emotional health. Unlike a previous study, paranoid schizophrenics were not found to differ from non-paranoid schizophrenics in either level or change in object representation. ACKNOWLEDGMENTS In memory of my father, Eugene Green (1919-1983), I would like to thank two people whose support and criticism made this work possible, my Chairperson, Dr. Bertram Karon, and my wife, Dr. Carolyn Manson. Dr. Karon‘s clinical and research work has long been an inspiration to me. His sensitivity to the life of graduate students is no less remarkable. I would also like to thank my Ph. D. Committee. All three members, Drs. John Hurley, Norman Ables, and Bertram Stoffelmayr, were supportive and helpful throughout the writing of this work. More importantly, all three of these psychologists are excellent psychotherapy researchers, as I strive to become. ii TABLE OF CONTENTS INTRODUCTION ........................................ REVIEW OF THE LITERATURE ............................ Object Representations and the Rorschach ....... Psychodynamic Definitions of the Term ”Object Representation" ................... Object Representation Scales and Object Relations Theory ................... Therapy and Changes in Object Representations ........................... HYPOTHESES .......................................... METHOD .............................................. Selection of Subjects and Assignment to Groups ................................. Subject Attrition and Administrative Complications ............................. Clinical Status Interviews ..................... Psychological Testing .......................... Instruments .................................... Scoring Procedure .............................. units 0f AnaIYSiS 00......OOOOOOOOOOOOOOOOOOOOOO iii Page 10 13 22 33 36 36 38 39 41 41 44 45 TABLE OF CONTENTS (Continued) Page RESULTS .............................................. 49 Intercorrelation of the Object Representation Measure ..................... 49 Tests of the Hypotheses .................... 50 Hypotheses 1 and 3 .................... 50 F+ Data ............................... 53 F- Data ............................... 53 Pooled F+/F- Data for the Krohn Scale ...................... 54 Hypotheses 2 and 4 .................... 55 Between-Group Differences in Outcomes on the Blatt and Krohn Measures .................... 55 Pooled F+/F- Data for the Krohn Scale ...................... 57 Hypothesis 5 .......................... 58 Paired E Tests ........................ 59 F+ Data ............................... 60 F- Data ............................... 61 Hypotheses 6 Through 9 ................ 61 Hypothesis 10 ......................... 62 DISCUSSION ........................................... 64 The Effect of the Experience Level Of the Therapist 0......OOOOOOOOOOOOOOOOOOOO 66 iv TABLE OF CONTENTS (Continued) Page Object Relations Theory and Psychotherapy Research ..................... 67 Limitations of the Object Representation Scores in This Study ....................... 68 Support for the Validity of Psychoanalytic Theory of Change ............ 70 Process of Change: An Example ................... 72 CONCLUSION ........................................... 78 APPENDIX 1 ........................................... 79 APPENDIX 2 ........................................... 86 APPENDIX 3 ........................................... 112 APPENDIX 4 ........................................... 120 FOOTNOTES ............................................ 121 BIBLIOGRAPHY 0......OOOOOOOOOOOOOOOOOO0.00000000000000 122 Table Table Table Table Table Table 6 Table 7 Table 8 LIST OF TABLES Page Patient Characteristics of Final Sample (n = 33) ....................... 37 Means and Standard Deviations for each of the Blatt and Krohn Measures at Four Times ................. 51 Mean Scores on the Blatt and Krohn Scales at Intake Versus 20 Months Later for All Treatment Groups ................................ 52 ANCOVA Comparisons Between Psychotherapy and Non-Psychotherapy Groups .............. 56 ANCOVA Comparisons Between Experienced and Inexperienced Therapists OOOOOOOOOOOOOOOOOOOO00...... 59 Paired t Tests Comparing Object Representation Score at Intake and at 20 Months for Experienced and Inexperienced Therapists .......... 60 Mean Scores for Paranoid Patients on the Blatt and Krohn Scales ......... 62 Pearson Correlations Between Object Representation Score and Clinical Status Ratings OOOOOOOOOOOOOOOOOOOOOOOO 63 vi INTRODUCTION Research in the psychotherapy of schizophrenia is shrouded in pessimism. In a review of the literature, Hollon and Beck (1978) concluded that drugs alone are the most effective method of treating schizophrenics. Smith, Glass and Miller (1980), reviewing the same literature, concluded that drugs plus psychotherapy provided a more effective treatment than either psychotherapy only or drugs only. With a few exceptions (Karon and VandenBos, 1980), other authors reviewing the same literature concur in the view that psychotherapy adds only negligibly to the treatment of schizophrenia (Klein, Gittleman, Quitkin, & Rifkin, 1980; May, 1975). Recently, Gunderson and Gomes-Schwartz (1980) have looked at exploratory psychotherapy with schizophrenics from a more optimistic angle. These authors suggest that change occurring due to exploratory psychotherapy is seldom measured properly. The best way to measure change from exploratory psychotherapy, they suggest, is to evaluate improvement from the patient's vantage point. Gunderson and Comes—Schwartz (1980) argue in favor of the patient's perspective based on a preliminary study of exploratory psychotherapy with schizophrenics. They argue that outcome studies relying on traditional measures of improvement, i.e., discharge and recidivism rates, fail to document significant areas of change. Since these traditional measurements are taken from the perspective of third parties, they usually bear on sign and symptom change, and not on changes in the patient's experiences of self and interpersonal relationships. Moreover, they argue that the enhancement of life via increased self—understanding is the goal of exploratory psychotherapy: fulfillment of this goal needs to be measured more accurately with instruments that reflect the patient's perspective. The present study, using data from the Michigan State Psychotherapy Project (Karon and VandenBos, 1980) (one of the few studies reporting positive findings for the effect of psychotherapy on schizophrenics), extended Gunderson and Gomes-Schwartz‘ (1980) search for more appropriate outcome assessment in exploratory psychotherapy with schizophrenics, and encompassed also a more widespread concern: the concern for greater "specificity.” Regarding ”specificity,” Hans Strupp (1978)1 and other notable psychotherapy (Kiesler, 1967) researchers (Kiesler, 1967; APA Commission on Psychotherapy, 1982) have long called for a change in the focus of inquiry away from the global question, “Does psychotherapy work?" and towards more exacting questions such as, "Does this specific form of psychotherapy help this defined class of patients, under these circumscribed conditions, and most significantly, for these specified problems?" In psychotherapy studies of schizophrenia (Hollon and Beck, 1978), these questions regarding "specificity" have rarely been addressed. The present study used the suggestion of Strupp (1978) and of Gunderson and Gomes-Schwartz (1980), and addressed some of these questions as it assessed the effect of an exploratory type of treatment on the schizophrenic's experience of self and of others. In accord with Strupp, the specified form of treatment, a modified form of psychoanalytic therapy (Karon and VandenBos, 1981), which aimed at, among other things, helping schizophrenics differentiate their experience of self and others, was assessed in terms of a specific and appropriate outcome, the level of differentiation of self and others the patient shows on Rorschach object representation measures. The patient's object relations, and changes occurring in them, were assessed in this study through looking at the human percept on the Rorschach, before and after treatment. Evaluating the human percept is one way to look at changes in object representations on the Rorschach. According to theorists, the human percept reflects the patient's internal models or concepts of other people and self, and changes in the human percepts reflect changes in such models and concepts (Blatt and Lerner, 1983: Blatt, Wild, & Ritzler, 1976a: Lerner, 1983: Mayman, 1967: Urist, 1973). These changes in the human percepts were assessed both qualitatively and quantitatively. The Krohn scale, a qualitative measure, was used to assess the level of development of affective themes in the patient's perceptions of humans (Krohn and Mayman, 1974). The Blatt scale, a more quantitative measure, was used to assess the level of cognitive maturity in the patient's ”concept of the object” (Blatt et al., 1976a: Blatt, Brenneis, Schimek, and Glick, 1976b). The Krohn and Blatt scales were chosen to be used here from among a variety of recently developed Rorschach scales designed to assess the patient's developmental level of object relations (Blatt and Lerner, 1983; Kwarer, Lerner, Lerner, and Sugarman, 1980). In line with recent trends in psychoanalytic theory (Kernberg, 1975; Kernberg, 1976: Masterson, 1976; Meissner, 1981), many of these scales were designed to understand primitive and, especially, borderline personalities, their defenses, and their object representations. Paralleling developing trends in psychoanalytic writing (Schafer, 1968: Schafer, 1976), these scales attempt to capture on a more experiential level the patient's perception of self and of others. Along with other object representation scales (Kwarer et al., 1980), the Blatt and the Krohn measures have been applied to diagnostic, developmental, and psychotherapy outcome studies. Positive findings from these sources amply support the construct validity of ”object representations.“ The Krohn and the Blatt scales were used by Schwager and Spear (1981) to clarify some of the changes that exploratory psychotherapy can accomplish with schizophrenics. Using the test-retest records of 10 consecutively admitted schizophrenics treated intensively with psychoanalytically oriented therapy, they examined changes in object representations as well as changes in other ego functions. Their discussion of the contrast between changes they found on object representations and changes they found in less specific Rorschach measures of ego functioning sheds interesting light on how useful object representations can be. To assess general ego functioning, they used Schafer‘s (1955) Rorschach indices of impulse and affect modulation and also his measure of thought organization. Interestingly, these more traditional measures of ego functioning did not show evidence of patient change.2 According to these indices, the patients had not improved. When they examined the Blatt and Krohn scale measures, however, on similar intra- and inter-group comparisons, they did find evidence of patient change. Furthermore, their findings were consistent with the expectations of most object relations theories of change. The comparisons were between two diagnostically distinct groups of schizophrenics: paranoid versus non-paranoid or undifferentiated schizophrenics. The patients were tested within one month after admission and then again at discharge, 12 to 18 months after admission. At retest, on the Blatt measure, the paranoid patients showed significantly less object differentiation, but the non—paranoids showed greater differentiation than they showed earlier on the first testing. If one takes an object relations oriented vieWpoint, one assumes that paranoid patients, on a defensive level, have overly-defined boundaries between themselves and others (Meissner, 1978: Searles, 1962). One also assumes that effective treatment should, at least temporarily, loosen up these boundaries to allow dissociated feelings to be integrated (Spear and Hymowitz, 1981). Using these assumptions, the results of this testing may reflect a desirable therapeutic effect. The results from the non—paranoid schizophrenics in this study appear to reflect similarly desirable therapeutic change from the object relations viewpoint. These patients can be thought of as having severe problems in blurring and merging their experience of the boundaries between themselves and others (Searles, 1959). Diminished blurring of boundaries resulting from psychotherapy would seem to be reflected in these authors' report of increased differentiation on the Blatt scale after treatment. The changes found on the Krohn measure were also consonant with this interpretation of opposite yet desirable change for these two groups of schizophrenics. Data from the Michigan State Psychotherapy Project (Karon and VandenBos, 1970, 1972, 1981) was used in the present study to extend the findings of Schwager and Spear (1981). The present study investigated several interrelated hypotheses about what changes schizophrenics show in their Rorschach object representations after psychotherapy. Among many methodological weaknesses, Schwager and Spear's study had a very small sample, lacked comparative treatments, and was limited to analysis of outcome at two points in time (admission and discharge). The Michigan State Psychotherapy Project, a comparative treatment study of schizophrenia, provides more informative data because the methodology used in the study avoids many of the methodological problems in the Schwager and Spear study. The Michigan State study, for example, used repeated measures (four points in time), had comparative treatment groups, had a follow-up, had a larger total sample 5 (33), and assessed outcome on many more criteria. Just before closing this introduction, it is interesting to note that a previous outcome study used the Michigan State Project's Rorschach data to assess a very different constellation of ego functions. This study reported changes in ego functioning as would be predicted from the psychoanalytic theory of regression in treatment (Glatt, 1971). Although Glatt examined changes in impulse control and reality testing, he did not examine changes in object relations, which are often considered as measures of ego functioning (Lerner and Blatt, 1983: Kernberg, 1975, 1976; Jacobson, 1964). This is true even though object relations theory informed the treatment in the Michigan State study (Karon and VandenBos, 1981). In closing, the present study assessed change in the schizophrenic's Rorschach human percept as one facet of his object representations and related the findings to a theory of change consonant with object relations oriented theory. In addition, this study attempted to validate object relations theory of change by examining changes in object representations in treatment subgroups of schizophrenics: psychotherapy only (PO) versus medication only (MO) versus medication plus psychotherapy (MP), and paranoid versus non-paranoid patients. REVIEW OF THE LITERATURE Object Representations and the Rorschach Rapaport, Gill and Schafer (1945, 1946) began using the Rorschach to systematically evaluate changes in ego functioning resulting from psychoanalytically oriented treatment. The term "ego functioning” refers to the quality of formal thought patterns, the organization of the defenses, and the ability to modulate affects and impulses (Schafer, 1955), and also to other adaptive and regressive features of the ego as dictated by psychoanalytic ego psychology (Hartmann, 1939). Recently, researchers within the psychoanalytic tradition of Rorschach interpretation (Blatt and Lerner, 1983: Lerner, 1983: Mayman, 1967) have argued that traditional ego measures on the Rorschach contribute little to understanding the patient's subjective experience of his interpersonal world. Moreover, the traditional ego-function approach ignores the contributions made by the object relations oriented theorists, such as Fairbairn (1952), Guntrip (1969), Jacobson (1964), Klein (1952), and Winnicott (1958), towards a better understanding of severe psychopathology. 10 Lerner (1983) criticized traditional Rorschach ego assessment for being unsuitable to evaluate and enhance understanding of the patient's experience of self and others. He has called attention to how remote ego-assessment language is from the experience level of the patient. He remarked that traditional ego-assessment language ”is couched in a mechanistic, natural science framework of impersonal structures, forces, and energies . . .' (p. 314). His comments are in agreement with the opinions of many critics of psychoanalytic metapsychology, but most notably with those of Schafer (1977), who has forcefully argued that psychoanalysis needs to develop a theory and a language for human actions to replace its archaic metapsychology. In this light, it is relevant that a spate of recent studies have related Rorschach object representations to important aspects of the clinical situation: to the transference and to the patient's ability to enter a psychotherapeutic relationship (Blatt and Lerner, 1983). Psychodynamic Definitions of the Term ”Object Representation” Psychoanalytic theorists such as Beres and Joseph (1970), Novey (1958), and Sandler and Rosenblatt (1962) believed that object representations evolve from dynamic and largely unconscious processes. As such, they are the product of the same conflicts, drives, and affects that 11 influence ordinary cognitions, dreams, and mental symptoms. Yet, as somewhat enduring features of the psyche, they must be distinguished from memories and sensory experiences or other short—lived perceptual events such as images (Sandler and Rosenblatt, 1962). This viewpoint has been developed further by researchers who study object representations on the Rorschach. Hatcher and Krohn (1980) have described the "governing concept” in the developing trends of Rorschach research (those trends will be reviewed below) as follows: During development, especially the development of the ego by means of its relationship with others during the first five years of life, there develop a set of internal structures that reflect the individual's early experience of important others, structures that we may very roughly call mental images of people. These structures filter, select, and organize the experience of other people and the actions, thoughts, and feelings of the self. Thus, an individual's experience of others will only be as differentiated or varied as are the internal representations with which he can match them up. (Pp. 299-300) In an impressive theoretical paper on schizophrenics‘ object representations integrating findings from experimental, cognitive-developmental, and psychoanalytic 12 psychology, Blatt et al. (1975a) theorized that object representations originate around recurrent ”action-sequences” where primary caregivers either satisfy or frustrate fundamental emotional and physical needs. Blatt et al. (1976a) cite one line of empirical evidence for this in the form of research that relates frustrating care-giving relationships to the development of object constancy (in the Piagetian sense of the word 'object') and person constancy (again in the Piagetian sense, that is to say, recognizing a person as the same person after they have temporarily disappeared) (Bell, 1970: Provence and Lipton, 1962). This research reported that frustrating care-giving relationships in the first months of life were associated with person constancy developing after object constancy. He cited the same research to show that the reverse is true also: that gratifying care-giving relationships are associated with person constancy developing before object constancy. This seemed to suggest to Blatt et al. (1976a) that frustrating care-giving situations lead to mental or object representations in which persons are perceived in a characteristically less constant fashion than when gratifying care-giving experiences predominate. Blatt et al. (1975a) elaborated on the qualities of early object representations and compared them to the qualities of schizophrenic object representations. They described a perceptual-developmental process that is very similar to werner's (1948) developmental theory of 13 perceptual and cognitive functioning, and they then related this theory to their research on object representations in the Rorschach. In Werner's (1948) theory, the infant perceives the world in an amorphous, undifferentiated fashion, and the infant's mental representations tend to capture primarily the undifferentiated need-gratifying or need—frustrating qualities of the environment. Blatt et al. (1976b), Blatt and Lerner (1983), and Ritzler, Zambianco, Harder, and Kaskery (1980) showed that Rorschach human percepts that are organized along the lines of Werner's primitive and immature mentation are significantly related to diagnoses of regressed pathological states and to children--that is, to persons at a hypothetically low developmental level of human perception. According to their research, as the child matures, his developing mental representations involve increasing differentiation of self and of object. In adults, an absence of this differentiation is associated with immaturity and psychopathology. Object Representation Scales and Object Relations Theory Kernberg's (1976) review of the object relations literature summarized that increasing maturation and mental health is accompanied by an increasing ability to internally represent significant others as unique persons having independent identities, distinct interpersonal strivings, 14 and, most significantly, as individuals having unique interpersonal needs. Over the last decade, research on object representations on the Rorschach has generally supported Kernberg's summary. (This research is reviewed below.) In addition, the concordance of findings using different measures and different methods to assess object representations lends the measures themselves the promise of both convergent and discriminant validity, as outlined by Campbell and Fiske (1959). One such scale, the "Mutuality and Autonomy Scale" Urist, 1973, 1977), has provided construct support for the idea of object representations. Several studies (Urist, 1973: Urist and Shill, 1977) correlated Urist scale scores on Rorschach object representations with Urist scale scores on object representations taken from other data sources, such as autobiographical information and staff ratings of behavior. Urist (1973) developed the scale to measure the developmental level of the subjects' object relations. Raters using the scale assign a scale point to the interactions of objects (animal and human figures) on the Rorschach or to such interactions occurring in other sources of data. Subjects whose object representations reflect themes of magical control and coercion, for example, are hypothesized to be at low levels of object relations development, whereas individuals whose Rorschach object representations reflect reciprocity and mutuality are hypothesized to be at a much higher level. 15 In two studies (Urist, 1973; Urist and Shill, 1977), ratings taken from the Rorschach correlated with ratings taken from autobiographical material (r = .63), and also with ratings based on staff ratings of behavior (r = .53). In a retrospective psychotherapy study of 70 children who were former inpatients, the Mutuality and Autonomy Scale was correlated positively with treatment outcome. In this study, Teuber (1983) compared 35 Rorschachs of children who were later rehospitalized, with 35 Rorschachs of children whose records indicated that they had not received any further psychiatric services. Criteria for inclusion in the latter group was that the child had been discharged at least five years earlier from the residential treatment center. When he applied the Urist scale to the Rorschachs, Teuber (1983) found a significant positive association between high object representations scores and the avoidance of rehospitalization. He also found significant associations between low object representations scores and the need for future hospitalization. In a nearly parallel line of research, the Krohn (1974) “Object Representation Scale for Dreams” has been applied to the Rorschach. Like the Urist scale, the Krohn scale focuses on the thematic (that is, content) interpretation of object representations. For example, at level one, the human percept is filled with themes of voidness and lifelessness. Each level Krohn assumed 16 (Krohn and Mayman, 1974) represented a develOpmental advance in object representation. Raters apply these thematic levels to the human figures in the material being rated. The first study of the scale (Krohn and Mayman, 1974) demonstrated construct validity for "object representation“ due to significant intercorrelations between ratings taken from dreams and ratings taken from other measures: the Rorschach and early memories. Further indications that this scale supports the validity of Kernberg's (1976) summary (stated above) came from the significant correlations Krohn and Mayman found between scores on their scale and scores on another measure of mental health, the Luborsky Health Sickness Rating Scale (Krohn and Mayman, 1974). A correlate of Kernberg's (1976) summary is that different types of psychopathology are associated with differences in the thematic and affective qualities of internalized object representations. This tenet received partial support from a Rorschach study comparing obsessive/paranoid borderlines with hysterical/impulsive borderlines on the Krohn and Blatt measures of object representations. Spear (1978, 1981) studied 55 inpatients who were divided into groups according to DSM-ll criteria for schizophrenia, borderline, and character disorder. He further classified the patients into obsessive and hysterical character styles, along the lines described by Shapiro (1965) for classifying neurotic styles. After 17 the patients were sorted according to these criteria and their Rorschach object representations inspected, Spear reported, among other findings, significant differences on the Krohn scale between the obsessive and hysterical borderlines. The hysterical patients had higher scores on the Krohn scale than the Obsessives. Although these findings were somewhat weakened when an analysis of covariance controlling for the correlation between the Blatt and Krohn measures was computed, this initial finding deserves further exploration, as it coincides with theoretical progress in differentiating among borderline character pathologies (Kernberg, 1975). A less ambitious study employing the Krohn scale (Grey and Davies, 1981) examined the relationship between mental health ratings derived from a fairly descriptive interpersonal adjustment scale, the Midtown Mental Health Rating Scale, and object representation scores on the Krohn scale derived from the patient's dreams. The subjects in this study were 30 young adult females undergoing psychoanalytic psychotherapy at an urban psychoanalytic clinic. The correlation between ratings on these two scales was high (eta = .83). Given that the Midtown Mental Health scale has been shown in at least one longitudinal study to have predictive validity (Srole, Langer, Michael, Opler, and Rennie, 1962), the relationship between the object representation scores and scores on this symptom-adjustment oriented scale seem rather 18 impressive. Grey and Davies noted that object representations on the Krohn scale, even from as random a source as dreams, may indicate the general level of mental health. Their finding is buttressed by Hatcher's and Krohn's (1980) study of 25 outpatients' dream ratings on the Krohn scale. They found that for patients who were rated as neurotic on the Luborsky Health Sickness Scale (scores of 50 or above), their dream object representation scores correlated with therapist ratings of the patients' capacity for intensive psychotherapy. The latter rating, made by the therapist, pertained to the patients' propensities for acting out, tolerance of ambiguity in self, etc. The Krohn scale was also positively correlated with Global assessments of psychotherapy outcome by Frieswyk and Colson (1980). They applied the scale to the Rorschachs of 35 adults, who were part of the Menninger Treatment Evaluation .and Follow—Up Project (Wallerstein, 1968). In addition to the Krohn scale, the Blatt scale was used in the present study. This scale is applied to structural properties of the human percept on the Rorschach. Blatt et al. (1976b) created the scale to measure the cognitive-affective development of internal “templates" (representations) for object relations. Three structural properties of the human response are rated by the scale: (1) ”differentiation," a classification of the wholeness of the human or quasi-human percept; (2) "articulation," 19 a classification of the percept according to its perceptual attributes, such as its size or clothing, or else according to its functional attributes, such as its role or its sex; and (3) ”integration," the way the human percept is integrated into a context of action and interaction with other objects. These three structural aspects were derived from Werner's (1948) developmental cognitive theory of mental maturity. Blatt et al. found (1976a) that as normal individuals approach adulthood, they give responses in which human percepts are relatively more differentiated, more articulated, and more integrated. This was discovered in a retrospective 20-year longitudinal study of normal subjects. Each subject was tested four different times between the ages of 11 and 30. The authors summarized their findings by noting a marked increase with age in the number of accurately perceived, well-articulated, fully human figures involved in appropriately integrated actions, and in the relative improvements on each of these Wernerian dimensions with age. Blatt et al. (1976a) compared this group of normals to a group of inpatient adolescents diagnosed as having a thought disorder. The authors discovered an interesting pattern in the responses of the psychotic adolescents. On well-perceived human figures (F+), the disturbed adolescents tended to show greater malevolence, passivity, and incongruous activity. These responses were at a very low developmental level on 20 Blatt's scale. On poorly perceived human figures (F-), however, the pattern reversed itself. The more disturbed group gave significantly more developmentally advanced human responses; these responses reflected greater articulation and integration of the object than the normals showed, despite controls for the number of responses in this category. Blatt et al. (1976a) interpreted this to mean that the psychotic patients used higher ego abilities to articulate, differentiate, and integrate when describing fantasy material, i.e., poorly perceived (F-) areas of the blot, than when describing more realistic material (F+ areas of the blot). These findings were replicated by Ritzler et a1. (1980) in a similar study comparing various groups of adult psychiatric patients on the Blatt scale. In this study, Ritzler et al. (1980) further explored the psychotic object representation pattern by comparing schizophrenic to non-schizophrenic psychotics. This comparison revealed that the schizophrenics displayed more of the psychotic pattern found by Blatt et al. (1976a): that is, the schizophrenics showed higher developmental levels on inaccurately perceived human responses than non-schizophrenic psychotics showed on inaccurately perceived human responses. Lerner and Lerner (1982) used the Blatt scale in a comparative study of defensive structure in neurotic, borderline, and schiZOphrenic patients. These authors 21 found that borderline patients produced significantly more quasi—human and quasi—human detail responses than a neurotic group. In addition, when the borderline group was compared with the schizophrenic group, the results indicated that the borderline patients offered responses with a greater number of humans with more articulation, and more differentiation than the schizophrenics offered. These results support the developmental view of psychopathology, especially that of Kernberg (1976), which asserts that neurotics, borderlines, and schizophrenics have progressively less well-developed internal object representations. This study also showed that the human figure could be scored for borderline defenses such as “devaluation," 'idealization," “projective identification,” and "Splitting." Using just the human figure to evaluate these defenses, Lerner and Lerner discriminated borderlines, neurotics, and schizophrenics, each from the other. A case study reported by Lerner (1983) suggests that changes in object representation on the Blatt scale may result from psychotherapy. Lerner describes a teenage girl whose behavior was described as symbiotic, sometimes self-mutilating, given to acting out, and possibly brain damaged. On the Blatt scale, her human percepts were initially inaccurately perceived, poorly differentiated, and generally at low developmental levels. This is aptly illustrated in Lerner's report of her response to Card 3 22 of the Rorschach ”Two cartoon people--not really people but some form of life-—looks like a head and body, kind of deformed, though, [examiner's inquiry] didn't have two eyes and stuff.” Her human responses exhibited remarkably more developmentally advanced features on the Blatt scale following a course of psychotherapy centered on her attachment to early malevolent objects. Lerner attributed this change in perceiving the Rorschach human figures to her acceptance of feelings of loss and to relinquishing pathological elements of her earlier attachment. It is noteworthy that subsequent to treatment this patient no longer showed any signs of being brain damaged. Therapy and Changes in Object Representations How to help the schizophrenic experience himself and others more realistically has been the subject of numerous discussions (Blatt, 1980: Hoedmaker, 1967: Lidz, 1973, 1980: Searles, 1965). According to psychoanalytic developmental theorists, in order to experience himself and others more realistically, an intermediate stage must take place: the schizophrenic's object representations must alter to become increasingly realistic, cohesive, and continuous. But before we can understand how to alter the schizophrenic's object representations, we have to understand why the schizophrenic's object representations became distorted in the first place. 23 Many psychoanalytic theorists see these distortions as defensive. In order to be healthy, the infant with a yet unformed and vulnerable identity needs a parent with his or her own realistic and autonomous identity, with whom he can have a healthy symbiosis (Mahler, 1965: Searles, 1979). If the parent has his own regressive needs to fuse (Jacobson, 1964), then the infant has a harder time relinquishing his need to fuse in the maturation process. The infant also experiences terror of his parent's intrusion on his precarious boundaries (Blatt, 1980; Jacobson, 1964). Thus, the infant, without a mature ego to modulate his responses, remains vulnerable to feeling an impending disintegration of his self due to the threat of fusion, murderous rage at the intrusion, and threat of completely destroying the other or being destroyed by the other (Klein, 1952). In order to preserve the illusion of a satisfying symbiosis (goodness) and to reject the fusion, rage, destruction, etc. (badness) that is so threatening, the infant uses the primitive defenses of splitting, denial, projective identification, and idealization. These defenses become part of his way of encoding the world and are structured into his internal object representations (Fairbairn, 1952: Klein, 1952: Kernberg, 1975). From then on, no experience occurs without being filtered through this encoding process. Stated differently, no experience occurs without being filtered through the structures of his/her split and 24 idealized, internalized object representations (Lerner and Lerner, 1982.) Other psychoanalytically oriented authors write from a more interpersonal perspective about the psychotherapy of schizophrenia (Karon and VandenBos, 1981: Lidz, 1973, 1980: Sullivan, 1953, 1962). These authors emphasize the recurrent, life-long experience of aberrant relating in the schizophrenic's family. These authors report that the parents totally eclipse a more rational view of reality by continuously modeling irrational ways of perceiving and communication (Lidz, Fleck, & Cornelison, 1965; Lidz, 1973, 1980). These strange modes of communication reinforce the pre-schizophrenic's experience of living in a world of strange categories, irrational communication, and, consequently, a world where the experience of self is never allowed to mature: a world where the experience of self and others is alternately pathologically fused or else fragmented (Blatt, 1976b, 1980: Lidz, 1973). Despite differences in emphasis on the causes of underdeveloped inner objects, both the intrapsychic and interpersonal psychoanalytic theorists accept that changes in object representations can only result from change in the experience of actual object relations. Both the interpersonal and intrapsychic viewpoints emphasize the schizophrenic's precarious sense of individuality, and the need to live on fantasized interpretations of reality. 25 Writers from both orientations agree that the schizophrenic patient desperately needs a therapist upon whom he can depend to understand his confusion, desperation, and rage, and upon whom he can focus his symbiotic needs without the threat of abandonment or annihilation. If the schizophrenic patient comes to trust the therapist, and to feel safe sharing his precarious senses of identity, then the therapist becomes established as a benevolent and relatively undistorted inner object (Blatt, 1980: Lidz, 1980). The theoretical intricacy of how the therapeutic process modifies the internal world of object representations is the subject of highly abstract discussions of "internalization" (Schafer, 1968: Meissner, 1981). Meissner (1981) and Wachtel (1977) point out that psychoanalysis lacks a formal theory of learning. Without such a theory, therefore, it is difficult to construct uniform models for how one might modify object representations through psychotherapy. Hence, discussions about changing object representations must be cast within the framework of internalization theory. Very simply stated, internalization is the process whereby the psyche develops internal structures that control, regulate, and modify behaviors that were originally controlled by environmental forces (Meissner, 1981). Freud used the term ”internalization" primarily when theorizing about the structuralization of the psyche. Taken from his work Ab 26 Outline of Psycho-Analysis (1940), Freud's oft-quoted passage illustrates how he saw the formation of the superego (which is in a general sense an extremely important inner object representation): A portion of the external world has, at least partially, been an object and has instead, by identification, been taken into the ego and thus becomes an integral part of the internal world. This new psychical agency continues to carry on the functions which have hitherto been performed by the people . . . in the external world. (P. 205) Psychoanalytic theoreticians such as Meissner (1981), Schafer (1968), and Loewald (1962) have refined Freud's explanations for internalization and have applied psychoanalytic metapsychology to an understanding of the transformation of object representations. These authors generally agree that modification of the patient's internal representations result from the process of working through the patient's transference. In the psychotherapy of schizophrenia, a somewhat parallel process occurs (Karon and VandenBos, 1981). The schizophrenic endows the therapist with qualities from the patient's inner world. Distortions, projections, idealizations, and split-off feelings that can be seen in the patient's object representations become externalized in the way the patient perceives the therapist. The therapist examines, clarifies, and interprets these 27 distortions. If this has been done effectively, the patient internalizes a more realistic view of the therapist. Through a process of accretion, the therapist continues, so to speak, to help the patient form more secure and realistic boundaries between the patient's experience of self and others (Blatt, 1980). The qualities and the sort of integration that one might find changed in the internal representations of schizophrenics has not been systematically examined. One can surmise, however, what changes to expect, partially on the basis of developmental theory, partially on the basis of the Rorschach studies discussed above, and partially on the basis of the changes reported by therapists, such as Karon and VandenBos (1981), Milner (1969, Rosenfeld (1965), and Searles (1963, 1979), who have extensively treated schizophrenics. The psychodynamics behind omnipotent thinking and idealizing fantasy contribute substantially to the theory which guides psychotherapy with schizophrenics (Blatt, 1980; Karon and VandenBos, 1981: Lidz, 1980: Rosenfeld, 1965). The object relations of schizophrenics typically involve themes of omnipotent control, and themes of idealized and devalued selves and others. Usually these themes form around primitive and drive-laden images of significant others. Blatt et a1. (1980) note that themes of |'urgent orality or hypersexuality“ predominate in the object representations of schizophrenics. The press of these 28 drives, according to these authors, manifests specifically on the Rorschach in “fantasies of relationships based on preformed social stereotypes." One sees these fantasies not only in Rorschach object representations, but also in object representations gleaned from clinical work. Where these themes stem from is a question many authors address. Blatt et a1. (1980) attributed the stereotyped idealizations and images of omnipotent control to the schizophrenic's restitutive efforts. They found evidence of these restitutive efforts specifically in the poorly perceived portions of the blot (F—), where the human percept is at a higher developmental level than it is on the F+ or well-perceived portions of the blot (as described above). They see these findings as corroborative of Jacobson's (1964) and Rosenfield's (1965) clinical theories of restitution in schizophrenia. According to these theorists, the schizophrenic's delusions are frequently formed around fantasies of blissful reunion with omnipotent maternal figures. Indeed, as it is so often noted by object relations theorists (Klein, 1952: Rosenfeld, 1965), the idealized, split-off, and projected wishes for an all-powerful maternal savior are at the heart of the psychotic transference. Blatt et al. suggest that it is in the F— portion of the Rorschach human percept that schizophrenics manifest these active efforts towards restitution just as Jacobson (1964), Klein (1952), and Rosenfeld (1965) describe. 29 All of these authorities observe that the therapeutic relationship necessarily entails working through and clarifying distortions based on these themes of omnipotence and idealization. Karon and VandenBos (1981) maintain that schizophrenics need to see the therapist as ominpotent in order for treatment to begin, but that when treatment is progressing, the therapist must slowly become an equal in the eyes of the patient. Lidz (1980) suggests, however, that the therapist should from the very start emphasize his limitations and separateness. Blatt et a1. (1980) add to Karon's view and say that the idyllic, blissful, and simplistic fantasies play a vital role in the schizophrenic‘s sense of security. As suggested by the Rorschach findings, these fantasies are constructed with the schizophrenic's highest cognitive abilities to articulate, integrate, and experience. According to Blatt et a1. (1980), the vicissitudes of the transference in regard to the patient's idealized images of the therapist are noteworthy. At first, the patient hides from the therapist his idealized images. He does not yet trust the therapist and therefore has greater need to sequester his restitutive fantasies of an all-powerful savior from the yet untrusted and threatening therapist. Eventually, the schizophrenic patient begins to trust the therapist more, and thereby the patient projects onto the therapist his island of security. The transference begins to reflect the patient's 30 idealized images. Later, in treatment, the therapist confronts the unrealistically idealized transference and helps the patient substitute a more real differentiated and trusting relationship for his previously magical thinking. The therapist's efforts are aimed, at this point in the therapy, at working through the idealized and omnipotent distortions underlying the transference. The central task becomes mourning the painful memories of earlier maternal empathic failures and relinquishing the restitutive efforts made in response to these failures. Until this time in treatment, these empathic failures have been warded off from awareness by means of the primitive distortions, most of which can be seen in the schizophrenic's delusions. It was useful in this study to surmize that when treatment was successful, based on the above, object representations of idealized figures (distorted F- percepts) would, theoretically, lose their prominence. Interestingly, in accord with Blatt's description of successful treatment, we theorized that the Rorschach would reflect these changes throughout. Specifically, the Rorschach human percepts would show a reversal in the ”psychotic pattern" observed by Blatt et al. (1976b) and by Ritzler et a1. (1980). That is, ironically, the developmental level of the human figure on poorly perceived portions of the blott would lose some of its more advanced features. As the patient had less investment in escape 31 from reality and in restitution for very painful reality, poorly perceived human figures would decrease in numbers and in sophistication (on the Blatt scale). Schizophrenics, perhaps more than all other groups of patients, have been described as lacking in a basic feeling of humanness (Karon and VandenBos, 1981; Searles, 1979). Searles (1979) and Milner (1969) illustrated this through case histories of schizophrenics undergoing intensive long-term psychotherapy. They describe how amorphous and inhuman the schizophrenic's experience of his body is at the start of treatment. When treatment is successful, the schizophrenic, according to Searles, feels less and less like an alien or a part-human and identifies less with the nonhuman world. When treatment is successful, the patient's problems become more human, that is, more neurotic than otherwise (Karon and VandenBos, 1981). On the Rorschach, this would translate into greater articulation and integration of the well-differentiated human percept (F+). This improvement from psychotic concerns to neurotic concerns would also be reflected in higher developmental interpersonal themes on the Krohn scale. In the present study, the diagnosis for schizophrenia was based on psychodynamic considerations, as described above. Thus, the "schizophrenics" treated in this study showed a chronic history of impaired interpersonal relations, and were in many other ways typified by 32 internal fragmentation and thought disorder, which has been described so vividly by Searles (1965) and by other writers from the interpersonal school of dynamic psychiatry. From this perspective, schizophrenia is primarily a social ”illness“ rather than a medical or primarily biological disease. HYPOTHESES Schizophrenics treated with psychotherapy and compared pre- and post-treatment will show: A) A significant increase on the Blatt developmental scale ratings on well-formed human percepts (F+). B) Significant decreases on the Blatt developmental scale ratings on poorly formed human percepts (F-). Schizophrenics treated with psychotherapy and compared pre- and post-treatment with schizophrenics who were not treated with psychotherapy will show: A) Significantly greater increased scores on the Blatt developmental scale ratings on well-formed human percepts (F+). B) Significantly greater decreases in the Blatt developmental scale ratings on poorly formed human percepts (F-). Schizophrenics treated with psychotherapy and compared pre- and post-treatment will show significant increases on the Krohn scale, that is, the affective and thematic ratings on the human percept will indicate a higher developmental quality. 33 34 Schizophrenics treated with psychotherapy and compared pre- and post-treatment with schizophrenics who were not treated with psychotherapy will show significantly greater increases on the Krohn scale, that is, the affective and thematic ratings on the human percept will indicate a higher developmental quality. The experience level of the therapist will be associated with changes in object representation scores for F+ and F— data. Paranoid schizOphrenics treated with psychotherapy will show a significant tendency for their scores on the Blatt developmental scale to fall during the middle stage of therapy and then to improve over their original level at follow-up. Paranoid schizophrenics treated with psychotherapy will show a significant tendency for their scores on the Krohn scale to fall during the middle stage of therapy and then to improve over their original level at follow-up. Non-paranoid schiZOphrenics treated with psychotherapy will show a significant linear trend toward increased scores on the Blatt developmental scale over time. Non-paranoid schizophrenics treated with psychotherapy will show a significant linear trend toward increased scores on the Krohn scale over time. 10. 35 Object representation scores on F+ percepts will be linearly correlated in a positive direction with clinical status ratings at each of the testing intervals. METHOD Selection of Subjects and Assignment to Groups The sample of 33 schizophrenics whose Rorschachs were analyzed as part of the Michigan State Psychotherapy Project (Karon and O'Grady, 1969) served as subjects in the present study. These participants were selected from among the first admission inpatients at the Detroit Psychiatric Institute (DPI), an institution servicing a primarily poor, inner city, black clientele. Table 1 shows the patient characteristics of the schizophrenics used in the present study. These patients were selected as follows: Preliminary case histories and full medical evaluations were completed before treatment group assignment on approximately 35 patients per week. Each week during a four-month period, three patients with similar pathology were selected and then randomly assigned to one of three treatment groups: psychotherapy alone, psychotherapy with medication, or the.hospital comparison group. Research assistants (those not administering treatment) assigned patients to groups and selected the participants according to the following criteria: (a) unquestionable schizophrenia, (b) onset of psychotic symptoms within 36 37 .conmHEHoa an coucHuomm 585.2 comma E $2 29888 .8ch2 :83. are» 362 .32 .momcmecns .m .o can Gonna .m .m >2 0396 m0 pagoda. one uanmHnHHONEom mo ammuguocowmm 59E o H o o o H mNma +mv N o o o o c was N4 B N H m H o H H mama mN 8 H N N H o H o mama I 3 o a mH m e e H mcoz Hoofloum mo new um czocu mm :oHumuHHmuHamom msoH>mnm o o o o o o +ONH o H o o H o ONH Cu oHH m m m H o H OHH Cu om n o o N N N om Cu om N m m H m o om 0u o asHHmolwaofiofiv 0H HouoHoeooch H H H o o o NuHmuo>Hca HouoHoeoov H. m N N m H H88 :91: 13398:: m a N N o N H88 :9: H m m o N o m Cu m mocmno H H o o H o m 0u H mocmuw coHunUSDQ m m m H o H oquz m oH m m m N xomHm comm I HH N m v N oHnEmm m 0H m H N H mez xmm cowHummEou mHoucmEHuomxm mmocHoHB HomH>HoQJm moocanH HomH>qu5m oHumHumuonunzu H3331 888 m @295 < 90.8 Amm u no oHoEmm Hmch wo moHumHuouomumnu ucoHumm H OHQNH 38 three months of admission, (c) no signs of organic contributions to the psychoses, (d) no history of alcohol or drug addiction, (e) no history of electroshock therapy, and (f) no prior psychiatric admissions. The principal investigators (Karon and VandenBos, 1981) noted that although these selection criteria were strictly applied, many of the subjects had even more serious pathology than detected at the time of subject selection. As treatment progressed, it became apparent that many of the subjects had been psychotic far longer than the three-month criteria. Indeed, treatment revealed that one-third of the patients had prior psychiatric hospitalizations and three patients had organic complications, as discussed below. (For a more detailed description of subject selection see Karon and VandenBos, 1981, pp. 385-390). Subject Attrition and Administrative Complications As reported in Karon and VandenBos (1981), two out of what was originally 36 schizophrenic subjects died, both from undiagnozed embolisms. One, patient, whose death occurred before randomization, was replaced. The other patient, a member of the non-medication group (Group A) died after therapy began and was not replaced; replacing her would have contaminated the randomization procedure. In fact, as aspects of her history came to light, it appeared that she would have been disqualified for three additional 39 reasons: (a) she had been addicted to heroin, (b) she had a history of ECT treatments, and (c) she abused alcohol. A third patient, who had an impaired gait and other motor abnormalities, was later diagnosed as having multiple sclerosis (MS). Since her MS symptoms seemed to follow a course separate from her schizophrenic symptoms, the principal investigators included her data in the final analysis. (The authors report that analyses excluding her data did not materially alter the findings.) A fourth patient, who complained of visual hallucinations of colored animals, was dropped from the data analysis when his therapist discovered that he had been abusing brain damaging quantities of seconal, dexedrine, benzedrine, nutmeg, and other substances. Finally, one other patient's data was excluded from the final analysis because staff interference in his treatment resulted in his being transferred to another hospital. (For a more detailed explanation of subject attrition, see Karon and VandenBos, 1981, PP. 385-390.) Clinical Status Interviews A psychoanalytically trained psychiatrist interviewed each of the patients at four intervals corresponding to the intervals set up for psychological testing. This psychiatrist was blind to the treatments given to the patients. All psychiatric interviews were tape-recorded and later rated by two graduate students in clinical 40 psychology. Both students were blind to the treatment administered to the patients. (In addition, portions of the tape had been deleted if they contained references to the type of treatment the patients received.) These graduate students were instructed to integrate 11 criteria into an overall mental health rating. The criteria employed were: (a) ability to take care of self, (b) ability to work, (c) sexual adjustment, (d) social adjustment, (e) absence of hallucinations and delusions, (f) relative freedom from anxiety and depression, (9) amount of affect, (h) variety and spontaneity of affect, (i) satisfaction with life and self, (j) achievement of capabilities, and (k) benign rather than malevolent effect on others. To integrate these criteria and to arrive at an overall mental health rating, the graduate students employed a scaling technique described in a separate paper by Karon and O'Grady (1970). Interjudge reliability between the clinical status raters was .82. Internal consistency as measured by intrajudge reliability for the two raters was .83 and .37, respectively. The predictive validity of this Clinical Status rating is indicated by the negative correlation found between the number of days of hospitalization during the six months following each Clinical Status Interview and the rating itself (-.71 for T1 and -.64 for T2), even though this interview was not part of the data available to the ward staff. (See 41 Appendix 1 for a copy of the Karon and O'Grady paper describing the rating procedure used for the clinical status of each patient.) Psychological Testing All patients were tested as close to the time of admission as possible (usually one to three weeks after admission), just prior to being randomly assigned to one of the three treatment groups. The Rorschach, the TAT, and several tests of intelligence were administered on the same day to each subject. All subjects were tested four times: first, within a few weeks of admission (Tl), second, six months later (T2), third, 12 months later (T3), and fourth, 20 months later (T4). Using standardized conditions, the same doctoral student administered all of the Rorschachs except for those given at T4, at which time another doctoral student in clinical psychology administered most of the remaining Rorschachs. Instruments In addition to the clinical status ratings, the central data in the present study included the Rorschach human percepts and the diagnosis of "paranoid” or ”non—paranoid.” These diagnoses were based on the hospital records for the subjects. For four subjects, however, the hospital records gave insufficient information about the diagnosis. For these four cases, the present 42 investigator and a clinical psychologist (Scott Haas, a licensed clinical psychologist experienced in diagnosing schizophrenia), independently diagnosed the subjects after listening to the taped interviews, which were also used for the initial clinical status ratings by the previous investigators. A diagnosis was then assigned based on the judgment of the two raters. As stated above, the human percept was scored and analyzed according to its structural and affective/ thematic features. To assess the former, the present investigator employed the Blatt scale (described in the theoretical context above). The Blatt scale assesses the maturation of the concept of the object and is based largely on the comparative developmental psychology of Werner (1948). The Blatt et al. (1975) scoring system rests upon Werner's developmental/comparative psychological principles of ”differentiation,” ”articulation,” and ”integration." "Differentiation” in the Blatt system refers to how complete a human percept is selected. The possible percepts that a subject might select on this dimension range from the most differentiated percept, the whole human response, to the least differentiated percept, the quasi-human detail. ”Articulation“ refers to the subject's elaboration of attributes describing a human or quasi-human response. Blatt et a1. (1975) scored two types of articulation: "perceptual” and "functional.” The former criterion 43 (perceptual) refers to such attributes as the size, clothing/hair style, and posture articulating the percept. "Functional" articulation refers to the sex, age, role, or specific identity ascribed to the human percept. Finally, the dimension of “integration“ is scored for three types of human activity characterizing the human percept, ”motivation of action,“ “object-action integration,“ and ”integration of interaction with another object.“ Each of these three types of activity is indexed according to specified developmental features evident in the activity. (See Appendix 2 for a detailed description of the scoring criteria for these types of hwman activity, along with a copy of the complete Blatt scoring system, attached.) In brief, scoring involves the use of an ordinal system of points for each of Blatt's three dimensions. Developmentally advanced features are placed at a higher scale value than are less developmentally advanced features. For example, a whole human response is scored 4 on the dimension of ”differentiation," while a quasi-human detail on the same dimension is scored 1. Previous research has found interrater reliabilities on all of Blatt's scoring criterion to be in the range of 71-91% agreement. (Blatt et al., 1976a: Ritzler et a1, 1980: Spear, 1978). At the time of the present study, a number of studies support the validity of this scale, as reviewed above. To assess the affective/thematic aspect of the human percept, the present study also employed the Krohn “Object 44 Representation Scale of Dreams.“ As discussed above, this scale has been successfully applied to Rorschach human percepts in a variety of clinical studies. The scale assesses the developmental level of the Rorschach representations according to a seven-point scale. Each scale point represents a deve10pmental level of object representation. (See Appendix 3 for descriptions of the scale points, along with two sample dreams. Reliabilities for the Krohn scale as applied to Rorschach was reported by Spear (1978) to be 83.1% for exact agreement by two judges using Krohn's seven-point scale. ScoringProcedure After the present investigator underlined the Human percept responses in the test protocols, a clinical psychologist (Scott Haas), along with.an advanced graduate student in clinical psychology (Judy Larsen), rated the human percepts. Reliability was established on both the Krohn and Blatt scale after training sessions were completed. Training sessions utilized schizophrenic Rorschachs from a different source. After the raters were sufficiently practiced such that interrater reliability reached the level of previously reported estimates (about 70%), the primary rater (Judy Larsen) independently scored all of the Rorschach protocols. The total number of human percepts scored in the sample was 262. To help preserve blindness, not only of group, but also of order 45 of administration, protocols were assigned randomly to both judges. Scott Haas (the secondary rater) scored two-thirds of the data for a reliability check. Interrater reliability was substantial on the Blatt scale, r = .90, p < .001, and moderate on the Krohn scale, r = .78, p < .01. Previous investigators had reported that interrater reliabilities were over 71% agreement for both the Blatt and Krohn scale (Blatt et al. 1976: Krohn and Mayman, 1974; Spear, 1978). Units of Analysis The principal unit of analysis was the arithmetical mean of each individual's score on the Blatt and Krohn Rorschach scales (for object representations or human percepts). According to the Krohn and Blatt scales, the lowest score that can be assigend for any human percept is one. Scores of zero were assigned to subjects who gave no human percept. The rationale was that the absence of human responses occurs more frequently in disturbed persons (Exner, 1974). Since human percepts are weighted on the Blatt and Krohn scale, and since there is evidence that the weights are linearly and positively associated with global ratings of mental health (Krohn and Mayman, 1974: Blatt et al., 1976), as previously described, it seemed reasonable to weight the absence of human percepts at the lowest end of the scale. 46 On both the Blatt and Krohn ratings, each individual subject's score for object representations was derived by summing the total object representation score on a given Rorschach protocol and then dividing this summed score by the number of human percepts on the protocol. This calculation gives the mean object representation score; this was the principal unit of analysis. Mean object representation scores on the Blatt and Krohn scales were also the main units of analysis in the Spear and Lapidus (1981) investigation and distinguished among various subgroups of borderline and schizophrenic patients. An effort was made to estimate how closely the mean scores in the present sample matched those of other schizophrenic samples, specifically, the sample from Spear and Lapidus's investigation. One 5 test for independent samples was performed on the scores from the Blatt scale and a similar 3 test was conducted on the scores from the Krohn scale. The mean Blatt score for the present sample (5 = 7.54) and the mean Blatt score for the Spear sample (M = 6.12) were not significantly different, 5(46) = 1.04, p<§ .25. The mean Krohn scores for the present sample (M = 3.26) and the mean Krohn score for the Spear sample (5 = 3.31) were also not significantly different, 5(46) = .08, Efl< .25. These 5 tests were performed by using the statistics (mean, standard deviation, and n) provided by Spear and Lapidus. Taken together, the results of these 47 .E tests suggest comparability of the present sample of schizophrenics with the Spear and Lapidus sample of schizophrenics, on these object representations scales. In addition, it was thought that a mean or average object representation score would be the most appropriate control for response productivity. The latter refers to the biasing effect of those subjects who give an unusually high frequency of human percepts. To determine whether using the mean scores successfully controlled for response productivity, Pearson correlations were computed to test the degree of association between the Blatt mean scores at intake and at 20 months with the total number of responses to the Rorschach at intake and at 20 months for each subject. These correlations were not significant, r(31) = .24, p < .15 intake, r(31) = .18, p < .25, 20 months, suggesting that the mean score provided an adequate control for response productivity. (If this correlation had been significant, then group analyses would have required the use of ANCOVA, with the number of responses as a covariate). Since the majority of the analyses in this study entailed only F+ data, both of these correlations were then computed on F+ data. For the Blatt F+ data at intake, the Pearson product moment correlation was not significant, r(31) = .19, p <, .25: for the Blatt F+ data at 20 months, the Pearson product moment correlation also was not significant, r(31) = .21, 2 <1 .25. These same correlation 48 coefficients were also computed for the Krohn scores and response productivity at intake and at 20 months. For the Krohn scale at intake, this correlation between response productivity and Krohn scores was not significant, r(31) = .22, p <. .20. Similarly, for the Krohn scale at 20 months, this correlation was not significant, r(31) = .27, 2 <1 .15. The same pattern of nonsignificant results also held for Krohn F+ intake data, r(31) = .25 p < .15, and at 20 months, r(31) =.18, p< .25. Meaningful correlations between response productivity and object representation scores could not be derived from the F- data because the ratings on F- data contained too many zeros to permit a meaningful Pearson product moment correlation (over 50% of subjects did not give human responses that were scored F-). RESULTS Intercorrelation of the Object Representation Measure The data in this study was divided into two categories of response: (a) accurately perceived human responses, which were scored F+, and (b) inaccurately perceived human responses, which were scored F-. Each was analyzed accordingly using the Krohn and Blatt scales. The two scales correlated highly on both the F- data, r=.88, p < .001, and the F+ data, r=.92, p <; .001. As discussed above, the Krohn and Blatt scales were originally designed to measure different aspects of object representations (Krohn 8 Mayman, 1974: Spear & Lapidus, 1981.) Their high correlations observed here, however, suggest that the two scales largely actually measured a single construct when applied to these Rorschach data. After these unexpected correlations between the Blatt and Krohn measures were found, it seemed reasonable that all predictions made separately on the Blatt ratings also should apply to the Krohn ratings. Thus, additional predictions identical to those stated for Blatt's measure but based on the Krohn scale were added to the study. (Prior to the discovery of their high correlation, all predictions 49 50 concerning the Krohn ratings were identical in every way to the predictions made on the Blatt ratings, except for one important aspect: the predictions for Blatt ratings distinguished between predicted patterns of response for F- data and predicted patterns of response for F+ data, but the predictions for Krohn ratings did not distinguish between predicted patterns of response for F+ data and predicted patterns of response for F- data. The discovery of such high correlations between the scales obviated the need to sharply distinguish the predictions according to which object representation rating scale was being used. Tests of the Hypotheses Hypotheses l and 3: Schizophrenics treated with psychotherapy and compared pre- and post-treatment will show significant increases in object representations scores for accurately perceived (F+) human percepts, and significant decreases in object representation scores in inaccurately perceived (F-) human percepts. To test Hypotheses l and 3, paired 5 tests were used to compare initial evaluation data with outcome data taken 20 months later. Table 2 shows the three treatment groups' means and standard deviations on the two (F+ and F-) measures for each of the Blatt and Krohn measures at the four times of testing. Table 3 presents the results of paired E tests of the difference between intake evaluations and 20-month 51 .ozouo coHuooHomaimsHmizamumzuozomma may monoconmo 32m: .osouo mHCOIcoHumoHomE may moumcmHmoo =02: Hum.Hv Avo.Nv HNH.HV How.Hv m>.o mh.H om.o NH.N Hmm.ov HNo.ov Hum.ov Hon.HV NH.o mN.o mH.o w¢.H Hem.HV Hmm.ov Hw¢.Hv Hmo.Nv om.H mm.o hH.H HH.H um cronx 8a.: SEN. 30.3 898 om.N oH.N mN.m om.N Hem.Nv Hom.Nv ANN.NV Hmo.Nv wm.m wH.m Hm.m HN.N H313 GENO $5.8 893 mH.m mH.m mm.N no.m +m ccoux oN NH m o mun—”H. mosaa.usom um monsoon: acoux can uumHm 2.» NO 208 HOW mCOwHMMNVa UHMUCBW UCM 209‘ «asoum >H¢01momumcuoco>mo on» moumconoo com: “ouoz 88.3 13.8 8E3 Roi mm EN EN NEH RR 2 HNH u 5 oz Home 15.: 18.: 3.2 mm N5 $5 $5 3.... z HNH u 5 E SNi 64.3 5.2 13.: NW NN.N NEH 2am EN 2 a n 5 8 ....H HumHm osouo 3N5 8TB HHHHV SNB elm. Sam 2;. NH $8 2 NH u 5 oz 18.8 1N3: 18.3 1N3: elm 85 33 HS Nme 2 NH u 5 so ENE 35 5.: 13.8 W NH; EN EN HH.HW z a u 5 8 E 33m ozone ON NH 8 o EH82 95H. N Home 52 Table 3 Mean Scores on the Blatt and Krohn Scales at Intake Versus 20 Months Later for All Treatment Groups Time Group 2_ 0 mo. 20 mo. 5 Blatt F+ PM 12 6.52 8.00 + .21 PO plus PM 21 7.20 7.75 + .44 MO 12 6.49 5.76 - .39 Krohn F+ PO 9 3.07 3.15 + .14 PM 12 2.21 3.58 +2.52* PO plus PM 21 2.58 3.40 +2.03 MO 12 2.54 2.80 + .32 Blatt F- PO 9 2.33 2.22 - .16 PM 12 3.19 0.25 -3.04* MO 12 5.33 2.10 -2.63* Krohn F- PO 9 .11 1.34 + .30 PM 12 1048 0.17 -2047* PO plus PM 21 1.32 0.67 -l.42 M0 12 2012 0079 -2047* Note: All 3 tests are two-tailed comparisons. .05. 53 evaluations for the Blatt and Krohn data. Table 3 presents the data for the two psychotherapy groups combined (the hypotheses were originally formulated with respect to all psychotherapy patients). (Herein, as shown in Tables 2 and 3, the three treatment groups will be referred to as follows: The group that received only psychotherapy will be designated by ll'PO," the group that received medication plus psychotherapy will be designated by "PM," and the group that received medication only will be designated by 'MO.") F+ Data: For the Blatt F+ data, Table 3 reveals that none of the groups significantly increased their mean scores in the intake versus 20-month E-test comparisons. The pattern on the Krohn F+ data is identical, with one exception: the PM group showed a significantly greater Krohn F+ mean at 20 months (M = 3.58) over their Krohn F+ mean at intake (M_= 2.21),‘£(ll) = +2.52, p,<< .05. F— Data: Table 3 shows that the PO group did not significantly decrease their scores on either the Blatt or Krohn F- data. In contrast, both groups that received medication did reduce their scores on F- data. For the PM group, the Blatt F- mean at 20 months (M = 0.25) was significantly less than the Blatt F- mean at intake (M = 3.19) £(ll) = -3.04, p <§ .05; the Krohn F- mean at 20 months (M = 0.17) was significantly less than the Krohn F- mean at intake (M = 1.48), £(ll) =-2.47, p <..05. A similar pattern was obtained for the MO group: the Blatt F- mean at 20 months (M = 2.10) was significantly less 54 than the Blatt F- mean at intake (M = 5.33), £(ll) = —2.63, p < .05; the Krohn F- mean at 20 months (0.79) was significantly less than the Krohn F- mean at intake (M|- 2.12), 3(11) = -2.47, p < .05. It is clear from these findings that psychotherapy was not necessary in reducing F- scores. Although the pooled PO plus PM group showed a significant decrease on Krohn F- scores, an inspection of the Krohn F- means for the PO group reveals that this latter group made a negligible contribution to the pooled psychotherapy results. That is, the decrease was almost entirely due to the PM group. Pooled F+/F- data for the Krohn scale: In order to test the unrevised predictions that psychotherapy patients (Groups PO and PM) would increase their scores on the Krohn scale (no distinction on F+ and F- data), paired £_tests compared intake data with the data at 20 months. For the pooled psychotherapy sample (PO combined with PM, n = 21), the results were not significant, 5(20) = +1.27, for the PO group, the results were not significant, 2(8) = .85; however, for the PM group, the results were significant and in the predicted direction. For group PM, the mean at intake (M = 2.4) was significantly less than the mean at 20 months (M,= 3.5),'£(11) = +1.94, 2 ‘< .05. These findings are the same as those on the F+ Krohn responses considered separately. (This is not surprising because there were many more F+ than F- responses.) 55 Hypotheses 2 and 4: Schizophrenics treated with psychotherapy and compared pre— and post—treatment with schizophrenics who were not treated with psychotherapy will show: significantly higher scores on well—formed human percepts (F+) and significantly lower object representation scores on poorly perceived human percepts. Between-Group Differences in Outcomes on the Blatt and Krohn Measures Four separate 1 x 2 analyses of covariance (ANCOVA) were conducted to directly test group differences in treatment effectiveness. Since there were significant linear relationships between prescores (0 months) and postscores (20 months), prescores were used as the covariate and post scores were used as dependent variables. These analyses compared outcomes obtained on the four outcome measures (Blatt F+, Blatt F-, Krohn F+, Krohn F-) by the pooled psychotherapy sample. Table 4 shows that none of these analyses of covariance reached statistical significance. Thus, Hypotheses 2 and 4 were not supported. The previous discovery using paired 5 tests of the MO group's significant reduction in the object representation scores on poorly perceived (F-) human percepts raised the question of how potent only medication might be. To better estimate the effect of medication on F- object representation ANCOVA Comparisons Between Psychotherapy Variable Blatt F+ Krohn F+ Blatt F- Krohn F- 56 Table 4 and Non-Psychotherapy Groups Pooled Psychotherapy (n = 21) N011- Psychotherapy gf (n = 12) 20 mo. Mean* 7.75 20 Mo. Mean 5.76 l 2.81 1 2.10 1 .791 1 *Table means were not adjusted for intake object representation score. .952 .83 .550 .366 scores, a three-group ANCOVA was used to simultaneously compare the MO group with the PM group with the PO group. As in the above ANCOVAs, the scores at intake were used as the covariate for the scores at 20 months. The results of this three-group ANCOVA were not significant for the Blatt F- scores, _F_(2) = 1.85, p < .17, but the results approached significance for the Krohn F- scores, {(2) = 2.69, p, < tests on the three groups were not significant at the .05 level, further between-group comparisons are not reported. .08. Because the results of these overall E The same three-group ANCOVAs were performed for F+ object representation scores. These ANCOVAs were not US ns NS NS 57 significant and the p values for the 5 tests did not indicate that there might be a trend favoring a particular group, 2(2) = .87, p < .43-Krohn: and {(2) = .63, ‘2 <1 .54 Blatt. Since these 2 values for these three-group E tests did not approach significance, separate two-group comparisons on F+ scores are not reported. Pooled F+/F- data for the Krohn scale: To test the original predictions on the Krohn scale that psychotherapy patients would show greater increases on the Krohn scale than non-psychotherapy patients (with no distinctions between F+ and F- data), the means at 20 months for the two groups were compared. The original prediction stated that on the Krohn scale (F+ and F- data pooled), psychotherapy subjects (groups PO combined with PM, 2 = 21) would show greater increases than non-psychotherapy subjects (group MO, 2 = 12). An inspection of the 20-month means in Table 4 revealed that this comparison could not be significant. The psychotherapy group (P0 combined with PM) mean (3.40) for the Krohn F+ data was greater than non-psychotherapy (MO) group mean (2.81), however, for the Krohn F- the non-psychotherapy group mean (.791) is greater than the psychotherapy (PO combined with PM) group mean (.67). Given that the differences between the group means for the psychotherapy and non-psychotherapy groups are in opposite directions (as seen above for the Krohn F+ data and the Krohn F- data, 58 respectively), pooling the F+ and F- means would not result in a greater difference between the groups. Hypothesis 5: The experience level of the therapist will be associated with changes in object representation scores for F+ and F- data. To determine the effect of therapists on outcome (Hypothesis 5), therapists were divided into two groups, two who had considerable experience treating schizOphrenics and seven relative novices at treating schizophrenic patients. Previous research using this sample of patients and therapists (Karon and VandenBos, 1981) demonstrated the importance of the experience level of the therapist in producing an overall healthier outcome. Karon and VandenBos showed that whereas the inexperienced therapists seemed to effect specific improvements, the experienced therapists tended to produce improvements in all areas of functioning. Thus, Hypothesis 5 stated that the experience level of the therapist will be associated with significant differences in outcome for F+ and F- data. To determine the effect of the experience level of the therapist on outcome, ANCOVAs were computed to compare the outcomes (20-month object representation scores) of patients (n = 7) of experienced therapists with outcomes of patients of (n = 14) inexperienced therapists. Correlations between object representation scores at 20 months were controlled by using the scores at intake as covariates. Four separate ANCOVAS were performed on the F+ and F- data. Table 5 59 Table 5 ANCOVA Comparisons Between Experienced and Inexperienced Therapists Experienced Inexperienced df F Sig of p Variable 20 Mo. Mean* 20 Mo. Mean Blatt F+ 10.48 6.39 l 3.60 .07 Krohn F+ 4.20 2.89 l 4.20 .05* Blatt F- .29 1.50 l .64 ns Krohn F- .71 .64 1 .87 ns *Table means were not adjusted for intake object representation score. p_= < .05 reveals group differences clearly favoring the experienced therapists. The results show that patients treated by experienced therapists made significantly greater Krohn F+ gains than patients treated by inexperienced therapists, .§(l) = 4.20, p <, .05. The corresponding Blatt F+ differences closely approached statistical significance, [2(1) = 3.60, 2 <1 .07. No significant differences were found between the patients of experienced therapists and the patients of the inexperienced therapists on the F- data, Blatt F-, {(1) = .64 p <; .05: Krohn F-, 3(1) = .87, p_ < .25. Thus, Hypothesis 5 was supported for F+ data, but was not supported for F- data. Paired t Tests: Paired 5 tests were used to compare intake data with data taken at 20 months for experienced 60 and inexperienced therapists, respectively. As in the above analyses, scores were computed separately on accurately perceived (F+) and inaccurately perceived (F-) human percepts. F+ data. Table 6 shows that for accurately perceived human percepts (F+), the difference between the means at intake (M = 2.62) and 20 months (M = 4.40) for patients of experienced therapists approached significance for the Krohn data, 3(6) = +2.18, p_<; .07, but not for the Blatt data, 3(6) = +1.19. Patients of the inexperienced therapists, however, did not show significant changes on either of the object representation measures. Table 6 Paired E Tests Comparing Object Representation Score at Intake and at 20 Months for Experienced and Inexperienced Therapists Experienced Therapists (N=7) Inexperienced Therapists (N=14) Blatt Krohn Blatt Krohn Intake 20 Mo. Intake 20 Mo. Intake 20 Mo. Intake 20 Mo. Vhrhflfle g3; M 7.11 10.48 2.62 4.40 7.24 6.39 2.56 2.89 SD 5.73 4.61 2.10 1.02 6.96 6.20 2.48 2.45 .E +1.19 +2.18 -.72 +4.82 ‘pf ns .07 ns ns g:_ M 2.36 .29 1.36 .71 3.05 1.50 1.30 .64 SD 3.50 .76 1.98 1.89 4.07 3.50 1.84 1.21 g -1045 -055 -200 -1061 ‘p ns ns .06 *All.£_tests are two—tailed 61 F- data. Table 6 shows that for (F-) object representation scores, patients of both groups of therapists (experienced and inexperienced) showed insignificant decreases in scores on F— percepts. However, the 3 tests for the patients of the inexperienced therapists approached statistical significance for the Blatt data. Hypotheses 6 thrgggh 9: Whereas subjects diagnosed as paranoid will show quadratic trends on the F+ data over four points in time (decreasing scores from 0 months to 6 months and increasing scores from 12 months to 20 months), non-paranoid subjects will show linear trends towards increasing scores over the same four points in time. An inspection of the means for each of these diagnostic groups (Table 7) revealed that over the four Rorschach evaluations (0 months, 6 months, 12 months, and 20 months), there was barely any variation in the means, with the largest difference between any two means being less than 1.25. Most importantly, it is quite clear by inspection that the paranoid patients did not change in the predicted direction. The prediction of a quadratic trend meant that the means at 6 and 12 months would be less than the means at zero and 20 months. As seen in Table 7, the data for the paranoid patients clearly contradicts this quadratic pattern. For example, the means for the Blatt scale did not decrease from intake (7.28) to 6 months (7.97). Similarly, the data for the non—paranoid patients clearly contradicts the linear 62 Table 7 Mean scores for Paranoid Patients on the Blatt and Krohn Scales Months 0 6 12 20 Paranoid Blatt 7.28 7.97 8.07 7.53 (M = 14) Krohn 2.65 3.30 3.20 3.50 Non-Paranoid Blatt 4.77 5.48 4.35 5.30 (M = 7) Krohn 1.50 2.28 2.11 2.22 prediction that stated that the means would show a steady increase from intake to 20 months: the Blatt means show a decrease from 6 months (5.48) to 12 months (4.35). Hypotheses 6 through 10, therefore, were not supported. Hypothesis 10: Object representation scores on accurately perceived human percepts (F+) will be linearly correlated in a positive direction with clinical status ratings. To assess the degree of association between the object representation measures and clinical status ratings, Pearson product moment correlations were computed for the F+ percepts at each of the four time periods corresponding to the times when the Rorschach was readministered (zero months, 6 months, 12 months, and 20 months). As seen in Table 8, none of these correlations reached statistical significance. Thus, Hypothesis 10 was not supported. 63 Table 8 Pearson Correlations Between Object Representation Score and Clinical Status Ratings Months 0 6 12 20 Blatt 016 -003 001 019 KrOhn 005 -013 004 .00 DISCUSSION The significance of the present study lies mainly in two areas. First, the present study expands the literature on the clinical usefulness of object representation measures, specifically, the Blatt and Krohn scales. The literature to date already had suggested that these scales were clinically useful in distinguishing between schizophrenic and borderline diagnostic groups. This study shows that, in addition, the Blatt and Krohn object representation scales can be used to document change within a single group of patients over time. This positive change is defined by the scales as an increase in the accuracy and organization of a patient's object representations. Moreover, this observed improvement supports the suggestion made by Gunderson and Comes-Schwartz (1980) that investigators explore the subjects‘ experience of self and others pre- and post-treatment in evaluating a mode of psychotherapy aimed at changing the self. Second, the significance of this study lies in the fact that the measures used gg document change in patients secondary to treatment. The change influenced by therapy plus medication, as noted above, is in the direction of 64 65 increased accuracy and organization in the patients' object representations (F+). The change influenced by medication is in the direction of decreased organization and elaboration of distorted object representations (F-), as indicated both by the psychotherapy-plus-medication group and the medication-only group. However, the three-way ANCOVA fell short of statistical significance. While these changes associated with 20 months of treatment were statistically significant for each group, the difference between the three groups fell short of statistical significance. Overall, the present data are in accord with the conclusions of Smith, Glass, and Miller (1980). Through the use of a meta-analysis, these authors integrated findings from 112 research studies that compared the benefits of drugs against the benefits of drugs plus psychotherapy and psychotherapy only. They concluded that psychotherapy plus drugs offers the most benefit to psychotics. Furthermore, they determined that drugs only were somewhat more effective than psychotherapy for psychotics. A parallel pattern of effects was observed in the present data: the patients treated by psychotherapy plus drugs showed the most positive findings, improving on both F+ and F- data. The patients who were treated by drugs only showed the next number of findings, improving on the F- data only. Finally, the patients treated with psychotherapy only failed to show improvements. Based on 66 the observed order of effects, our data closely follows the pattern of effects determined by the Smith et a1. meta-analysis. It may well be that treatment by medication is accompanied by a decrease in distorted fantasy life and inappropriate object representations, but that an increase in differentiated and appropriate object representations occurs only when psychotherapy is added. It is puzzling that psychotherapy alone did not seem to have a potent effect on these measures. It should also be noted that in the psychotherapy-plus-medication group, four of the patients (those treated by Dr. Tierney, the experienced therapist) were taken off their medication within two months of therapy. In evaluating the benign effect of medication, it should be noted that these effects were measured after 20 months, an insufficient period of time for possible negative neurological changes induced by medication (Breggin, 1983). The Effect of the Experience Level of the Therapist The present results showed that when the patients treated by inexperienced therapists were compared with the patients treated by experienced therapists, the latter group made greater gains in their F+ object representation scores. This finding extends the findings of Karon and VandenBos (1970). Karon and VandenBos found that on other measures of outcome, e.g., clinical status ratings, experienced 67 therapists were associated with superior outcomes. The patients of the inexperienced therapists did show an improvement on the F-data which approached statistical significance: however, the change on the F-data for experienced therapists was not significantly different. Object Relations Theory and Psychotherapy Research The results of this study challenge theoretical assumptions made by object relation theorists. Most object relation theorists (Fairbarin, 1952: Guntrip, 1969: Klein, 1948) assume that change in a person's developmental level of object representations occurs only after prolonged and intensive psychotherapy. The present results showed that with appropriate psychotherapeutic interventions, considerable gains were made by patients who averaged only 70 treatment sessions. A recent finding by Ryan (1984) buttresses this challenge. Ryan found significant changes in the develOpmental level of object representations in a group of inpatient schizophrenics who had been treated for about nine months with psychoanalytically oriented methods. From the more general viewpoint of psychotherapy outcome research, the results of this study also challenge the pessimism of May (1975) and Holon and Beck (1978). Their reviews of psychotherapy outcome concluded that psychotherapy adds very little to the psychological growth 68 of schizophrenics. As shown by the present results and Ryan's outcomes, this pessimism appears misguided. In the future, a fruitful line of investigation for psychotherapy researchers might lie in trying to answer the difficult question of exactly how the psychotherapy mediates changes in object representations. This question might be answered through research that directly examines process variables (Orlinsky and Howard, 1978). Limitations of the Object Representation Scores in This Study A question arises about the meaningfulness of the object representation measures due to their failure to correlate with clinical status ratings, an outcome measure that in previous studies correlated highly with a number of important change variables (Karon and VandenBos, 1981). Thus, it is of some concern that the present work's clinical status ratings did not correlate with object representation scores. A possible explanation for this failure is that the scales themselves may not be as sensitive to variation in overall health, as suggested by previous studies (Spear and Lapidus, 1981: Schwager and Spear, 1981). This latter explanation is supported by the failure of the present study to detect any distinctive pattern of comparison between change in the two different subgroups of schizophrenics, paranoid and non-paranoid. Schwager and Spear found that during the first stage of treatment, paranoid schizophrenics show a i— 69 decrease in object representation scores, but their scores increased with further treatment. In contrast to Schwager and Spear's findings, the present study found negligible temporal variation in the scores of paranoid patients. The present findings, therefore, constitute a failure to replicate Schwager and Spear's findings. It may be, however, that the object representation measures assess a different dimension of functioning than those included in the clinical states rating. The present findings suggest that the object representation measures are sensitive to change, but that their range of application needs to be further explored. Whereas the Krohn scale was designed (Krohn and Mayman, 1974) as a measure of affective and thematic qualities of object representations, the Blatt scale (Blatt et al., 1976) was designed as a measure of the cognitive and structural features of object representations. The high correlation between the two object representation scales in both the present sample of schizophrenics (r = .90) and the sample described by Spear and Lapidus (1981) (r = .65) negates the meaningfulness of the distinction between affective/thematic object representations and cognitive/structural object representations, at least for our sample of schizophrenics. The loss of this distinction implies a loss in the overall validity of these two scales for assessing the differences between cognitive and affective features of object representations. werner's (1948) developmental theory 70 suggests a possible explanation for the correlation found between these two scales when applied to schizophrenics. An additional explanation for the high correlation between the scales may be found in Werner's (1948) comparative developmental theory. Werner's theory holds that, as compared with normal adults, persons with severe psychopathology generally show less differentiated perception from an affective point of view and also from a cognitive point of view. An application of werner's theory to the present data is possible if we assume that (a) the Krohn scale does in fact measure the "affective differentiation" of object representations, and (b) the Blatt scale does in fact measure the “cognitive differentiation” of object representations. Then, according to Werner's theory, BEEB scales should reflect relatively little ”object differentiation” in schizophrenics. The high correlations between the two scales observed in both the present sample of schizophrenics and in the sample described by Spear and Lapidus are congruent with Werner's theory, i.e., both scales revealed low scores or little "object differentiation." Support for the Validity of Psychoanalytic Theory of Change The results of the present study showed that schizophrenics treated with psychotherapy and medication advanced their developmental scores (i.e., their object representation scores) on accurately perceived human percepts. The proposition that more accurate object 71 representations occur simultaneously with overall ego development is supported by two psychoanalytic theories: 1) the traditional theory of ego development of Anna Freud (1937) and 2) the object relations oriented theory of Otto Kernberg (1976). Empirical work related to Kernberg's theory was discussed earlier (Blatt et al., 1976: Mayman, 1967; Ritzler et al., 1980; Spear and Lapidus, 1981). As suggested by Anna Freud's theory, more accurate object representations may result from improvement in other areas of ego functioning. Or, alternatively, other areas of ego functioning improve as object representations become more accurate, as propounded by Kernberg (1976). Previous empirical work (Karon and VandenBos, 1970: Glatt, 1971) based on the present sample of psychotherapy patients supports both of these psychoanalytic ideas about change. These previous investigators (Karon and VandenBos, 1970; Glatt, 1971) discovered, among other findings, that schizophrenics who were treated with psychotherapy showed greater ego functioning in at least three areas: (a) less thought disorder, (b) greater cognitive flexibility, and (c) greater tolerance of ambiguity. If looked at from the orthodox ego-analytic viewpoint of Anna Freud (1937), these aspects of increased ego strength will facilitate more accurate self and object representations. Anna Freud believes that as the ego is strengthened, there is concomitantly less need to distort inner experiences of the self (e.g., with ego mastery of aggressive instincts, there 72 is less need to "identify with the aggressor"). Conversely, if the present findings are looked at from the more object relations oriented viewpoint of Otto Kernberg (1976), then changes in object and self—representations facilitated the change in other areas of ego functioning. In Kernberg's theory, self and object representations themselves are primary: they themselves are psychic structures that provide strength to the ego. As the individual integrates self and other images, this integration reflects in more anxiety tolerance--an ego strength. Process of Change: An Example Two responses from a patient treated with psychotherapy will be used to illustrate the process of change. Otto Kernberg's (1976) developmental object relations theory of the character pathology will be the guiding theory in explicating the observed changes. The first of these responses is from intake, and the second is from the data at 20 months. Both responses were to Card 3 of the Rorschach. At intake, the patient reSponded to Card 3 with: "Two people-~two Martians or two girls--red and black.“ On inquiry, the patient responded, "Tits, head, body, dong (7), you know, 'ding-dong'--but it's not . . . I've seen them around.“ When asked about the Martians, he replied, ”I don't know what they look like and I don't want to go up there and find out.“ 73 What is most striking is how this patient merged, within the same portion of the ink blot, both male and female sexual attributes. Specifically, the patient merged the male “ding-dong" precept--an association to the blot's phallic contours--with the female ”tits” percept. Both were blurred into a single object representation. Also striking is the way in which the patient's perception vacillates between the sexualized human percept and the nonsexualized but paranoically-tinged (Martian) quasi-human percept. How does this apply to Kernberg's developmental of character pathology? According to Kernberg's developmental theory of character pathology, this patient carries the diagnosis of schizophrenia and is, therefore, at a primitive or developmentally less advanced level of character pathology. According to Kernberg, object representations at this level of development are impaired because the individual can neither master nor modulate the influence of the drives, especially aggression. Kernberg (1978) explains (p. 146), Excessive pregenital aggression also causes a pathological condensation of pregenital and genital conflicts with predominance of pregenital aggression and is evidenced by sadistically infiltrated, polymorphous perverse infantile drive derivatives which contaminate all the internalized . . . object relations of these patients. 74 Kernberg also noted that patients at this level of character development are very prone to process rapidly many contradictory object representations. This is both because the splitting defense predominates over other defenses in these individuals (thus the contradictions in their perceptions), and because these individuals are fending off unusually intense drives such that they need to recruit constant defense. The patient begins by seeing two ”people" in terms of sexual images, and in terms of polymorphous (fused male and female) sexual images, as we said above. Kernberg's description in the quotation above applies in that he describes these individuals by saying of them, "polymorphous drive derivatives contaminate all the internalized object relations of these patients." To start with, this patient's percept “people” (his object representation of self and other) is limited to a "polymorphous drive derivative." Furthermore, Kernberg says these “polymorphous . . . drive derivatives“ are ”sadistically infiltrated" and ”perverse." The blatant primitive aggression (infantile sadism) in this first image is suggested by the demeaning childish references to male sexuality (”ding-dong") and to female sexuality (”tits"). Lastly, Kernberg describes these drive derivatives as ”infantile." Clearly, the infantile, primitive quality of the human percept ("people") here can be seen in the primary process and regressive clang association noted in his word for the male phallus 75 "ding-dong.“ In sum, polymorphous perverse infantile drive derivatives can be said to be present, as can be seen in this patient's response. And such "sadistically infiltrated,” polymorphous infantile drive derivatives could be said to contaminate the internalized object relations of this patient, extrapolating from the data of this response. Continuing on to consider how else Kernberg's developmental theory of schizoid and schizophrenic characters may apply to the present study, let's look at the next percept this patient reports in his response at intake: "Two Martians or two girls." The patient's rapid shift from people to Martians and back to girls shows a rapid processing of disparate representations. The contradictory quality of his object representations can be seen in his rapid shift in focus from the genital (sexual) to the paranoid (or projected aggression). I say projected aggression because his responses to inquiry about the Martian ('I don't want to go up there and find out") clearly indicate that he is afraid of the Martians and that he perceives the Martians as threatening aggression towards him. Furthermore, in that the aggression is clearly his own within the first percept of “people,” but in that, as his response to this card progresses, he shifts to seeing the aggression as coming from the outside, in the second percept (coming from the Martians), one could easily say that he begins to project his aggression. Kernberg's 76 theory can be said to further apply in that he splits off (or projects) the aggression from his self to the other (the Martian). Thus the splitting, or splitting off, defense is the predominant defense employed here. The rapid shifts Kernberg describes can be seen as due to the intensity of aggression the patient experiences. Accordingly, as aggression floods the patient's sexual image, he needs to rapidly shift to a Martian in order to get rid of, or rid himself of, the aggression that flooded him by attributing it to someone else. Finally, the source of the flipping from sexual to paranoid and back to sexual images may be an attempt to separate and therefore protect ”good objects" from bad objects." At 20 months, this patient's response to the same blot is, "Two women pulling on something . . ." (Inquiry) ”Bust, back-end, legs, high-heeled shoes, head." As can readily be seen, the patient had moved from a low-level undifferentiated human (a low score on the Blatt scale) to a pOpular response involving two well-defined and interactive women (a much higher score on the Blatt scale). Clearly, the patient was still focused on the sexual and narcissistic attributes of the percept. But this percept is, after 20 months of treatment, now clearly of one gender, no longer polymorphous perverse. The sexual attributes described are no longer infantile. Although the females are not involved in a well-defined activity, the patient describes them as separate people. The erotic or genital 77 image is no longer infiltrated with aggression or sadism. There is no longer obvious evidence of splitting in that there is no splitting off, or projection of aggression, nor does he shift from all ”good“ to all "bad” objects. Finally, there is considerably less aggression evident in the percept. Appendix 4 presents several other representative examples of change. CONCLUSION This study represented an attempt to extend the literature on changes in object representations for schizophrenics. First, schizophrenics object representations improved with psychotherapy plus medication, there was increased object differentiation, along with a decreased need to embellish object relations with fantasy. Second, this study offered support for the belief that the effect of exploratory psychotherapy with schizophrenics can be illuminated by instruments that are sensitive to changes in inner experiences. Third, this study challenges the pessimism or theorists who conclude that psychodynamic psychotherapy adds very little to drug therapy for schizophrenics. Finally, this study suggests that when schizophrenics are treated by experienced psychotherapists, psychoanalytically oriented psychotherapy can be effective at helping these people develop healthier internal models of other people. 78 APPENDICES APPENDIX 1 ‘ r .l? k. A " d Consulting and Clinical Psychology .04 y... :4. No. 1. 119-23: QPPQ b\d\ ;( .1 “7:5; QUANTIFIED JUDGMENTS OF MENTAL HEALTH FROM THE . 'i- ‘- 73“ RORSCHACH, TAT, AND CLINICAL STATUS INTERVIEW BY [5,? MEANS OF A SCALING TECHNIQUE ‘ -. .1 a ‘ BERTRAM P. KARON 2 AND PAUL O'GRADV 1' . Michigan Sate University ‘3 _ . ft; . Meaningful measures of mental health derived from clinical instruments such ‘ '. as the Rorschach, TAT, and Clinical Status Interview (C5!) are not ordinarily 6"" quantifiable without sacrificing the subjective, clinical, and global nature of )7 - the data. A practical technique for sealing clinical judgments ls described. “*3" Judgments of 35 schizophrenic patients yielded intrajudge reliabilities from to 'L .79 to .96. and interiudge reliabilities from .81 to .95. Days hospitalized during J g‘ the subsequent six months correlated —.63 with the TAT scale. -.58 with the '2 ‘Jf Rorschach scale, and --.7l with the CSI scale. Days hospitalized subsequent to ;_g.‘..fj, a later testing. correlated -.64, -.49, and -.64 with scales determiner! from if that later testing. The scaling technique thus was found to he reliahle and 1 [T J valid for measuring mental health. jamming changes from morbid to post- .id stages following therapeutic treatment '7 a complex issue. Methodologists re- it quantitative data for statistical analy- .' ,1 but clinicians find that those measures * flick are most easily quantified may be clini~ "any meaningless. The Rorschach, TAT, and finial Status Interview (CSI) have been ,5] in the past by clinical personnel as mea- .' ‘ of mental health. But such data are not $.15“,in quantifiable without sacrificing the .fijective, clinical, and global nature of the '“ appropriate evaluation of such observa- lt is proposed that a simple method of . , .. may be used to reduce such data to ,, . (native measures which are clinically i ' n'ngful. 'wn is assumed that while there are obvious : i ' cities, nonetheless, a unidimensional W of “mental health” may be meaning- : ”employed. This assumption may be justi. on the basis of data presented by Lu- T”, (1962), who found that the first - fjpal component of a factor analysis of :1 . t . ¢ IL 5‘ . ," rm investigation-was supported in part by the ' . State Psychotherapy Project. National In- , of Mental Health Grant Mil-08790. The au- ' ”m to express their appreciation to the inter- , Leon Berman; the projective tester. Stanley . uck; the clinical raters. Erma Alperson, Ross '0'“ Charles Glatt. and Frank Long; and Pa- ‘3 3m. and Adeline Reith for their assistance. _ {fawn for reprints should be sent to Bertram t m Department of Psychology. Michigan mini-shy, East Lansing. Michigan 48823. . g. 3 at '. -'l .: a _ I .o ,4 i, .- ratings with respect to 14 specific aspects of psychopathology accounted for 60% of the variance. Thus, one dimension meaningfully accounted for a great deal of the variation in emotional health. Similarly, llieri, Atkins, Briar, Leaman, Miller, and Tripodi (1966) found that experienced clinicians, ‘given the discrete classifications of standard psychiatric nosology, may resort to some underlying di- mensional basis in discriminating among cases. Menninger, Mayman, and Pruyser (I963) criticize the value of standard psychiatric nosology and have advanced a classification of their own which appears to be a dimension of pathology, ranging across five categories from least to most disturbed. These categor- ies are "nervousness, neurotic, social acting out, psychotic, and severely disturbed.” it is the present authors’ contention as well as the previously cited authors’ that there is evidence that despite the obvious multiplicity of symp- toms, much of mental illness may be sum- marized in terms of a unidimensiunal con- struct. Moreover, a sophisticated judge can take into account the relative severity of alterna- tive symptoms and the implications of en- vironmental context. Thus, while ability to hold a job is a central criterion of adjust- ment, :1 patient who has a job waiting for him may not be necessarily ln-ttcr adjusted than a patient having difficulty in obtaining work de novo. Similarly, an increase in the 39 79 D.- 80 i 2.ltl Butt't'ttAM l’. KAmm AM) l'mn. U'tiuanv severity of a symptom like manifest anxiety which, everything else being equal, would in- dicate a worse adjustment, might mean move- ment toward health if it represents the results of the patient’s abandoning a more patho. logical defense, for example, the disappearance of hallucinations. The proposed scaling pro- cedure allows such factors to be weighted sub- jectively, and the resulting quantitative di- mension should be more successful at sum- marizing health-pathology on a single dimen- sion than the necessarily linear combination of symptoms in the factor analysis of Ln- borsky. The procedure presented in this paper con- sists of obtaining global ratings of emotional health from a clinically trained rater by pre- senting him with the protocols (Rorschachs. TATs, or CSI) of two patients. First, the rater is asked to judge which is the healthier indi- vidual. Using that healthier person as the standard, he is then asked to estimate the proportion determined by the ratio of the emo- tional health of the sicker individual to that of the healthier. Procedures for selecting the pairs to be rated, obtaining the ratings, com- puting the scale values, calculating coefficients of internal consistency (i.e., a measure of unidimensionality), of interrater reliability, and of validity are described. METHOD Thirty-five schizophrenic patients were seen by a psychiatrist who interviewed them for to minutes and a psychologist who administered a ZO-card TAT and a Rorschach test. The interview and the TAT were tape recorded. The TAT was then transcribed by a typist. The Rorschach responses were recorded by hand. and the protocols scored by the Beck system. All three sets of clinical materials (interview tapes. Rorschach protocols, TAT protocols) were given to two pairs of raters who were advanced graduate students in clinical and personality psychology. One pair rated emotional health from the projective ma- terials and one pair rated it from the interview re- cordings. liowever, the TAT ratings were made separately from the Rorschach. That is, the two raters first rated the TAT and then the Rorschach protocols. Usually, while the raters rated the pro— tocols independently. they were permitted to clarify the content of the protocol prior to the rating and to discuss the ratings afterwards. l'be raters were asked to compare a pair of patients at a time. They were instructed to rate “mental health" becau'e protesting had indicated that “mental health" is judged more consistently than “mental illness." it seems as if there is no clear base line of “health" from which to measure “illness," but that in rating “health” the base line seems to be something like “no functioning at all." The raters were instructed not to use a formal weighting system for the specific criteria of emo. tional health provided, but to use such criteria clinically to arrive at a global. relative judgment of emotional health. These criteria, for the clinical status interview, were ability to take care of self. ability to work, sexual adjustment, social adjust. ment, absence of hallucinations and delusions, degree of freedom from anxiety and depression, amount of affect, variety and spontaneity of affect, satisfaction with life and self, achievement of capabilities, and benign versus malignant effect on others. For the projective tests, in addition to those cri. teria above, the following were used: the length of protocol, absence of stereotyped responses and more varied material, presence of benign fantasies, helping nurturant parental figures, self-confidence, reality testing, and direct representation of problems Presence or absence of primary process material on the projective tests was not to be used a.» an indi. cation in itself of health or sickness, inasmuch a, presence of primary process material has different significance in the context of differing treatments The dimension of health was further clarified by having the raters read a case history by “it-55mg}, and Ricks (1066, pp. 145473), entitled, “Winn, a happy man." This is a case history, complete With projective protocols, of an extraordinarily healthy- individual who has not had therapy. ' After reading or listening to a pair of protocols, 1 rater made a judgment of the ratio of the emotimul health of one person to the health of the other. Tin.- judgment was recorded on a BO-centimenler line The rater judged who was healthier of the two patients and placed him at the rightmost point of the line The leftmost point represented an absence of menm health, or the zero point. That is, the length of the line represents the health of the healthier individual The second decision involved taking the less healthy member of the pair and deciding, by placing a marl. at some point along the fine, what proportion of the first person's health the other person had. in order to get consistent judgments, it is necessary to require that the healthier person be the length of the line, so the observation is always a preportinn. with the healthier person being the denominator and the less healthy person the numerator!I if all possible pairs had been compared, there would have been (lo .«7 .lS)/Z comparisons, or 0,.) judgments for each rater on each type of material To reduce the work, it was decided to select 2(a- l) comparisons or 70 judgments. Previous work on “This is based on preliminary work on this scal- ing technique using psychopbysical stimuli which had been carried out by Donald P. Estavan. The logari:|.. mic solution and its application to clinical mate-m! is the responsibility of the authors. however —-..—.—-—.-.—- .d Junostsms or MENTAL HEALTH Wehophysical stimuli had suggested that stimuli m to each other provided the most stable judg- geats. However, it was necessary to order the stimuli (patients' protocols) before such judgments could be m. Therefore. a comparison stimulus (person) 16 chosen and all other stimuli were judged against gm person. that is. using the constant stimulus “thod. These first n-t judgments were used to mice the stimuli. and then adjacent pairs were de- guiaed to be presented to the judges for ratings. fie tentative scaling was not discussed with the ‘ Jugs, and the pairs were randomized before pre- gptatlon to the raters. The raters reported that it ' n! easier to make the judgments using a constant mules when rating the interview recordings, but ill the adjacent comparisons were easier to rate M dealing with projective protocols. In order to illustrate the calculations involved in mining the scale values, suppose there are three ' “ts—A, B, and C. A is compared with 8. found .he .60, and B is rated to be .80 of C. if these W are meaningful, we can then predict that A id be .80 x .60= .48 of C. When the rater actually '3. compare A with C. we get a check on whether he ratings form a ratio scale. This check can be armed simultaneously over the whole matrix of m as described below. The judgments of a single rater may be organized m a matrix with each stimulus being represented 5, a row and a column as shown in Table l. which goes all possible comparisons of three stimuli. Each 33 consists of the column stimulus divided by the .1 stimulus. In the matrix there are missing en— nit wherever the numerator would have been larger in the denominator. The missing entry is obtained '3, taking the reciprocal of the comparison. If A and _|are compared. A is seen as .60 of B. in the writ. A over ii is recorded as .60. in the same Quiz, 3 over A is missing and is determined by “puting the reciprocal of A over B, that is. the ’udprocal of .60 or 1.67. (The diagonal entries are [5, definition 1.00.) A complete matrix is shown in ‘ nu: 2- ‘ father computations can be simplified by trans- to logarithms. The logarithm of a ratio ”was the logarithm of the numerator minus the f TABLE 1 A Maratx or JUDGIIENTS stimulus .\ u ' c e I). 2‘ A A A A A '3 s' B u u n a I: <.; C c c c 81 231 TAlllJ". 2 AN lt.t.tn:'ru.vrtvn (Tourer-21+: Marius Stimulus A It (I A 1.011 1.07 2.08 ii .0" Lilli l.25 C .43 .811 1.00 logarithm of the denominator. that is, log All! equals log A -log 8. The matrix at ratios now becomes a matrix of differences, which makes the solution for the scale values arithmetically simple. Moreover. the logarithms of the judgments tend to. have more nearly equal errors of measurement for extreme and nonextreme judgments than the raw judgments and tend to be more linearly related to other variables as will be discussed below. Those who are familiar with scaling techniques will recognize that the data analysis to be described is similar to that for Thurstone’s well-known Case V of the Method of Paired Comparisons (Guilford. 1954. pp. 154-178; Mosteller, 1951). In Thurstone's Case V, the entries in the matrix are obtained by taking the normal deviate corresponding to the percentage of times one stimulus is seen as greater than another in a large number of trials; in the present teclmiQUe. the entries in the matrix are ob- tained by taking the logarithm of the ratio ob- tained from a single judgment. Beyond that point, the computations are identical. This computational procedure and its logic may be best understood by examining Tables 3 and 4. in Table 3. the logs of the ratios are substituted [or the ratios in Table 1. it will be recalled that the log of a ratio is numerically identical with the diflerence between the logarithms of the numerator and of the denominator. As seen in Table 3. it the entries in a column are averaged. the log scale value of that column stimulus can be determined as a deviation from the average log scale value of all the stimuli. Since the zero point of the log scale (i.e.. the unit of measurement of the original scale) is arbitrary, one may set the mean log scale value at zero. in which case the result simplifies to the log scale value being the average of the column. This is a leastosquares solution for a complete data matrix (all possible pairs compared). and is algebraically identical (but not experimentally identical) with Mosteller's (1951) derivation for Thurstone's Case V. Numerically. this is illustrated by Table 4, where the logs of the en- tries in Table 2 are entered and the scale values com- puted. A complete comparison of all possible pairs re- quires n(n -1)/2 comparisons, where n is the num- ber of stimuli. However. a - 1 comparisons are suffi- cient to determine the scale and 2(n ~1) compari- sons. as used in this study, provide sufficient data to test internal consistency. A more general least-squares solution. for both' complete and incomplete data matrices. may readily “ .- ‘b ‘O -.- 82 232 Baa-rams P. Kaaou mo PAUL O'Gaaov TABLE 3 Mann: or Looutrnus or juoourzu-rs Stimulus A B C A logA — log A log B - log A log C -- log A B logA-logB logB—l-g" logC-logB C log A — log C log B — log C log C — log C Sun 310gA Jiogli 310gC -(logA+logli+logC) -(logA+logB+logC) —(logA+logB+logC) Average log A log B log C _(logA+logB+logC _(logA+logB+logC _(logA+logB+logC) 3 3 3 a Note—1f the avenge log scale value ( h. A +10; 8 + '0‘ C) la 3 tologA.10gl.andiogC.rupectlvely. be ulculated as follows. For any one stimulus j, a least-squares solution for L, (the logarithm of the scale value of j) is given by the following equation: 1 1 “'2V5§”"'+173§“ [‘3 where Du is the logarithm of an observed ratio of scale defined as equal to 0. then the column averages simplify value is set at zero. If, however, the column is not complete, the solution seems indeterminate at first glance. However, one may compute an iterative solution using any trial values of L; to compute the next approximations. The solution rapidly converges, in the empirical examples, 4 to 12 iterations sumeed, To make the procedure'more explicit: Sum the columns of observations (after trans. stimulus j to stimulus i, that is, Du is an empirical 1. estimate of L, - La.“ and a" is a value of r' for which forming to logarithms). the comparison Du was observed. It should be noted 2. Divide by the number of observations in that that for these purposes, if 0.. was observed, Dr, is column. considered to have been observed since it is deter- 3. Sum the trial values of 1., for all stimuli With mined by the same judgment and equals -D,.. which. that stimulus j (the column stimulus) A little algebra will readily generate the solution was compared. for the complete data matrix from Equation 1. in 4. Divide by the number of observations (um any incomplete matrix in which a whole column number as in Step 2). (and row, obviously) is determined, the value for 5. Add the results of Steps 2 and 4 to obtain the the stimulus is determinate. That is, as with the next estimate of L1. complete data matrix, the mean of the column of log 6. Do this for all stimuli (i.e., all columns of the observation is the log scale value if the mean log matrix). TABLE 4 Locamrnus Coaazsronomo 10 mt: Eaten-:5 m TABLE 3 Stimulus A B C A 0 .22185 .31}? 9.77815 - 10(=-—.22185) 0 .0969! C 9.68124 - 10(= -.31876) 0.90309 - 10(= —.0‘)691) 0 Sum — .54061 .12494 .4156? Log scale value (column average) -.18020 ”“65 .13856 Anti-log scale value .66 1.11 1.38 ; mt Steps 3 through 6, using the new values 5, 's of L:- ‘Q: ‘ Slop when the values of L, do not change ap- ‘3’ ’ predlbly. 3 an trial values of the L, to begin 14”.: rmij=Q Some iterations can be A by using better initial trial values determined 5" g—l of the comparisons. But no matter its; t8 trial values, the solution will converge. 99...,“ be noted that one of the scale values is f" (as in paired-comparison scaling) so that 511; n ,3 the average L, or the value of one. par- _ as gqual to zero to suit one's convenience. 231-“ incomplete data matrices, it tends to be {are .m to set one of the L, equal to zero. utations yield the logarithm of the -- . 1b MP . .. . . “ Intrarater reliability for the logarithmic be csfimaied by a method Similar to that I? . . by (3an and Tukey (1957) for as- ""1: the reliability of the Thurstone paired-com- ,‘ abalone. The total sum of squares (the ‘3. (the squares of the observations over half the .‘i' . either above or below the diagonal) may be j" by the total degrees of freedom (the number far" Wt observations) to determine 7', an - , of the total variance. i“ a. scale values computed by a scaling 1. " one can derive a theoretical value of what Maori ought to be. if the scale values were M if the scaling method worked exactly. If M the theoretical value of an observation ' bfllifiauy observed counterpart. one obtains ii 1"“ mummy. By squaring these errors and ' over half of the matrix (either above or .3 diagonal), one obtains the discrepancy a ‘ flares. Thus, the total sum of squares can 5" w into two parts: the discrepancy sum of -. .. and the sum of squares accounted for by the {'7' J d j; tint . _ mu! degrees of freedom may be secondarily 'hul. . .. Since "—1 scale values were computed #3 ~ a“, ”—1 degrees of freedom were used - . ° the scale values. The total degrees 01 ,i’i (the number of independent observations) ‘3 lb number of scale values determined from i— u “—1) yields the degrees of freedom for I! or discrepancy variance. I. a” wepancy sum of squares divided by the _ degrees of freedom yields 0, the dis- ~V . , variance. Internal consistency, R.“ may now “ed as follows: 'I' — I) p . H Rss =3 ".1?“ [ll t 's ”r 1' ha noted that all failures of the scaling .W J Wracies of judgment, unreliability of the d lack of unidiniensinnality increase 0 and .. reliability and validity may be deter- .‘eg the ordinary product-moment correla- scale values of two raters or .1“ use! the 108 83 ‘ . JuooMiims or MENTAL HEALTH 233 between the log scale values and an external cri- lerion. Data for the present study were taken from the pretreatment and follow-up evaluations of patients in the Michigan State Psychotherapy Project (Karon a O'Grady, 1969). The patients were administered a battery of tests of intellectual functioning at each evaluation as well as the Rorschach, TAT, and CSI. These data were gathered for research purposes. iienc, tne psychiatric interviewer, testers, and raters were personnel not connected with the inpatient service or the treatment of patients. They did not know the patients before evaluation, and did not know the course of treatment. The data were stored outside the hospital. Hence, the results of these ex- aminations were not part of the ward staff’s basis for determining the course of treatment or discharge. RESULTS The data to be presented in this section can be subsumed under three categories. The cate- gories will be (a) evidence for intrajudge re- liability, (b) evidence for interjudge reliabil- ity, and finally (c) some evidence for the validity of the scaling procedure. The first rater using the clinical status interview had an internal consistency of .83, while the second rater had the surprisingly low internal consistency of .37. The two 'l‘A'l‘ raters had an internal consistency of .86 and .81, while on the Rorschach they had .96 and .92. One advantage of the R33 coefficient is that it immediately reveals whether a judge’s rat- ings are consistent. and unidimensional. lf R33 is not satisfactory, one can search for and remedy the cause of the inconsistency. The low internal consistency of one rater led us to question his data. This rater was blind, and had to depend on someone else to transcribe and label his judgments. An error or errors of transcription seemed likely. Since scale values can be determined from n - 1 judgments, and since the raters carried out two sets of n -l judgments, scale values were determined from his first set of judg- ments and from his second set of judgments independently. lie was asked if the rankings of these scale values were reasonable. Those determined from the first it -— l were. mean- ingful to him. Those determined from the sec- ond n — 1 were meaningless to him, so he was asked to redo those judgments. After redoingr them, his internal consistency was .70. The traditional mode of assessing reliabil- -m..— “p-~—- - on... 84 234 ity in rating situations by correlating the scale values between the two raters was examined. Scatter plots of the logarithms of the scale values seemed to be more nearly linear in their relationship to each other and to ex- ternal criterion than the scale values. More generally, the use of log scale values, rather than anti-log, for statistical analyses seems indicated by these data. The interjudge reli- ability between the clinical status raters was .82, using the initial judgments, despite the low internal consistency of one rater! The in- terjudge reliability on the TAT and Ror- schach was .94 and .95, respectively. Both of these latter figures are considerably higher than would normally be expected on a rating task. The first correlation of .82 might be expected to go even higher if the intrajudge reliability of the first rater were raised. Iii- deed, this was the case; when the second half of the ratings was aednne by the rater who had the original negative internal consistency, the resulting interjudge reliability was found to be .87. For each instrument (CSl. Rorschach, and TAT), the 35 log scale values obtained by averaging the ratings of the two raters were then correlated with various independent cri- teria to get a measure of their validity. First, among themselves, the three scales correlate "moderately highly. The interview ratings cor- relate .44 with the Rorschach ratings, and .55 with the TAT ratings. The TAT correlates :64 with the Rorschach. inasmuch as the patients were psychotic at the beginning of the project, much of the variation in Porteus Maze scores and WAIS scores represented degree of functional im- pairment at the time of testing rather than intelligence. The patients had also been exam- ined on the Drasgow-Feldman Visual-Verbal Test (VVT), a concept formation task that is designed to be specifically vulnerable to the schizophrenic thought disorder. The Porteus Maze correlated .57 with the interview rat- ings, .47 with the Rorschach ratings, and .44 with the TAT ratings. The Wechsler corre- lated .43, .58, and .55 with the 35 patients’ scale values on the three assessment proce- dures. The number of errors on the VVT cor- related --.50, -.49, and -.41 with the three ratings. Bantams P. KARON AND PAUL O’Gnnov Finally, the number of days spent in the hospital during the following six months was used as one appropriate predictive criterion measure with which to correlate the three scales obtained from initial clinical status ma. terial. The 'l‘A'l‘ correlated —.63 with day, hospitalized, while the Rorschach correlated -.58. The clinical status interview correlated -.71. Surprisingly enough, when the new, more internally consistent, ratings for the second rater of the clinical status interview were substituted, the correlations of the interview ratings with the various criteria were essen. tially the same, that is, with 'lie TAT, .57; with the Rorschach, .41; with the Porteus Maze, .60; with the WAIS, .40; with day, hospitalized in the next six months, -.71. Thus, even when Rm is relatively low, the least-squares scale values still yield high inter. rater reliability and high validities. They are thus robust with respect to inconsistent or nonunidiinensional judgments. The procedure was replicated with the in- terview, protocol ratings, and test data ob- tained from the examinations of the same pa. tients after six months of treatment. R3,. for all ratings was .78 or better, and interrater reliabilities were .82 for the C81, .81 for the Rorschach and TAT. The CSl correlated .55 with the Rorschach and .57 with the TAT. and the two projective ratings correlated .30. Validity coefficients against the intellectual tests were (in the order CSI, Rorschach, TAT) .35, .47, and .45 for the Porteus Mazes; .46, .74. and .65 for the Wechsler-Bellevue [[5 and -.S7, —.81, and —.74 for the VVT. Cor. relations with days hospitalized in the six months subsequent to this evaluation (i.e., sixth to twelfth month of treatment) were ‘ ‘- ——-— ’-'*~-~ O‘cmmoluw M -.64 for the C81; --.49 for the Rorschach: , and -.64 for the TAT. DlSCUSStON The method yields high intrajudge and interjudge reliabilities and high validities for global clinical ratings of all three types of material. Surprisingly, the inclusion of some inter. nally inconsistent judgments in the ratings of one judge did not affect the scale values suffi. ciently to lower interrater reliability or Valid. ‘l Jumsumrs or MmrM. limorrt ly sppreclably. Similarly, using approximate fictions tor the scale values, sttch as one simi- ‘h to that suggested by Guilford (1954) for imp-stone’s Case V or as described in O’Grady id Karon (1968), was found to yield consid- 3“lily lower intrajudge reliabilities, but the ximate scale values were highly corre- md (r = .87 or better) with the least-squares g values, and the validities were reduced 'dy slightly. Nonetheless, the least-squares dutien is sufficiently easy to compute that “would recommend its use. ‘ 0f some interest are the raters’ reports in: the constant stimulus judgments were .» laid to make than the adjacent comparisons _ ’3 the clinical status interviews, but that impsrisons of adjacent stimuli were easier it: the Rorschach and the TAT. This differ- . i probably is due to the fact that the inter- fin were recorded on tape and presented a W formidable problem of recall than the Jam projective protocols. This task was Wed if one of the stimuli was repeated pin and again, and hence was more firmly ' 'tted to memory. This scaling procedure may be contrasted " other scaling procedures. The simplest of rating procedure, the method of abso- judgments, has two drawbacks: it re- !” the judge to have an internalized scale it end points and units before judgment be made meaningfully, and the judgment [3 experienced as very difficult by the , The technique Used by the authors, r, requires the judge to have neither . aplicit end point nor unit, but simply to Lye one human being’s functioning with . ’s, using one of the persons as the ° This concrete judgment is experienced we judge as much simpler. Moreover, in- “ consistency and degree of unidimen- rm may be examined empirically, when ‘w~'fl-‘9T&£"'§~ 54'5””; 7" ' ' 85 US judgments are made on pair-wise comparisons. Thurstone’s luth'etl-rutitpttt’lsmt technique. while requiring only a "which is greater?” judgment, has the fatal drawback that it re. quires each pair of stimuli to be compared by a large number of judges (50 or more) or by the same jmhze a large number of times. But if clinically sophisticated judges are required, one cannot for practical purposes obtain a large number of them, nor will one sophisti- cated judge cooperate in making the same judgment a great many times. REFERENCES Brent, 1., Arxms, A. I... Bar/tn, S., LEAMAN, R. 1... Mates, 11., 8: Tammi, '1‘. Clinical and social judge- ment. New York: Wiley, 1966. Gunman, 1. P. Psychometric methods. New York: McGraw-llill, 1954. Gueuxsen, 11., 8: Tuxrzv, 1. W. Reliability [or the law a] comparative judgment. (Tech. rep.) Prince- ton. N. 1.: Princeton University, 19"]. Know, B. P., & O’Gnaov, P. 1. Intellectual test changes in schizophrenic patients in the first six months of treatment. Psychotherapy: Theory, Research, and Practice. I969, 6. 83—96. Lvnoasxv, L. The patient's personality and psycho- therapeutic change. In H. Strupp & L. Luborsky (Eds), Research in psychotherapy. Vol. 2. Wash— ington, D. C.: American Psychological Association. 1962. Mcnnmcrzn, K., anuan, M., St Pnuvsea. P. The vital balance: The tile process in mental health and illness. New York: Viking Press, 1963. Mosreuen, C. F. Remarks on the method of paired comparisons: I. The least squares solution assum- ing equal standard deviations and equal correla- tions. Psychometriha, 1951, 16. 3-9. O’GRADY, P. 1., 8: Karma, B. P. Quantified judg- ments of mental health of thirty-five schizophrenics on the Rorschach. TAT. and Clinical Status Inter- . view by means of a new simple scaling technique. Paper presented at the meeting of the Ontario Psychological Association, Windsor. February 1968. Wrssaum. A. F... & thxs. 1). F. Mood and person- ality. New York: Holt, Rinehart 8: Winston, 1966. (Received February 25, 1969) APPENDIX 2 ' _ - ~— —.—-.—~- ‘O‘m ”901:1- 1 WWW”. ANALYSIS OF THE CONCEPT 0? THE OBJECT 0}! TIE RORSCBACII Sidney J. Blatt, c. Brooks Brenneis, Jean 6. Schizek Yale University and liar ion Click Southern Connecticut State College to Oonrighc. Dace-ha 1. 1975. 86 87 mm or cos-um: —‘ l. 3oloction of Responses A» lumen and quasi-hunch responses I. Aninnl responses with explicit hunan qualities 1!. 8coring Procedures A. Accuracy of the response I. Diiiercntintion l. iunnn responses 2. Quasi-human responses 3. lucan details A. Quasi-human details CL Articulation l. Perceptual characteristics a. Size or physical structure b. Clothing or,hnirstyle c. Posture 2. functional characteristics a. Sex 3. Age c. Iolo 4. Specific identity D. Integration 2. Motivation of Action n. Dnaotivatcd activity 5. Inactive motivation c. Intentional aotivation 2. Object-Action integration v. .‘1. 11 f .w -.'."' 7F:- 13'. i fun—v.» .—'n-s-. if, ,. . -.e ‘-‘.t0 - ‘ ‘p-t ivy-or..- v. ‘up—F-p-vw-II— w—_ . D H‘Y'i" t v . w‘ ""'.'.‘.‘."'"Iv- 1.-. 8E3 ““ “I.“ m. 87 a. ihsion oi object and e¢t1°fi b. lneongrnen: integration of object and action c. lonspeciiic integraticn of object and action " c°fltrvent integration'oi object and action ' 3. Integration oi interaction with another object a. lature oi interaction 1. Active - passive 2. Active - reactive 3. Active - active 5. content of interaction 1. Malevolent 2. Benevolent . ,’w 11 _ “...‘ -—" V .V'T 89 The ilportanee oi the hunaa response on the lorrchaen h.. ‘.cn noted oiten in a variety of contexts. but generally with a Ilntau. oi theoretical elaboration. Aspects oi these responses nay have particular relevance ior the studyuoi; .the development of the con- cept oi the object and its iepsirnentl in psychopathology. This scoring systole is an attenpt to apply developmental principles oi differentia- tion. orticuletion. and integration (Werner. 1948; Herner & Xaplan. 1963) to the study of human responses given to the Rorschach. Bitiggensiation is dciined as the nature of the response with hulan content; Articulation is deiincd as the degree to which the response was elaborated. and Integration is defined as the way the concept oi the object is integrated into a context of action and inter- action with other objects. Uithin each ei these areas. categories were established along a continues bastd on developacntal levels. Hithia each category. ratings ranged iron dcvcloprcnrnlly lower to develo;mcn:ally higher levels. fificomrs or .I.\'M‘!$!S 1::3 Koontz: P2fif!;!".'7‘..“s .. SBLECTlOR 0F RESPONSES A. figggnVand quasi-hutrn rtrncqggi. All huaan and 10831 5033“ I“ and (“)1 responses are stored. human and quasihunan details are secred ii they 1) involve huLan 8Ctivity, e. (e.g., talking. pointing. struggling) or 2) involve . 8Cbszsntial ,c- portion of the card and are not just a stall rare or edge detail and 3) contain some description 0: explicit 56:33 or husenoid characteristics. Thus. independent of their 3°“‘3°“. the (ollouing responses would be scored: '0» iace...oi an old I» with wisps oi hair on a, “do. l. 9() L‘AA_ 89 ' a use with sunglasses on" ”a girl's head” ”a baby's face" “baby's hands with nittens on" ”face with a large hooked a..." ”iaces of 2 angels” ins es ns in sane rare instances. animal responses are classified as quasi-human ii the aninal is explicitly given qualities that only a huzan could. . have. The exceptional quality of this classiiication must be eaphasired. It is‘ng£,neant to include all responses scored Aainal flovcsent 73. Though the iollowing responses night he scored PM. they would not be in- cluded as a hunan or quasi-human response: Hanan-like actions which could be achieved as the result of special training and which night. therefore. be expected in the context of a circus act. Activities which huaans perform. but which can also be periorned by , ini-als (e.g., rubbing noses). The hunan content oust be explicit. li. for exanple. ”Bugs bunny” is given as a response. it is scored only if Bugs Bunny is engaged in a clearly hunan action. Thus. lugs lunny crying or talking would he scored as a quasi-huzan [ (H) ) response. Applying these criteria. the following anieal response would be scored as quasi-hucan: ”a hookah seeking caterpillar...fron Alice in wonderland.” “two drunken penguins leaning on a lanp-post...they're definitely .ICUh.‘ e 9 921 'two lobsters caning out of sisaloon...and they kind of have their area around one’snother.” 'sea gull...laughing.‘nahihg'fun of sozebody.” “two frogs...tcteoa-tete...tuo angry frogs. their months are downcast.“ 11. 8003130 PROCEDURES ‘oiA££2£2£2.2£.£h£.1££222229 3ltponses are classified as perceptually accurate or inaccurate (F+. Ft. t;. r-). r» or r: responses are classified as accurate and r— responses and 8; responses are classified as inaccurate (Rapaport. Gill s Schafer. 1945; Allison. Ilatt s Zinet. 1968). tow Iere responses are classified according to types of figures perceived; whether the figure or subject of the action are quasi-huaan details (Id). hunan details Rd; full quasi-huzan figures (8); and full husan figures. 3. l. £3553_§g§£22§g§, To be classified as a human response. the figure must be whole and clearly human. Examples: ”People“ “Ion“ u..byu ”African natives” 2. [92231;33§32_;ggggg§£§, Here the figures are whole but less than human or not definitely specified as human. lxacples: "rm-warms - ' - .-..- .. 92 ”Hitches" ”Dwarfs” ”Two opposing forces. sticking out arms and hands. Opposing forces. pitted against each other... inching at each other. ‘Hith conplicsted...of talons. appendages. arns raised in conbat...Persen naybe...standing there. being very offensive and attacking.” 3. human details. Here only part of a hunan figure is specified. Exanples: “bands strangling" “faces staring at each other” d. Quasi-huznn details. here only part of a quasi—huaan figure , is specified. Examples: "angel‘s face” “witch's head”. ”devil's face“ C. Articulation . lere responses are scored on the basis of types of attributes ascribed to the figures. A total of seven types of attributes are considered. These types of attributes were selected because they seen to provide information about hunan or quasi-human figures. The . analyses are not concerned with the sheer detailing of features or with inappropriate articulation. The analyses are only concerned with articulations that enrich a hunen or gunsi-huzan response. that enlarge a listener's knowledge about qualities which are appropriate to the figures represented. for example. a response which states that a nan has a head. hands, and feet does not enlarge the listeners' . g e 2 . av I .f“ .N to 'C _ . W Q’ee‘ s‘ ‘ e n: ‘ . 4. 1.; lb . ‘ ‘t ' .- a. .‘ -e l . _ ' . _-_——* AM A“ A-‘A‘_4 .- 4 .‘n‘ A!—~ “_ .— A “"““ ‘.‘"“ ... .u.. u _-_A h.‘_._ __.__ 92 knowledge about the nan. Possession of these fast“... .. presupposed by the initial response. "nan.“ An artieui..... such as “a use with wings“ is not scored as an articulatiog'g.. cause it is an elaboration which does not add to the.specificati... oi the buses or quasi hunan features of the figure. There are two general types of articulation: the articulation of l) perceptual and 2) functional attributes. 1. [grccgtual characteristics. a. giro or physical gtructure. for this aspect to be scored as carticulated. descriptions of the figure nust have adjective status. Thus. no credit is given in a response where an craninee only says that a one has feet or that a hand has fingers. Sire or structure is only scored as articulated if there is a Qualitative description of aspects of body parts or the whole body. Descriptions of bodies or body parts as ”funny” or ”strange” are not scored as indicating articulation of body structure. Certain aspects of facial expression can be scored as articulations of sire or structure. included in this category are responses like ”eyes closed“ or "south open” in which the description of facial expression anounts to something lore than just a description of physical appearance. Applying these criteria. the following responses would be scored as articulations of size or physical structure: 'alin nan“ “big feet” “the tsp of the body is sort of heavy and her legs are real. real teeny” *- : " Wtwmvmwuv .. e car-twice?” "firearm £34 1' I. I ... *. ‘A AA A... ”slanted eyes” . “china protruding down iron the “¢¢" “eyes closed” . ' ‘noutha open? . “tongue was sticking out” '1 ¢°atf¢$ts the following responses are 92; scored as articulations of size or structure: 'wonen with breasts“ "they're shaped like people” “eyes. nose. nouth” “woman doesn't have a head” “a pervert with bunny ears" “person with wings instead of area” b. flashing or haigsgyle. for this aspect to be scored as articulated. there has to be a qualitative description of acne aspect of either clothing‘gl_hairstyle. It nust enrich the description of the figure. Sinple nention of itecs of clothing inplied by the response does not enrich one's understanding of the figure and is. therefore. not scored as an articulation. Using these criteria. the following responses are acorable as articulations of clothing or hairstyle. “sons kind of nonstache..right above its nouth” “girls with ponytails” ”hair and the things sticking out of then. feathers” ’their°pants would have to be shintight and when they lean down. their jackets go pointing out. nakcs it look like a very tight jacket.“ "a couple of witches with red hats” mu-no- --- , 95 "wearing a black coat and e honburg hat. Black coat 1. go“ . , . i. of billowing behind hits..." , '...e full-tailed coat" “two little girls. all dressed up in their nother's things'f 'Gey 90‘s type wonen...loth wearing a long bustle and feathers in hair.” “All Merican Indian in sons ceremonial costue with wings and paraphernalia.” "a nan...with sunglasses on.” by contrast. the following responses would m; be scored as articulations of clothing or hairstyle: “m wonen with shirts on." “shoes on” e. [ogturg Posture _ia scored if the response contains: a) a description of body posture which is separate fro: the verb describing the activity of the figure. or b) a description of facial expression that goes beyond acre articulation of the physical appearance of features in that it con- tains a sense of novesent or feeling. Posture is £3_t_ scored if body posture is inplied in the verb rather than being separately articulated or if it is ainply a description of a figure's position in space (e.g., facing outward). thus. the following responses are scored as articulations of posture! 'aru flung 91". "head tilted” 'standing with legs spread apart'.‘ Wfi’?9."1'm' "P ".1"“'II’U'.‘HV“.3"Y.‘.’6‘;'1YI'.? PM.“ name. Mu 0- , ~ v - - ~ ‘m‘ " E96 9S “leaning on a leap post“ “shoulders hunched” “sonebody h“““““"“"'¥§f§.é9¥“' droopcd. fornless. shapeless” ‘ '1 “eyes look piercing“ “grittiog teeth“ “soiling“ ' the following responses are §g£_tonsidercd articulations of posture: “sitting“ . “standing“ “doing a high dive“ “back to back“ “facing outward“ “south closed“ 2. zgnctional characteristics. a. ggg. For sex to be scored there either has to be a specific nention of sex of the figure or an assignment to an occupational category which clearly inplies a particular sexual identity. If the final sexual identity is not decided but alternatives are precisely con— sidered. sex is scored as articulated. lf. however. the indecision is based upon a vague characterisation of the figures with an enphasis upon the sexual nature of the figure as a whole. sex is 22; considered articulated. In the following responses. sex is scored as articulated: “Han“ “Girl“ 3. “Hitch“ O I I g... t 3: f?“ .‘yf-J 2.1;: 7'3 ~‘L--.M;:. “3‘1“; R59». : 'f‘,{§{“"“.?'° . . fl}. I»- ' ‘5: ‘3.) 3Y1?“ .A 3"“ V V; x ‘g r J .- ”é“ . . ’ _ p ‘ ..‘ " ‘ “p“ .‘. . t - ‘.A. u. r. .' I'.‘ '; ,_ . ,‘ , \ . 8'- 97 “lather“ “Priest“ ,{. “v , “either an old nan or an ugly wonan" “2 boys putting on a disguise kit or a girl with he: nakeup kit“ - ' Iy eontrast. sea is‘ng;.ecored as articulated in these responses; , “Hell. these look like two hunan figures. 1 think when you look at the breasts there. they're girls. Then down here could look like phalluses. I don't know. It's rather ambiguous. confusing...protrusions fron the thorax. you know.“ “Looks like two people. Could be a wonan or a nan. l debated _this for a ninute. (scant) Hell. this fora could be uonen or.the costuning of son. (1) Hell. I guess it would be tights and sort of loose shirt. I don't know exactly.“ “Tun people beating druns in a way like both night he semen. la another way. like nan. Doesn't see: to be any real indication whether they are nale or fenale. The rather ex- tended cheots seen to represent breast of wonea and protuber- ance on botton seens to be leg. In these respects it has a bisexual appearance. There is sonething barbaric about the figures. Seeas to be something of a representation of gods or soaething like that. They see: to be wearing high heel shoes. both of the figures seen to be very awkward and look as though they're doing sane clunsy novenents in beating the druns. the heads also don't look hunan--look as though they‘re sons kind of bird's heads.“ £98 97 b. ‘55:. For this aspect to be scored. specific reference lust be wade to sons age category to which the figure belongs. Thus. age is spanned to be delineated in the following ra- spouses: ‘ “child“ “baby“ “old won-n“ “young girl“ “little boys“ “teenagers“ ly contrast. although sone indication of age is inplied in the following responses. the references are not specific. Thus. age is 332; scored in these responses: 0““. . “girls“ who’.w “priest“ c. .1213, "hen figures are human. a clear reference to the work a figure does (occupation) is scored as an articulation of role. ‘Hith regard to quasi-human figures. role is scored if the aanner in which the figure is represented inplies that it would engage in certain activities rather than others. Thus. role is assuned to be articulated in the following responses: “soldier“ u'umu) 1 —_ l "hen sexual identity is clearly indicated in a role designation. both sex and role are scored as articulated. Such a situation exists in tho following responses: “nether.“ “witch. “ “priest.“ 99) “fpsnish dancer“ “ballet dancer“ “Princess“ “bother“ “witch“ “devil“ 4‘ ”u'wx- .9" . “elves“ lole is‘22§,scored in the following responses because there is no clear indication that they refer to occupation rather than a nonentary activity. “dancer“ “singers“ d. gpecific identity. Bern a figure uust be naned as a specific I character in history. literature. etc.2 Examples: “Charles Decaulle“ “Theodore Roosevelt“ S-Wissflasien. This is the sinple enumeration of the total number of types of features articulated. In the preceding section. seven types of attribution were described (sire. clothing or hairstyle. posture. sex. age. role and specific identity). Thus. for any single Rorschach response. a total of seven types of features could be articulated. The average nunber of features taken into account in each hunan or quasi- huaan response constitutes the score for the degree of articulation of individual figures. If. for .xsfibii. a subject gave four hunan responses "a...ln.i To the degree that age. sog:§:hg‘occupation are clearly indicated in the specific identity. these fehtures are also scored as articulated. Thus. in the response. “Charles Decaulle.“ sex and occupation are specified. Such is not the ease in the response. “piglet.“ 100 w—*---’—- 'v -._.._.__,...._{~ ‘ __.__-._- .~——~~ “.. and attributed a total of ten types of attributes to;rh... 51. score for degree of articulation‘is}1.$. D. tio ‘.L 94'2‘ integration of the responaa.was’btdred in three ways: a) the a..... .‘ internality of the notivation of the action (unnotivated. reactive. and intentional). b) the degree of integration of the object and its action (fused. incongruent. nonspecific. and congruent). and c) the integration of the interaction with another object (nalevolent- benevolent and active-passive. active-reactive. and active-active). These analyses can only be applied to figures engaged in hunan activity. 1. flgtivation of action. ’ The articulation of action in terns of native inplies a develop- Iantally advanced perception of action as differentiated fron but related to the subject. Moreover. notive can be ascribed in two ways: as reactive or as intention. Reactive explanations involve a focus on past events and behavior is explained in terns of causal factors: one assunes that. for certain prior reasons. an individual had to do a certain thing. by contrast. intentionality is proactive and inplies'an orientation toward the present or future. The individual chooses to do sonething to attain a certain and or goal. The ability to choose between natives and to pur- posively untcrtake an activity inplies a greater differentiation between subject and action than is the case when an individual is inpalled to take an action because of past occurrences. Tor this reason. the analysis of action‘will consider whether or not a native was provided.aed;whether the notivation was reactive (causal) or intentional. 101 100 0- Wain... lore action is described with no explanation of why it occurs. Examples: “Two people kissing each other.“ “Bones looking at each other.“ “Ben leaning against'a hillside.“ ., W lore perceived activity is described as having been caused by a prior situation (internal or external) and the subject is seen as having little choice in his reaction. Ixsnples: “A.Gernan soldier on guard duty. 1 think he sees something and points his gun at it.“ .“Arabs recoiling from an Israeli boob.“ ‘VA person afraid of a snake. standing on a rocky cliff 'with arns upraised as if he's going to hit it‘with sonething.“ “Two women struggling over ownership of a garaent.“ cw Tor notivation to be scored as intentional the action lust be directed toward sons future noaeat and the subject nust be seen as. in some sense. choosing his action rather than having to react. Examples: “Halloween witches. naking incantations over the fire. in preparation for all hallows' eve.“ “An.orchestra conductor. his arms raised. about ready to begin.“ i. ‘1‘ 1' - °' d ,.1."‘- 0’ "~' (‘0' v-t ~a'. s- ' ,A ,. 'I“&. ”-1-; ‘g". -v-,“i‘...-v\ ‘- awrb"._o-‘r . 1‘.g“o'. 5" ‘7; ‘3: ,. .113 .3“: 3‘ 4‘. : ‘Y ' “ ~, f ‘ g.‘“~ "" .oa ‘ - " :7. a. .‘ '1 - 1. , e ‘ .. ’.' _. . . f . .‘ h 3‘" "2:. -_ I" 'AV.‘ “_, ‘4‘ --“3.- . ~ . ’.1 '~ fl ' - .' ~ . (3W 1 9‘1”“ .7 ‘- '- " a H‘ ‘ -",.o "‘ .34’ I' . . 1()2 . .-.. e__. __‘__ ‘AQAM_‘ A 2. ect-ac n e t on in this analysis. four levels of integration of the object with its action are distinguished (fused. incongruent. nonspecific. and congruent). - a. Eggion of object and action. for a response to g. ‘nclu‘.‘ within this category. the object must be amorphous and only the activity articulated. In such situations. object and action are fused. The object possesses no separate qualities of its own. it is defined only in terns of its activity. This type of response is exemplified below. In both instances. nothing is known about the object expept what it is doing. lxanples: “Two opposing forces. sticking out arns and hands. Opposing forces. pitted against each other...looking at each other. filth conplicated...of talons. appendages. arns raised in conbat...Person naybe...standing there. being very offensive and attacking.“ “figure there with hands. standing with legs spread apart. reaching out with hands as if trying to grab sonething.“ b. ‘nggggrugnt integration of object aggfactiggL Tbr a response to be included within this category. there should be none separate articulation of object and action. Sonething must be known about the object apart fron its activity. levertheless. the activity is incongruous. unrelated to the defined nature of.the object. The articulation of action de- . , . tracts from. rather than enriches. the articulation of the AA -‘-———A A-“ -A‘ M“ - A .____- 103 .e object. Isa-plea: ' “A great big noth. dancing b‘llet.‘ “no figures. one half human and one half animal holding we 6 - two sponges.“ “A.littlstaby throwing a bucket of waggg.' “A.satyr-thing bowling.“ “Two sphinxea pulling a decapitated woman apsgg,' “Two beetles playing a flute.“ c. Eonspecific integrating of object and actiog. Inclusion‘within this category also requires some sepagggg articulation of object and action. however. the relationship between the two elenents is nonspecific. The figures. as de- fined. can engage in the activity described but there is no special fit between object and action. "any other kinds of objects could engage in the activity described. Thus. while the articulation of action does not detract from the articulation of the object. neither does it enrich it. Exanpleot “One big person standing with arns raised.“ “A knight. standing ready to do his job.“ “Cavemen leaning against a hillside.“ “Two figures dancing.“ 1 “Two older women trying to pull something away fron each other.“ “Two nen fighting.“ “A nan running away.“ “A person. sort of a girl. standing on her toes.“ its; < WW7:— -.--r_‘__ -_ 104 103 d. n ent nte ra ion o ob t a ct on for a response to be assigned to this category. the nature of the object'sni the nature of the action oust b0 articulated separately. In addition. the action oust be . particularly suited to the defined nature of the object. Iy'way of contrast with the preceding category. the action mat not only be something the object night g; u must be something that the object would be especially likely to do. There is an integrated and particularly well:suited relationship between the object and the specified action. lioreover. the articulation of the action enriches the lungs of the object.3 3. t rat o f in erac ion w th an e act. a. nature of interactigg. This analysis applies to all responses involving at least two human or quasi-huaan figures. In addition this analysis can also pertain to situations where a second figure is not directly perceived. but its presence is necessarily iaplied by the nature of the action. l; Activefipassive interaction. Two figures can involve a representation of one figure acting upon another figure in an active-passive inter- , e In situations where the role definition of the object amounts to nothing nore than a literal restatencnt of the action. object and action are not considered integrated. Responses like "dancer's dancing. “ or “singer‘s sint‘flt' "¢ '¢°"‘ as nonspecific (level 3) relationships. however. responses such as “ballerina dancing“ or “character free a Rudolph falls opera. singing“ are classified ll l congruent (level ‘) relationship. \f “I \ .‘ . .' I :. i, _. , ‘ ‘ ‘ >¢ “ ‘. I” ' ':_‘ ‘ .,_ ‘ . Z‘. I . '-.‘< "r... .- 1‘ _ I; L‘. ‘. ,“. "... ' . .. ‘ ‘ 1 . .V . , , I . .. I -1 ‘ . , . a. . 8"". ' Km" 4“; .cc‘ l'é . O b a .' . 9‘, . "'7': . . T I _ o ‘V / ~ I’ . I a . I ...- . 5 . . ‘ - . .. ‘ , ‘ "“ ‘ 2" ' ' ' . " "' '1 on..." . 105 A u-“_‘A A- —A— .uI-‘A A“ action. One figure is active and the ‘u A— u“ a. IO! other entirely passive so while acted upon. it does not respond 1. .3, v.’. 1. Active-reactive interaction.r ls another type of interaction the figures nay be unequal. 'I“‘e. t ... One figure is definitely the agent of .‘lh '82'.’ a the activity. acting upon another figure. The second figure is reactive or responsive only to the action of the other. This is de- fined as an active-reactive interaction. 3. Active-active interaction. In a third type of interaction. both f igures contribute equally to the activity. and the interaction is nutual. b. Content of interaction“ l. flglggolent: The interaction is aggressive or destructive or the.results of the activity inplies destruction or harm or fear of hero. 2. benevolent: The activity is not destructive. harnful or aggressive. It nay be neutral or it nay reflect a warm positive relationship between objects. v—i A Attached are exaoples for scoring both the nature Iotations in the left hand nargin indicate scoring in action. [Active-Passive (A-P). Active-Reactive (Ari). locations in the ritht hand nargin indicate the scothd‘w interaction [Malevolent (x) and benevolent (3)]. .n‘ u..‘..‘ of interactions. r ‘5. .gture of the inter- ... gggive-Attive (a-a)], ‘., gha content of the A AL_- [stage of lateragtiog A-P A-P - A-P A-t A-I A-P A-I A-I A-I or or a-a a-t Ask A-A A-A A-A . A-A 106 '- -..-..-‘ - 105 Integration of Interggsjgn Regent of era on A couple of under'tzskers’ lowering babies into the p“. ‘ A prostitute rolling; drunk. It Crucified oath-H” U ,izt. . . II A nother holding-out her an and telling her kid never ' to eons back. Two sphinxa pulling a decapitated woman apart. n 2 no people kneeling down with hands extended toward and f touching other people. I African natives beating a dnna.-~liartians applaud.. he being tempted by a snake (snake seen on card) If Two people with hands up as if trying to ward off the two people casing to get them. Two guys with black capes... coming in to get the other people... H German soldier - think he sees something and points gun . at it. it An orchestra conductor. ants raised. just about to begin. A man running away. I! A woman crying out for something...two forces pulling her apart. one is depression. one is suicide. a A man trying to kill a little girl. who's running away. if A woman with a child looking up at her. I Soneons having intercourse. a non child and a woman child. trying to sake love but not knowing how. I One person there is pointing and the other is listening. I. Two people and two. nartians fighting. '1 -- ,uv- 107 Eturs of lntgrgctiog A-A A-A A-A bu...“ -.-——&:-“£—;—- 1C6 integration of lntegaction flat out of lgtggactioa in .:ac" the women having}. fight. calling each other nanes. M gremlins ready to'hit each other. Teopls pledging hands together - like victors. walking along like that. . ' I . . O I .37 . .':l m “"up‘ti ' ‘ '4‘ . ,.‘e k. ‘. V! v t.."‘3. .qs 4-4'5.“.‘ ' s .. ‘, (d . . ‘..‘l 1,.“ :t .5; 19.; Law}; ‘3'. 9' "'. $23.“; i .0 ‘ xx! 3:. :7 t :1 R 9‘s“ -\ '. . ' ‘ f . '._V“ k: ‘I.£ “e "1?... "‘J ‘ I" fi't!’ 'e x «a. '- nt 0 . I l- e'. e i .5. 2‘? ' v: -‘t{%':* ‘ - 3 I '1 ‘ z . ”‘_ _ I ‘* 7&fl3h. ‘ ' \ ‘1 (lat IO \’ i“ ' n . ‘ 'v .1 108 corin Outline o ies of 31 l. Accuracy of response (Ft or F-) ll. Differentiation one. of figures perceived) (1) Man (1) Quasi-hum“ (3) III-an detail (4) Quasi-human detail 11!. Articulation (a) Perceptual attributes (l) bias or physical structure (2) Clothiu or hairstyle (3) Posture (b) ”notional attributes (1) Sex (2) At. (3) hole (6) Specific identity (c) Degree of articulation U features articulated] 0 responses) 1'. Integration (a) liotivstion of action (1) Wotivated (2) leactive (3) Intentional 0) 3. integration of object and “no. a) miss of object an: lotion a, West action (3) Ion-Smith action (0 arenas: action 3" I ". ._' . “My; 3. ti. , ~ Se‘b "- “#1 Sl’ - D"). or"! ‘. 109 -—.._._... mm '7 . , I -._ A‘ _-— I I h .CI_--l~ M h~~.‘—-m *o.‘ “..*’.~h .108 (c) integration oi the interaction with another object (1) Iature of interaction is) Activev'assive 0) Active-reactive (c) Active-active (2) Content of interaction (a) liaieroient (b) icnevolent ‘. D Ida-low l09 Summa : Structural 0b ect Re resentations Scorin Outline Categories of Analysis l. Accuracy of response ll. Differentiation ii) "Mann (2) Quasirhcuan (a) lit-an «an (4) Quasi-human detail . ill. Articulation (a) Perceptual attributes . g (l) Size or physical structure (2) Clotning or hairstyle (3) Posture (b) Functional attributes (l) Sex (2) Age _‘(3) Role (4) Specific identity 1V. Integration Io action (a) activation oi action (I) Unnotirated (2) Reactive (3) Intentional ‘5’ 1‘. integration oi object and action ' (1) fusion of object and action '(2) Incongruent action §£2E££ F+ or F- (l-i). total, subcategory (i) (3) (2) (ll (0-7). total. subcategory (0-1) (0-1) (04) (0-1) (O-l) (O-l) (O-l) (ObIZ). total. subcategory (0. on all integration scores) (1) (2) (3) (l) (2) 11.1 llo (3) Non-specific action (3) ,i4) Congruent action ii) (c) integration of theiinteraction with another object (I) Nature of interaction (I) Active-passive (l) (bl Active-reactive (2) (c) Active-active (3) (2) Content of interaction (a) Halevolent (l) (b) Benevolent (2) lotal (1-23) J Staff: "is“: -’ 7' ~..~'"'.-.'f'i " Wit-sitar flaw APPENDIX 3 ._‘v-. r—‘_— A‘ ”H mm. APPENDIX - A” ~“-‘“‘** 3 111 Alan Krohn. PhD oaaccr-rerncscumxon SCALE roe. owns" GENERAL lliSlRUCTIOiiS: first. read over the scale. lhis scale attempts to register the sense the subject has‘of objects in his world - how primitive. shadowy. malevolent. anonymous, stereotyped. or one- dinensional they are on the one hand. or hon full. defined. feeling and ' in general human and complete they are on the other. iour task is to infer this overall quality of the patient's inner object world from his dreans. let the dreanuwash over you. let your clinical “feel” tell you how human. how real. hon whole. and how nature are the people as the dreamer experiences them. lhe scale rests on your global sense of the dream and there- fore l would like to encourage you to use your intuition and empathy in any any you can. ‘Each point on the scale is defined in three ways: a global description of the nature of the object world. some typical characteristics to look for in the dream. and a sanple dream. I would like you to rely most heavily on the global description of the scale points and the sample dream that illus- trates that description. The typical characteristics are features of the dreaus that tended to correlate with each of the global descriptions. lhe criteria should be used as aids to your intuition and are not intended to be used instead of your overall sense of the dream. Indeed. if you have a clear sense of the dream's location on the scale and the dream does not c ntain any of the characteristics. simply ignore the typical characteristics. hese characteristics are intended to be signposts to supplecent your intuition. not constrict it. SPECIAL SCORIKG “DIES: l. Hany dreams contain several characters. If a dream contains even one character that clearly meets the criteria in categories 2. 3 or 4. give par- ticular weight to that. _ 2. If the dream seems to have two almost equally important portions with very different kinds of characters in each. give a score for what seems to be the more salient quality of the dream and then a supplementary code for the quality of the object in the secondary portion of the dream. 3. Dreams that contain no peeple (other than dreamer) in then present some special difficulties. Handle them as follows: first. indicate that there are no ebjects explicitly represented in them by the code no. then try to code on the object representation scale by asking yourself if 50:3 inanimate obJect or animal in the dream seems to have human qualities. If something seems to represent. with some subtlety. a person. then score the dream in category 6. lhe final code would then be NO 6. These non-human elements which stand for ob- Jects 93! be buildings. the overall setting. animals or even place names. if nothing in the dream seems to represent an object. try to capture the overall object quality of the subject from the global descriptions in the scale and assign a code from categories l—S. ' ‘ As well. the sESlc has been applied to the Early Memories Test and Rorschach lest productions and by therapists to their longoterm psychotherapy patients (Krohn. A. and Mayman. fl... Bulletin of the l-ienninmr Clinic. W74)- 1212 113 - ~—--a-—--—-_.-~. .— -.._...— ‘ A ‘ ‘ *" - '-"-._.D "-‘c ”ll.- _ Page lwo ll2 SCALE POINTS: l. The subject's world seens to be completely "'9‘955- "C3“. lllcn. strange° it is a world essentially without people: he experiences the world as very start; and static or very fluid and fomlcss; in short. the world for him is an unpredictable desolate. often strange and bizarre place‘ that he only rarely understands. ' for Example: l. The dream is virtually devoid of people 0" “WW-"lie figures; if people are present as drean figures at all. they are nullified. extremely vague. and inci- dental to the action of the dream. ‘ 2. lhe dream setting is either very fluid or almost frozen (or both). 3. As the dream is read it nay 9W0 3'00 0 “15‘1“”! 'othervworldlyf unsettling feeling. Sample Dreams: I dreamt that l was walking into a forest. the birds were singing and the sun was out. then all of a sudden the trees. rocks and bushes began to melt. all run together with the ground. just everything all combining. ‘fhen the animals came running out of different places and caning after me and they too sort of started to dissolve into eacn other. to celt together. l'n walking underwater and Lind of half swimming and half walking through the water. it's all around ee. this fluid reybe a little thicker than water. and there are large. whale-lite thin;s. but sreller than whales - someone i told the dream to said they seemed lite they were like sperm. and as I walked through the water they part. just ahead of re. lhere may be more but that’s all I repeater of the drean. 2. The subject's internal world inclufes people, but not really alive. human. benevolent. Peeple are insubstantial figures, prone to seem “figment. brutal. nurderous. extremely cold. technical, less than hum". lhe subjezt's ”"59 of people seems bizarre and distortef. flora specificaIIy. it seems that under the pressure of morbid. sadistic. careers-as impulses and fantasies, people becom transferred into calculus. aninzlistic figures. There is no internal experience of real interaction anon; 5.20;“, for others are cherienced as little nore than the subject's own primtue icy-41595 incarnate. for Exauple: l. A figure in the dreal is half animal. half human; animal turns into a person. or vice versa. . 2. A dream character is de-dJfir-E-nabset to die. or killed - each in either a bizarre. brutal. very explicit or m-bud fashion. 3. A dream chant!" 0' "‘9 "0"“ l5 daing something which is bizarre. morbid or terribly 5W”'° 4. A person appears 1" the "n" ”*0 is labeled a witch. devil. robot ia malevolent. cold. 00“” ‘9‘ ““33““5 bizarre. hunanoid figure). Samle Dreams: . 1 was mine i- I mf'hgfih'nh ”her People in it. a. from the library was there. Others 6‘7. , t ‘3‘“ w“ '0 b”?! dnd asked if they could use the living room to study in. I 4 '4“ i“ ‘0' some reason i consented. They m irl looted lite a car 5 were ugly new“: . “not . , the ‘ " he had a nustache and short. _ he just I 9‘9 P “C. “strain , along hair “L in“ ”I, m“ ”, thcy “a“ s 9 they could stay 1 said no. . ta . . 'fhen a. came 7 until“ ‘- 9! m a mu. 9. 3' "my were Just about leaving _ , fl. i knew she was dead already. «J . M . kt um . ' tier body ”$31.91*”? mg to h" grave. She kissed me goodbye and t I“. e) {;‘.;‘.5-; g, I”, ékflgfi'v _> " :2”? . . .. A“ v.» -". 1 . -..,.'._ . .- gge .. . _ , {"3" K . ‘6 ~51”? ' . . t ‘ w" i . -2' r I - ‘- ‘a e o Is' . » 11L4 _ “flu...“ .—-——n-- -‘fi-.-...-—»"‘ ‘ Page Three “3 trotted out in a little gray coat. She didn't know it but she was 90ing to her grave. l renenber thinking whether they would make her lie down in the casket and just close it or would somehow put her to sleep first. l was with a grOup of people in the front yard of a big white house. we were just sitting around talking about hen nice the weather was. i don‘t remember how. but all of the people began to loot like animals. My friends were some of the animals. 1 could thll by their voices who they were. there was a gorilla. (my boy friend). a zebra (a good girlfriend of mine) and various other people that I know. I didn't know all of them though. but they were all animals. I remember thinking that they were going to kill me. but they hadn‘t made any attempts or anything. i ran away into a child's playground and was playing on things there and l was coming down a slide and when 1 reached the bottom there were two anirals whom i thought were waiting to 'get' 09. SO i ran and 1 saw a construction site on the side of the road so i jumped into this big hole which had a big pile of dirt right next to it.i remenber right after l Jumped into the hole i thought that if I stayed in the hole and 'if they"found me they woold bury me alive. l climbed Out of the hole and ran again; this time I ran back to the house where it all began. l rushed upstairs into a bedroom and sat frozen on the bed. i rerember hearing some voices coming from downstairs so i opened the door when i heard ”them" coming up the stairs. I ran to the stairway leading to the part of the house where they were at and the animals were people again. They had lOn; rnives in each hand and my boyfriend was leading the group. i ran back to the door and ran outside. It was raining, the sun was shining brilligptjg'and these little animals were climbing .all over me. clawing at my face and scratching my arms and i tried to get them off me. but every time l knocked one off there was another to take its place. 3. feople are experienced as insubstanial. fluid. core or less interchangeable. but are not experienced as calcrolent as in *2. lhough people do not seem bizarre or aggress against one another as in previous categories. the subject experiences others in a vague. fluid and undefined icSthS.th subject cannot really artic- ulate what someone means to him. because he has such an undifferentiated concept of what other people want. feel 0'.¢0~ Shit subjects ray be unsure who did what to whom. for his internal representations of Other people are so unstable. diminshed and distorted and because his sense of his cnn baunderies and the boundaries of other people is SO POOT- lhis C“°3“'¥ '5 it?! similar to 52 but different in important ways too: the subset! rfifirifnif3 Others as confusing. fluid and without any enduring char¢€i¢"§“",h? id" éfik on. but ngt_pervasively malevolent. weird. mysterious or strec;é- llilfhlrs :5 this kind of subject describe an interaction. one would C€9ilnualll "‘3? rc:~;r {laératt contradictions in the way other people are depicted; lhis arise: cut of the subject's limited capacity to form a meaningful '99“an °f ‘“*‘h“' F?'S°n. leaving him with a set of moment-to-moaent iflpf¢55‘°“$ ‘hai ‘“" ‘0 (ivlure what is enduring and salient about other people. for E:amgl;;rson changeS into another person. comes to i of £80 P”p °‘ is a gomglgztaggect of . dream character chcng 3: A dream character “ ‘ ‘°‘t‘“"'°“ °' resemble another. as in an unreal way in the dream. two innocuous. stereotyped. distant Public "9u'05 ° l"W'n gremlins, a , J*i “Cti2nazno;nimal uetamorohoses in some way. ctors ~o‘l . zarre ways. in inexplicit ‘Hi u.ually not including gory details. 'Vi' ' . rson is dead or killed. lug 5, “fir-hi lad" E "tugs” not id'ffllflt of the dreamer. ». i .t, . r-..:‘::‘v—_. ", 3‘").. a- -"_ . “a" ‘ . a '\ (1‘3. ;" ‘ fi' . “f .‘ . 3 "it‘ll ‘6: T , ‘-- . "an .. .3...- . .11.“. ,3.~.:1.~.-' ‘ -.. .- 11.5 Page-Four ll4 Sample Dream: 1 dreamed l was coming here to sop n, therapist and l walked in and he was sitting there as he usually does. but he was different because he was just like he usually is up to his shoulders and neat. but from his neck on he looked like a dwarf. as if his head were very small. 4.1he subject's experience of people is_t0 a very 9"?at cxteni "5hi0ncd around the need they can Quite directly gratify in the subject and/or around the needs the subject can directly gratify in the other. in this sense peOple are experienced in a very incomplete way: aspects of the other that do not bear directly on the exchange of gratification are only Partially perceived and ""derSt°°d' P”t another way. the tremendous inportanse of other people as gratifiers ‘9343 the subject to be only vaguely aware of qualities of the other that exist apart from its need-gratifying function. for Examle: l. Very minimal interaction of a drean.character with the dreamer. and what interaction there is involves the drean character satisfying emotional (in- stinctual) needs of the dreanzr endior the converse situation. 2. People interact predominantly on a feeling level with the dreamer. without much explicit interaction. 3. Dreamer is involved in self-directed activity exclusively - preening. ad- miring own body. practising something - with others watching or absent. Sample Dreams: - fly girlfriend B.and I are at the carnival at night with another couple. we are walking down this dark alley and i now notice how really sexy this other girl is. As if by my silent cugmand she walks over to he and embraces me. She is rubbing her beautiful legs up and dawn on mine as we stood there in the dark.- 1 looked up to see what Becky would do and this other guy came over to her and started to seduce her Bob. led. Carol and ilice-style. I got very mad and started punching this‘other girl; the drean ended there. ‘h‘ 39‘0““ “'93” "‘5 ‘b95t i ran"; girl in the hospital. The hospital beds were very comfortable and the food was 933:. lie girl did not seem to be suf- fering ouch. The girl's mil-9" was WU him! and very concerned while the girl was in the hospital. when she left the hospital the warm feelings faded. lhe mother and daughter went to the pier an: there was a lot of noise and con- ‘““°"- The ”other "35 being “to Pied“ tr ¢,r1n Offering her perfume. The girl kept thinking how happy she was at th~ hospital with the clean covers and the pretty. colorful bedspreads. 5. The subject's world is exocricn:id as DCleated with other people who are neither fluid nor massn'ely distort” U 913”; Integrated effects. but who do not have real identityo There '5 '7“"39 ‘LJ: people are more or less interchange- able ior the subjeCt- People either 59*“ 'try Shaded; and their motives are unclear to the dreamer. or else lb?) fife Gltcricnced in stereotyped ways. People do not really lute sense to the sthect: for example,he hears them but is often unsure 0‘ h0' the! "' "““'”9 95“ i"?! are saying; he is either deaf to what is implicit in what they "0 5‘f‘“? °' rcads a great deal of implication and hidden messages into what they Ore $01}fls- lhere is no real depth. specificity or uniqueness to the PCEPlO ‘“ l“? :bchCt s world. his world seems to be pop- ulated Hith”P“5°rs'hy' ’0 to‘spea‘. ”ho citlcr “iiier little from one another or who fall into 00¢ 0' ”"" '{9 5° ‘vvtr‘icial catcgories.lhese subjects 35y try to use M931! “‘“nuu‘ ued. at “‘3 ’sfiibolic" notions of people. I ‘u \ ‘ ’ .‘ \"- “' .y.‘ .. . . "a 1‘“ , . a~mr.- we , m,'“Y=€‘.-¥?Z'W'-3' v . . '$:: .i- . *"gr I '.‘ j e’ . . I . '. ' a :isa»--m z. ... 5« ..-.. -. a a a . a! . ~ 0 a; ~ ."'\“ " L. I. ‘H? 3. "' ' Fr —1.. l.'-‘ r'. .,..,. 1 , ,, . ; - "t “Vile-xi- " .- J: . n \ , . M g... :. :w' ... .vH-j fi' “may “HIV”: - .19.«1m‘ dy~k wm’anhfixifkeummatnhw.tfirxv F ' cl \ v I 1116 fihw Page Five "5 for Example: l- If peOple known to the dreamer a pear in the dream, the or involved in no thought. feeling or intention; the; may be d0"‘9 things. {at :ith no sense of goal or intent; there is no cgplicit interaction with the dreamer; people may even talk to the dreerer or rfininally interact with him. but the inter— action or conversation is really only part of some action in the dream. 2. All the people in the dream are e anyeous or nearly so. 3. Huch said and done by others in the dream SCCNS unclear, vague. non— sensical. “inqilicit' to the dreamer or highly syzbolic. 4. People in the dream a d 'h d as innocuous stereotypes 'laz me .' 'ltmlberjaclt types“). re cu“ c ( y n Sample Dreams: 1 was walking or driving down a street that seemed like c. road (near where I grew up). lhe houses were very ordinary. l especially noticed oany men standing around (i was alone). lhe can all seened to be wearing blue denim pants and shirts and cowboy hats. lhey all seezcd lite truck driver types. Soon I was with someone (I think ). probably my friend c. (forele). and we were driving away from them, but we still saw so..' along the way. lie drove on a winding country road 2 sunset. it was hilly and green. thhoagh the guys didn't bother us we were fearful (or i was) of them. A male person (i think it's he) is captive by a group of other males. all wearing white coats. lhere is a discussion of when to ask for ransom. Someone said at night. because newspapers wori faster at night. Then they decide to execute. Someone says it would be too obvious if the whole group did. because. everyone could tell who_did it by the unite coats. so one person is chosen to do it on his own. Someone says to do it in the day because “at the newspapers. the work lights shone brightest at midnight.“ lhen someone else said. “But it's only 4:30 in hlgeria.’ a a a a e a In the categories that follow the subject has a much richer experience of people, is more tuned into their needs. motives and individual differences, People become more defined and therefore core unisue and individual for the Subject. fantasies. fears. needs. guilts and conflicts asang them all have an irpect on the subject's feelings towards others. be: others alnays maintain a basic humanness and wholeness within the su:;c:t. Categories 6-8 try to tap differences in the amount and quality of internal catrerce with objects. Beyond feeling whole IHd human, how much range does there see: to be in the subject's inner reper- toire of objects? how ru:h and hen rtadily does the subject consider the feelings and implicit ains of those around hit? in gehcral. haw Subtle and differentiated is his experience of others? . 6. The subject ”3‘ ' ii'“ h°id °9 °53°“‘ ‘"d gcncfally CCnforms to the description above. however. he does not readily. whether for defensive or characterological reasons. try to und='$‘¢"‘ ‘“° ‘““°' “P°"°“‘¢ °‘ other pecple - their feelings. thoughts. wishes. ctC- 0“! i° “9”'°“‘ PVQCOOCCPiions and preoccupations. people do not feel to the 50559“ ‘° be ‘5‘“ ‘° ‘“‘"'°C‘ in caSY. mutual fashion. neurotic conflicts lead the subject ‘°"‘”‘ ‘° 5* Vii“ "COPlE. but in a parallel act- ivity or at a safe distance fra- OtherS- This tree 07 Subject is dealing with his conflicts by steerHD (‘93' 9‘ °ih9"- 'V0‘6‘09 the mare intense involve- ments particularly. that bring his conflicts to the lore and cause him dis- comfort. lhis subjcC‘ 'lght 5' ‘iit1°.'“°'° °i 3"! intense conflicts. being only in touch with an overall sense that his life is not as full as he wants it to ku,. ____.__ ..._. ._.—— “fl- "' " W'— " —".:"..T_'. — v _ _ __ __...._.. ' I _ I“... Mm 1:17 Page Six llb for Example: l. As you read tledreaa you get a sense that someone Specific is being represented in the dream. far less interchangeable than in lower categories. The dream characters OPE. in a global sense. believable. There is. however, very little explicit interaction of the characters in the dream and virtual]; no account 0‘ the thoughts. feelings or wishes of th: dream characters. 2. when people do interact in the dream it is fundamentally parallel activity (everyone painting. walking. etc.) rather than such mtual. cor.- plementary interaction. - ' 3. There will be some dreams in this category in which 22_people appear. but in which some inanimate object will lave object-lite properties. 0r. sorething about the setting (its detail. degree of differentiation.the presence of things done by people or made by them. etc.) which contains or implies this level of object representation capacity. Sample Dreams: 1 was in an apartment where apparently i lived. i lived there with several people but J. is the only one i can rerezhcr. he had a fight. He decided that he would move out. He took all of his stuff. and all of his books. many of which were mine. I was so relieved that i didn't even notice about the books till later. He had a small electric saber saw that he was cutting up the bookshelves with. Then he left; and i got an electric skillsaw. a considerably bigger one. and did the sat: thing. So: sane astcd re why i did this. and I said that it was because the saw J. had used has not big enough. (lhe follcwiig dream is particularly illustrative of Criteria 3.) “the dream concerns the nursery school I went to when l was 3 and 4 years old. 1 an coming back to it as an adult. l real life I have a violin and a classical guitar. Dream scene is at the nursery school and I an trading my violin with someone for a better one. I then play the iHStrenant _ much better than in real life. I then pick up my guitar and play it beautifully. though 1 don't know how to play it in real life and am afraid to try. Then 1 see myself in the house of the nursery school. I note the nice children.S toys on the shelves.‘ The house has many connecting rooms (as in real life) and the waod of the Hay]; and floors is very interesting because it forms patterns. Very thin plants of wood. lhere are objects covered with tar.:lin. Sc?33n2 says the school is closed Up now. The little Children don't cane there any;3r0_ 1 think about the plants _ haw beautifully they decorate the place. 7. The subject experiences P809}! Hith I good deal of sensitivity and acuity. lhey are unique. varied and rather well defined for him. Hg is aware of im- portant. soue differences among FCOPie. both ln terns of changes in magent- to-moment and dayoto-day $0935,¢"3 Oii'iuiii. as well as overall differences in character of those around hin. feople seen to be central to the subject's inner life. CV8“ if “’“'°"‘ conf““‘ ‘f‘d 5?” ‘° experience then in childish. transference—doninated ways. "9 ‘9633 to be in touch“ with people affectively. for [x le: IT‘Srean characters explicitly interact Uilh each other or with the dreamer. The thoughts and feelings of at least one character besides the dreamer are . notedé. '0“ 92‘ the sense that the dreaa characters are defined sufficiently n" that you might remain W 0' “We 0' the- in another dream (besides than by their nalrsl- 1' 3. ‘h. dreall in this catcsorr .3 HE' as those in #8 are very vivid. and the relationships mum ‘Wmm. mush peppered with the absurdity 2'7. .,4 at, . I r .I" ' '7 .e '- ‘r I ' 7" " ‘1v " W “on-- e-Wmfi'WQ‘c-«V' “1"“: . - - V ‘ 1:18 _...-.Q ‘ - ‘_A ‘L‘A ‘ A _A y ————— A‘ S Page even “7 of dream consciousness. is generally cogent and believable. Salple Dreams: ' "‘5 in ‘ ‘3'93 department store with my father. anther and youngest daughter. My daughter and mother had separated from dad and I. and we were just walling around. talking. and looking at various items in the store.we must havg £33" in the hardware section. because I remember looking at certain tools. etc. Hy dad looked over a wheelbarrow which he felt he needed. i remember thinking that he was too old to be purchasing a wheelbarrow. as well as the fact that he probably wouldn't be using it that such. we also looked and priced bathrosm 'OC‘lllieS. because he was thinking of adding another bathroom to his howse. for some reason we did not aeet my mother and daughter until we returned home (my parents'). I remember coming into the house and looking for them. fiy mother was in bed complaining of a backache or something. and my daughter was standing nearby. I felt that this was a way of telling me that she did not enjoy looking after my dau; ter in the store and that it had “got her dawn.' I dreamed l was getting married to a guy I know (but not all that well) named A. l know we went through the ceremony. which is rather vague to me now. "hat is vivid is the wedding celebration. It was a sit down dinner held in a large room.of some building (probably a hotel). lhere were a lot of people. Before the dinner 1 reaenber throwing ny bauQuet to the single girls over my shoulder and yelling 'byork.‘ My sister Debbie caught it. it didn't fly far. only to her. She was in front of the group. it was a white bouquet of flowers and l was wearing a beautiful white wedding gown that made we feel beautiful and happy. I threw the flowers to the girls who were surrounded by other people and also the orchestra....lhe neXt part was dancing (the next item on the agenda of a wedding reception). A. and l were to so in and “start the dancing.“ i picked my skirt up. gathered it together and we started walling towards the room. I then looked at A. in a quizzical fashion. asking him if I ever danced with him before. and then said. oh. yes. that one tire at Hargie's wedding (a real incident that did occur). l remember feeling sort of shy about the whole thing. lhen we were at the entrance of the social hall and cy parents were at the other side of the long hall that you had to come dcwn before yoo got to the room. lhey were telling us to first wait for them before wc began dancing. it was exciting and typically Jewish. i woke up so adhere around here....also i didn't know anyone at the reception. “at that they were strangers. just that they were 'people.“ 8. the subject lives in a lively world of fully human objects. There is a sense of rapport with people and a well-developed understanding of their thoughts. feelings and cggfljcts. lhere is a well-articulated internal model of people being involved with each other. inciud|09 an Understanding of why they form relationships, why they 981 03!.0‘ them. and what interferes with them. Others' behavior and personal characteristics are conSidered in perspective, regain open to reinterpretation. etc. lnere is a good deal of self and interpersOnal aware- ness. indeed. a psychological-nindedness. Relationships are. for the most part. not neurotically conceived. or Ex le: f ‘am: reflective consent is made in the dream report about what the dream is saying.or the self in the dream re:-ects during the dream on some aspect of his feelings toward another dreao cha ter. or on some aspect of that character's conflicts. Hora than reporting t 2 it n95 0? HIShes of a dream character. the subject spells out 1“ awareness 0' so. conflict(s) in the character. himself or their relationshiP- “‘5“ ‘ ‘a$--. - . fwdifl‘?'1’:"EJ131333“”5‘33“...13-343:m~~. :sfififlntfi'fiscem’um TI" new. «vo‘r’fu .‘P‘tfn in item H‘uraw-W' ”I'm”. . ~»..----~-—--— - ~ 1fl19 ,._.. M--- . ‘ . A L‘ . I Ian ‘7‘“- “w.-.-v~~. ‘ Page tight Us 2. There is a creative use of humor in the dPCBM. aroung the hunan characters in the drlamdi t i . here is something particular I 5 inc‘ V0 and subtle character noted by the subject. _ _ 'bOUt . dream 4. lhe dream characters and thalr interactions seen only minimally ciracted by neurotic concerns. lhere is a maturity about the dream characters. ‘ Particularly centerin9 Sample Dream: ‘ - I dreamt l was with L.. my teacher in the math section l'n.in. in the dream he had this rather absurd hat on. it reminded me in the dream of scathing I'd seen in the circus. He came up to me and someone else who was with me. but who isn't clear. and suddenly started to deliver this serious speech about the terrible state of the econch, and then about the importance of ecology and threw in. for 50:53 reason, something about the price 0‘ clover in Britain. 1 asked him why he was telling me all this and he asked me if i wasn't interested and l said no, l'n not. and that he seemed not to be either, that he seexad to be trying to be a clcz-m and serious at the sane the. like he wants to clam around but is aluays pulling himself back to being so serious. as if he thinks it isn't right to clean around (this is something l‘ve t-egun to think about this guy in real life). He seened to hear me in the dream. but changed the subject. as ii he didn't like what l'd said to him. APPENDIX 4 Examples of Changed Human Percepts Intake This looks like maybe birds with hearts and blood flying around. Heart's on table and there's bone in their hearts. (Card 3, Subject #11, Group Psychotherapy plus Medication [PM] ) Two men and butterflies and blood spots. Very abstract . . . (Card 3, #7, Group Psychotherapy Only [PO]) Something dancing--1ike a cartoon in a way, of a person, like people dancing, floating through the sky . . . a person bent. (Card 3, Subject #25, Group PM) Two people, two Martians or two girls--red and black, tits, head, body, dong . . . ding dong. (Card 3, Subject #23, Group PM) 120 20 Months This looks like two men standing there and objects flying around. Two guys pulling on something. Two natives dancing over a pot. Two women pulling on something, bust, back-end and high-heel shoes, head. FOOTNOTES FOOTNOTES lHans Strupp (1978) has characterized the challenge facing psychotherapy researchers as one of demonstrating that psychotherapy is not a ”unitary process." Patient variables, he states, must be selected according to their fit with other process and outcome variables. Moreover, he states that researchers frequently fail to take seriously the role of values in outcome research. Clearly, the present study has focused on a system of values that favors the patient's experience as a most important outcome measure. But even within the context of "measuring" the patient's experience, the present study imposes the values for "maturity" as more or less dictated by object relations theory. 2Interestingly, the study in which Schwager and Spear (1981) failed to find changes in traditional Rorschach indices of ego functioning was conducted at the Menninger Institute. As in the present study, the investigators were examining very disturbed patients who were being treated with exploratory psychotherapy. 121 BIBLIOGRAPHY BIBLIOGRAPHY APA Commission on Psychotherapy. (1982). Psychotherapy research: methodological and efficacy issues. American Psychiatric Association. Appelbaum, S. (1977). The anatomy of change. New York: Plenum Press. Bell, S. M. (1970). The development to the concept of object as related to infant-mother attachment. Child Development, 2;, 292-311. Beres, D., & Joseph, E. (1965). The concept of mental representation in psychoanalysis. International Journal of Psycho-Analysis, 2g, 168-182. Blatt, s. J., Wild, c. M., & Ritzler, B.A. (1975). Disturbances in object representation in schizophrenia. Psychoanalysis and Contemporary Science, 235-288. Blatt, S. J., Brenneis, C. B., Schimek, J. G., & Glick, M. (1976a). Normal development and psychOpathological impairment of the concept of the object on the Rorschach. Journal of Abnormal Psychology, §§, 364-373. Blatt, S. J., & Wild, C. M. (1976). Schizophrenia: A developmental analysis. New York: Academic Press. 122 123 Blatt, S. J. Schimek, J., & Brenneis, C. B. (1980). The nature of the psychotic experience and its implications for the therapeutic process. In J. Strauss et al. (eds.), Theypsychotherapy of schizophrenia. New York: Academic Press. Blatt, S. J., & Lerner, H. (1983). The psychological assessment of object representation. Journal of Personality Assessment, 1983, 31, 7-28. Breggin, D. (1983). Psychiatric drugs: Hazards to the brain. New York: Springer. Campbell, D. T., & Fisk, D. W. (1959). Convergent and discriminant validation by the multitract—multimethod matrix. Psychological Bulletin, 3g, 81-105. Exner, J. E. (1974). The Rorschach: A comprehensive system. New York: John Wiley & Sons. Fairbairn, W. R. D. (1952). Psychoanalytic studies of the personalipy. Boston: Routledge & Kegan Paul. Frieswyk, S., & Colson, D. (1980). Prognostic considerations in the hospital treatment of borderline state: The perspective of object relations theory and the Rorschach. In: Borderlineyphenomena and the Rorschach test. Kwawer, J., Lerner, H., Lerner, P., & Sugarman, A. (Eds.) New York: International Universities Press, 229-257. Glatt, C. T. (1971). Some Rorschach correlates of change in clinical status: an investigation of ego regression in schizophrenia. Unpublished doctoral dissertation, Michigan State University. 124 Grey, A., & Davies, M. (1981). Mental health as level of interpersonal maturity. Journal of the American Academy of Psychoanalysis Vol. 2, (4) 601-604. Gunderson, J. G., & Comes-Schwartz, B. (1980). The quality of outcome from psychotherapy of schizophrenia. In J. Strauss et al. (Eds.) The psychotherapy of schizophrenia. New York: Plenum, 1980. Guntrip, H. (1969). Schizoid phenomena, object relations, and the self. New York: International Universities Press. Hatcher, R. L., & Krohn, A. (1980). Level of object representation and capacity for intensive psychotherapy in neurotics and borderlines. In: Borderline phenomena and the Rorschach test, J. Kwawer, H. Lerner, P. Lerner, & A. Sugarman. (Eds.) New York: International Universities Press, 299-321. Hartmann, H. (1939). Ego psychologyand the problem of adaptation. New York: International Universities Press. Hoedemaker, E. D. (1967). The psychotic identifications in schizophrenia: the technical problem. In: Boyer B. L. Bryce, Giovacchine, P. L. (Eds.) Psychoanalytic treatment of schizophrenic and characterological disorders. New York: Science House. Hollon, S., & Beck, A. T. (1978). Psychotherapy and drug therapy: Comparisons and combinations. In: S. L. 125 Garfield and A. E. Bergin (Eds.), Handbook of psychotherapy and behavioral change: an empirical analysis, 2nd edition. New York: John Wiley & Sons, Inc., 437-490. Howell, D. C. (1982). Statistical methods for psychology. Boston: Duxbury Press. Jacobson, E. (1964). The self and the object world. New York: International Universities Press. Karon, B. P., & VandenBos, G. R. (1970). Experience, medication, and the effectiveness of psychotherapy with schizophrenics: a note on Drs. May and Tuma's conclusions. British Journal of Psychiatry, 116, 427-428. Karon, B. P., & VandenBos, G. R. (1972). The consequences of psychotherapy for schizophrenic patients. Psychotherapy: Theory, Research, and Practice, 39, 111-120. Karon, B. P., 8 VandenBos, G. R. (1981). Psychotherapy of schizophrenia: The treatment of choice. New York: Jason Aronson. Kernberg, O. (1975). Borderline conditions and pathological narcissism. New York: Jason Aronson. Kernberg, O. (1976). Objects relations theory and clinical psychoanalysis. New York: Jason Aronson. Kiesler, D. J. (1966). Some myths of psychotherapy research and the search for a paradigm. Psychological Bulletin, 154 110-136. 126 Klein, K., Gittelman, R., Quitkin, P., & Rifkin, A. (1980). Diagnosis and Drug Treatment of Psychiatric Disorders: Adults and Children (2nd ed.). Baltimore: Williams & Wilkins. Klein, M. (1946). Notes on some schizoid mechanisms. In: Envy and gratitude and other works. New York: International Universities Press. Klein, M. (1952). The origins of transference. International Journal of Psycho-Analysis, 33, 422-438. Klein, M. (1948). Contributions to psychoanalysis. London: Hogarth. Object representations Krohn A., & Mayman, M. (1974). in dreams and projective tests. Bulletin of the Menninger Clinic, 38, 445-466. Kwawer, J. S., Lerner, H. D., Lerner, P. M., & Sugarman, A. (1980). Borderline phenomena and the Rorschach test. New York: International Universities Press. Lerner, H. (1983). An object representation approach to psychostructural change: a clinical illustration. Journal of PersonalityAssessment,‘41, 314-323. & Cornelison, A. (1965). Lidz, T., Fleck, S., New York: International Schizophrenia and the family. Universities Press. Lidz, T. (1973). The origin and treatment of schizophrenic disorders. New York: Basic books. 127 Lidz, T. (1980). The developing guidelines to the psychotherapy of schizophrenia. In: J. Strauss et al. (Eds.), The psychotherapy of schizophrenia. New York: Plenum. Loewald, H. (1962). Internalization, separation, mourning, and the superego. Psychoanalytic Quarterly, 31' 483-504. Mahler, M. S. (1965). Mother-child interaction during separation-individuation. In: The selected papers of MaEgret S. Mahler (Vol. 2). New York: Jason Aronson 1979, 35-48. Masterson, J. F. (1976). Psychotherapy of the borderline adult: A developmental approach. New York: Brunner/ Mazel. May, P. R. A. (1975). Schizophrenia: Overview of treatment methods. In: A. M. Friedman, H. I. Kaplan, & B. J. Sadock (Eds.) Comprehensive Textbook of Psychiatsy (Vol. 2) Baltimore: Williams and Wilkins, 1192-1216. Mayman, M. (1967). Object representations and object relations in Rorschach responses. Journal of Projective Techniques and Personality Assessment, _§l, 17-24. Meissner, W., W. (1978). The paranoid process. New York: Jason Aronson. Miessner, W. W. (1982). Internalization in psychoanalysis. New York: International Universities Press. film”! 128 Milner, M. (1969). The hands of the living god. New York: International Universities Press. Novey, S. (1968). The meaning of the concept of mental representation of objects. Psychoanalytic Quarterly, 21, 57-79. Orlinsky, D. E., & Howard K. (1978). The relation of process to outcome in psychotherapy. In: S. L. Garfield & A. E. Bergin, Handbook of psychotherapy and behavior change: An empirical analysis. New York: John Wiley and Sons, Inc., 283-331. Provence, & Lipton, R. C. (1962). Infants in institutions. New York: International Universities Press, 1962. Rapaport, D., Gill, M. M., & Schafer R. (1945-1946). Dlagnostic psychological testing, 2 vols. Chicago: Year Book Publishers. Ritzler, B., Wyatt D., Harder, D., & Kaskey, M. (1980). Psychotic patterns of the concept of the object on the Rorschach. Journal of Abnormal Psychology, 89, 46-55. Rosenfeld, H. (1965). Psychotic states. New York: International Universities Press. Ryan, E.R., & Bell, M. D. (1984). Changes in object relations from psychosis to recovery. Journal of Abnormal Psychology, 23, 209-15. Sandler, J., & Rosenblatt, B. (1962). The concept of the representational world. The Psychoanalytic Study of the Child, ll, 128-45. 129 Schafer, R. (1955). Psychological test evaluation of personality change during intensive psychotherapy. In: Projective testing and Psychoanalysis. New York: Universities Press, 1967. Schafer, R. (1968). Aspects of internalization. New York: International Universities Press. Schafer, R. (1976). A new language for psyphoanalysis. New Haven: Yale University Press. Schwager, E., & Spear, W. E. (1981). New perspectives on psychological tests as measures of change. Bulletin of the Menninger Clinic, 45(6), 527-541. Searles, H. (1959). Integration and differentiation in schizophrenia. In: Collected papers on schizophrenia and related subjects. New York: International Universities Press, 1965. Searles, H. (1962). The source of anxiety in paranoid schizophrenia. In: Collected papers on schizophrenia and related subjects. New York: International Universities Press, 1965. Searles, H. (1979). Countertransference. New York: International Universities Press. Shapiro, D. (1965). Neurotic styles. New York: Basic Books. Smith, M. L., Glass, G. U., and Miller, T. J. (1980). The Benefits of Psychotherapy. Baltimore: The John Hopkins University Press, 1980. 130 Spear, W. E. (1978). The relationship of clinical diagnosis to structural and thematic aspects of manifest object representations in hospitalized borderline and schizophrenic patients. Unpublished doctoral dissertation. Columbia University. Spear, W., & Lapidus, L. B. (1981). Qualitative differences in manifest object representations: implications for a multidimensional model of psychological functioning. Journal of Abnormal Psychology, 90, No. 2, 157-167. Spear, W., & Hymowitz, P. (1982). An examination of SchiZOphrenic patients' object relations using a theory of ego states. Bulletin of the Menninger Clinic, fig, No. 6, 52-59. Srole, L. T. S., Langner, S. T., Michael, M. K., Opler, T., & Rennie, A. F. (1962). Mental health in the metropolis: the midtown Manhattan study. Volume 1, New York: McGraw-Hill. Strupp, H. H. (1978). Psychotherapy research & practice: an overview. In: S. L. Garfield and A. E. Bergin (Eds.), Handbook of psychotherapy and behavior chaags: An empirical analysis, 2nd edition. New York: John Wiley & Sons, Inc., 3-22. Sullivan, H. S. (1953). The interpersonal theory of psychiatry. New York: Norton. Sullivan, H. S. (1962). Schizophrenia as a humanlprocess. New York: Norton. 131 Teuber, S. B. (1983). Children's Rorschach scores as predictors of later adjustment. Journal of Consulting and Clinical Psychology, El, 379-385. Urist, J. (1973). The Rorschach test as a multidimensional measure of object relations. Unpublished doctoral dissertation, University of Michigan. Urist, J. (1977). The Rorschach test and the assessment of object relations. Journal of Personality Assessment, _]_..' 3-90 Urist, J., & Shill, M. (1982). Validity of the Rorschach mutuality and autonomy scale: A replication using excerpted responses. Journal of Personality Assessment, 46, 450-454. Wachtel, P. L. (1977). Psychoanalysis and behavior therapy. New York: Basic Books. Wallerstein, R. S. (1968). The psychotherapy research project of the Menninger Foundation: a semi-final view. In: J. M. Shlien (ed.), Research in Psychotherapy (Volume 3). Washington, D.C.: American Psychological Association, 584-605. Werner, H. (1948). Comparativepsychology of mental development. New York: International Universities Press. Winnicott, D. W. (1958). Throughspaediatrics to psycho-analysis. New York: Basic Books. MITIWICIWINIWSIWIWIUIWIIVIWWIIWWillss 3 1293 0306] 7837