Clinical Implications of Dispositional Self-Consciousness

Loyola University Chicago Loyola eCommons

Dissertations Theses and Dissertations

1981

John Smith Crandell Loyola University Chicago

Follow this and additional works at: https://ecommons.luc.edu/luc_diss

Part of the Psychology Commons

Recommended Citation Crandell, John Smith, "Clinical Implications of Dispositional Self-Consciousness" (1981). Dissertations. 2002. https://ecommons.luc.edu/luc_diss/2002

This Dissertation is brought to you for free and open access by the Theses and Dissertations at Loyola eCommons. It has been accepted for inclusion in Dissertations by an authorized administrator of Loyola eCommons. For more information, please contact [email protected].

DISPOSITIONAL SELF-CONSCIOUSNESS

John S. Crandell

A Dissertation Submitted to the Faculty of the Graduate School

of Loyola University of Chicago in Partial Fulfillment

of the Requirements for the Degree of

Doctor of Philosophy

January

1981 ACKOWLEDGMENTS

I would like to express my particular appreciation to Dr Nic olay, who has long supported my interest in this area of research.

He has truly served as a Director, not only shaping the research process but guiding me safely through the many obstacles to com pleting my graduate work. To Drs. DeWolfe and Shack go much of the credit for the fact that this project was able to bridge the apparent gaps between theory, research, and clinical application.

Field research requires unusual generosity and support from already busy clinicians. I am particularly indebted to Gerald Moz dzierz and Rulon Gibson for their advice and encouragement. The staff of the Mental Hygiene Clinic and the four units of Inpatient

Psychiatry at Hines Veterans Administration Medical Center were un failingly cooperative. They all have my deep gratitude.

iii VITA

The author, John Smith Crandell, was born January 15, 1951 in

Ann Arbor, Michigan.

His elementary education was obtained at City and Country

School in Bloomfield Hills, Michigan. He completed his public education

at Seaholm High School in Birmingham, Michigan. Undergraduate work in

psychology and philosophy was done at Kalamazoo College in Kalamazoo,

Michigan and Fourah Bay College in Freetown, Sierra Leone, West Africa.

He was elected to Phi Beta Kappa and graduated Cum Laude from Kalamazoo

in June, 1972.

He began his graduate work at Eastern Michigan University and

obtained the M.S. degree in experimental psychology in August, 1976.

While preparing his thesis on "the Impact of Differential Empathy on

Discrepant Trait Attribution to Self and Others", Mr. Crandell worked

as Director of Concord Counseling Service in Westerville, Ohio. He re

turned to graduate school at Loyola University in September, 1976, working in clinical psychology and completing a clerkship and intern

ship at Hines Veterans Administration Medical Center. Since June, 1979 he has worked at Northwestern Community Mental Health Center in Chicago,

Illinois.

iv Mr. Crandell has presented professional papers and workshops at the 1976 Ohio Drug Studies Institute and the 1980 convention of the

American Psychological Association. His publications include: "Review of Psychotherapy: Theory, Research and Practice", 1980; "Designing a

Computer-Assisted Medical Record to Meet Quality Assurance Standards

in a Mental Health Setting", 1980; and "Clinical Applications of a

Computer-Assisted Medical Record", 1981.

v TABLE OF CONTENTS

Page

ACKNOWLEDGMENTS iii

VITA •. iv

TABLE OF CONTENTS vi

LIST OF TABLES • ix

LIST OF FIGURES xii

CONTENTS FOR APPENDICES x-iii

INTRODUCTION • • • • . • • • 1

REVIEW OF RELATED LITERATURE 7

Conceptual and Theoretical Overview 7 Self-Awareness Theory • . • • . • . 17 Demonstrations of Objective Self-Awareness •• 21 Importance of Standards During Objective Self-Awareness . • • . • • . • • . • . • • 31 Clinical Implications of Research on Self-Awareness. 32 Avoidance of Objective Self-Awareness 33 Individual Differences in Self-Consciousness 39 Clinical Implications of Self-Consciousness 56 Ego Strength • • . • • • • • • • • • • • • • • 58 Ego Strength Scale and Prediction of Improvement in Psychotherapy • . • • . • . • • • • • • • • • • . 58 Ego Strength Scale Versus Ego Strength Concept • 66 Ego Strength Scale and Objective Self-Awareness 77 Premature Termination from Psychotherapy 78 Methodological Issues • • • • 78 Psychological Tests • • • • • • • • • • 88 Social Class • • • . • . • • . • 92 Other Patient Characteristics 95 Therapist Characteristics 104 Expectancy Match . 110 Other Factors 116 Self-Consciousness and the Treatment Dropout Phenomenon • • . • 119

vi Personality Typology . . • . • • • . • • • . • . • 123 A Psychodynamic Personality Typology • • . • 123 Personality Type and Level of Self-Awareness 133 Summary of Hypotheses 138

METHOD 142

Subjects • 142 Materials 150 Reliability of Personality Type . 151 Procedures • • • • 155 Experiment 1 155 Experiment 2 159 Experiment 3 160

RESULTS AND DISCUSSION 161

Self-Consciousness Levels of Psychiatric Patients 161 Comparison of Patient Groups . . . . . 161 Self-Consciousness Scores of Patients Compared with Normal Samples • • • • • • . • ...... 167 Factor Structure of the Self-Consciousness Scale 171 Self-Consciousness and Ego Strength • . 175 Self-Consciousness and Premature Termination from Psychotherapy • • • • • • • . . • • • • . • . . . 180 Self-Consciousness and Continuation in Treatment 181 Indirect Effects of Self-Consciousness Scores on Premature Termination • • • • . • ...... 187 Discriminant Analyses . • . • • • • . . . . . 189 Other Hypotheses about Premature Termination 196 Personality Type and Self-Consciousness Levels 199 Methodological Issues • • • • . . . . 199 Self-Consciousness Levels for Five Global Personality Types . • • . . . • • • • . . 200 Comparisons Between Specific Personality Types 209 Subscale Item Averages and Personality Type 213 Discussion . . • . • . • . • • . . • • . • • . . . . 218 Implications for Self-Awareness Theory . . . . 218 Implications for Further Research and Clinical Applications 222

SUMMARY 225

REFERENCE NOTES 228

vii REFERENCES 229

APPENDIX A . . . . 249

APPENDIX B 260

APPENDIX C . . . . . 263

APPENDIX D 265

APPENDIX E . . . . . 267

viii LIST OF TABLES

Table Page

1. Response Latency Under High and Low Self-Awareness Conditions • • • • . . • • • • . . . . • . . . 23

2. Effect of Self-Awareness and Discrepancy from Stan- dards on Minutes of Participation . . . . . 35

3. Means and Intercorrelations of the Private (PR-SC), Public (PU-SC), and Social Anxiety (SA-SC) Subscales 42

4. Correlations Between Social Desirability and Subscales of the Self-Consciousness Scale • • . • 46

5. Correlations Between Test Anxiety and the Self- Consciousness Scale 47

6. Percent of Self-Attributed Responsibility with High and Low Self-Awareness and Private Self-Consciousness 51

7. Percent of Self-Reference with High and Low Self Awareness and Private Self-Consciousness • . • 52

8. Percent of Patients Terminating Prematurely in Studies Reviewed . • . • . . • • . 79

9. Expected Rank Order of Personality Types on the Self-Consciousness Scale • • • • . • • • • • . 135

10. Demographic and Clinical Management Characteristics of Research Participants • . . • . • • • • • • • • • 148

11. Agreement Between Clinicians in Classifying Personality Types • • • . • • • • • • • 153

12. Self-Consciousness Scale Scores when Administered Before or After the Ego Strength Scale • • • • • • • • • 158

13. Mean Self-Consciousness Scale Scores for Patient Groups • • • • • • • . • . . • • . . • • • • • • • • 162

14. Standard Deviations of Self-Consciousness Scale Scores for Patient Groups . . • • . • . • • • • • • • • 163

ix 15. Intercorrelations of Self-Consciousness Subscale Scores for Patient Groups • . . • . • • • • • •• 164

16. Self-Consciousness Scale Means, Standard Deviations, and Intercorrealtions for Patients and Normals 168

17. Communalities and Factor Matrix for the Self- Consciousness Scale • . • . 174

18. Correlations of the Ego Strength Scale with the Self-Consciousness Scale • • . • • • • • • 176

19. Intercorrelations of the Self-Consciousness Subscales for Patients at Different Levels on the Ego Strength Scale • • • . • • • . • • 178

20. Self-Consciousness Levels of Decliners, Terminators, and Continuers 182

21. ANOVAs for Self-Consciousness Levels of Decliners, Terminators, and Continuers . 183

22. Total Self-Consciousness Score Intervals by Length of Stay . • . • • • . • • • • . . . • . • . . 185

23. Association of Self-Consciousness Scores with Covariates of Premature Termination . • . . . 190

24. Standardized Discriminant Function Coefficients for Continuers, Terminators, and Decliners 192

25. Standardized Discriminant Function Coefficients for Continuers and Terminators 193

26. Global Diagnosis and Personality Type • 201

27. Self-Consciousness Scale Means (with ANOVAs) as a Function of Diagnosis and Personality Type 202

28. Self-Consciousness Scale Means (with ANOVAs) as a Function of Unmixed Personality Type . . . . . 205

29. Distribution of Social Anxiety Scores by Personality Type • . . • • • . . • • • . • . • • ...... 206

X 30. Distribution of Total Self-Consciousness Scores by Personality Type • . • • . • . . • . • . • . • • . 207

31. Expected and Observed Rankings of Personality Types on the Self-Consciousness Scale . • • • . • • 211

32. Expected and Observed Subscale Ranks for Personality Types • • • . • . . . • . 214

33. Self-Consciousness Subscale Item Means for Person ality Types • . • . • . • • 215

34. Self-Consciousness Scale Factor Score Means for Personality Types . . • • • • • • • • . • 217

xi LIST OF FIGURES

Figure Page

1. Reviewers' Responses to Variables Associated with Premature Termination from Psychotherapy • . 96

2. Summary of Variables Associated with Premature Termination from Psychotherapy . • • 120

3. Characteristics of Different Personality Styles 127

4. Personality Types and Descriptions • • . 129

5. Number of Subjects Participating in Each of the Component Studies • . . • • ...... • 14 7

6. Percent Continuing among Patients in Successive Intervals of Total Self-Consciousness ...... • 186

xii CONTENTS FOR APPENDICES

Page

APPENDIX A Copies of Questionnaires and Forms Used 249

Self-Consciousness Scale 250

Ego Strength Scale 252

Intake Rating Scale • 256

Instructions for Completing the Intake Rating Scale 257

Consent Form 258

APPENDIX B Reliability Study Results • 260

Percent Agreement on Primary Type • 261

Percent Agreement on the Applicability of Each Given Type . • • • • . • 262

APPENDIX C Factor Structure of the Self-Consciousness Scale 263

Varimax Rotated Factor Matrix for All Factors with an Eigenvalue Greater than One 264

APPENDIX D Diagnoses of Research Participants 265

DSM II Diagnoses of Research Subjects •• 266

APPENDIX E Raw Data 267

Format of Raw Data 268

Raw Data 270

xiii INTRODUCTION

The ability to reflect on oneself is an important human

capacity that has received considerable attention from philosophical

and psychological theorists (Bugenthal, 1965; Mead, 1934; Sartre,

1956). In the last decade, self-awareness theory (Buss, 1980; Duval

& Wicklund, 1972; Fenigstein, Scheier, & Buss, 1975; Wicklund, 1975b)

has provided a basis for empirical research on self-consciousness.

Self-focused attention fosters a comparison between performance and

salient personal standards and feelings. The usual result of such a

comparison - the experience of dissonance - leads to one of two

resolutions to this internal conflict: either behavior or standards are altered to minimize the discrepancy or attempts are made to avoid focal self-awareness. Performance in simple tasks is improved in

the presence of mirrors, audiences, and video recordings of the

subject (Wicklund & Duval, 1971/1972). But there are negative sequellae as well, since self-consciousness arouses a state similar

to evaluative anxiety. If the personal importance of the task passes an optimal point, self-focused attention is associated with decrements in performance (Liebling & Shaver, 1973) and in self esteem (Ickes, Wicklund & Ferris, 1973). Especially if the disso nance cannot be eliminated quickly, stimuli to self-focused attention may cause escape behavior (Cummings, 1976; Duval, Wicklund & Fine,

1 2

1971/1972; Wicklund, 1975b).

To date, the primary application of self-awareness theory has been in social psychology. Duval & Wicklund (1972) have used it in discussions of such traditional topics as attribution, conformity, social facilitation, and self-esteem. There has been no systematic attempt to generalize the theory to phenomena of interest in clinical psychology, although Duval and Wicklund and subsequent researchers

(Fenigstein et al., 1975) have suggested that such an application would be fruitful. This dissertation will attempt to extend self awareness theory to three issues of practical and theoretical impor tance in clinical psychology.

Of primary concern is the level and structure of self consciousness in a psychopathological sample. Fenigstein et al.

(1975) have administered a Self-Consciousness Scale to samples of normal college students, validating it as a measure of the disposi tional tendency to focus attention on oneself. They report three components: attention to private aspects of the self, attention to the self as social object, and social anxiety in anticipation of negative evaluation. It may be that a clinical sample deviates from the norms in overall self-consciousness or in subscale scores.

Certainly, the generalized anxiety that typifies many diagnostic groups might well be expressed in the Social Anxiety subscale.

Similarly, the disruptions in intra- and interpersonal functioning found among those seeking psychotherapy may represent disturbances in the ability to reflect on their internal feelings and motivation, 3

or on their impact on others. Recent theorizing about self

consciousness has stressed the importance of the distinction between

awareness of public and private aspects of the self (Buss, 1980).

A more relevant distinction for a psychopathological sample may well

be between the evaluative and nonevaluative aspects of self-awareness.

For this reason the factor structure of the scale will be reexamined.

Lastly, since the clinical population is certainly not homogeneous,

self-consciousness will be examined as a function of ego strength

or degree of pathology. The first goal of this research will be to

examine the responses of a psychopathological sample to evaluate the

degree to which self-awareness theory, as developed for a normal

population, can be generalized to clinical situations.

The second goal of this dissertation will be to examine the

hypothesis that deviant levels of self-consciousness correspond with

a tendency to prematurely terminate psychotherapy. Duval & Wicklund

(1972) indicate that escape behavior is especially likely when there

is a large discrepancy between behavior or attributes and relevant

personal standards, when discrepancy reducing behavior is difficult, and stimuli to self-awareness are salient. The early stages of psychotherapy closely approximate this situation. The presence of

an expert observer and the all but exclusive focus on the patient's life guarantee sustained self-consciousness. The topic of discussion

is generally on problems of such severity, duration, and intracta bility that the patient is in considerable distress. The required

changes are usually characterological or otherwise involve a major 4

shift in lifestyle, changes that can take place only after extended

time or effort. This situation is closely analogical to the one

Duval and Wicklund suggest will motivate avoidance behavior, in

this case, failing to continue in treatment.

Surveys show that up to 65% of those beginning psychotherapy

terminate unilaterally (Malzer, 1980). A great deal of research has investigated the reasons for this high rate of dropping out.

Relatively few patient characteristics have been identified as related to premature termination (Garfield, Affleck, & Muffly,

1963). Instead, research has increasingly shifted to situational factors, and specifically to the interaction of patient and therapist

(Frank, Gliedman, Imber, Nash, & Stone, 1957). Surveys of terminators repeatedly find complaints about the quality of the relationship

(Hiler, 1958; Kamin & Coughlin, 1963; Saltzman, Luetgert, Roth,

Creaser, & Howard, 1976). Therapists were accused of being cold and aloof, lacking understanding or interest, and showing too little respect. These complaints less likely point to real shortcomings in the therapist than to an interaction pattern that fails to assuage the patient's initial doubts, anxieties, and self-criticism. What are perceived as therapist failures may instead be signs of excessive self-consciousness. Premature termination is, then, a major problem for which the existing explanations are consistent with an account based on the avoidance of excessive self-focused attention. The second study reported will evaluate the applicability of a specific hypothesis from self-awareness theory to the clinically important 5 issue of therapy drop outs.

The third goal of the research will be to examine differences in self-consciousness levels among distinct personality types. While the first two studies consider patients as a group, this experiment pursues the difference among subgroups of the clinic sample. Self consciousness may constitute an important treatment issue for partic ular types. The self-preoccupation of obsessive or passive patients may make them particularly vulnerable to paralyzing self-criticism if the therapist fails to carefully limit the amount of self-focused attention. Conversely, the lack of insight displayed by sociopaths or paranoid patients is likely attributable to a lack of self monitoring. This research expands on the second goal. Some types may drop out of treatment because therapy exaccerbates a preexisting excessive self-consciousness (e.g., passive types) while others may lack the requisite ability to reflect on their internal motivation that is necessary for insight oriented therapy (e.g., paranoid patients). If either of the first two experiments find self consciousness to have important clinical implications, this more microscopic examination of differences between psychopathological subgroups should help to clarify the results.

This dissertation is an initial attempt to extend self awareness theory to a new area: clinical psychology. From this point of view, it is an attempt to generalize a well established body of findings to a new population. Clinicians have, of course, previously considered the importance of self-focused attention. 6

Verbal therapies have historically considered insight to be a critical element in treatment (Greenson, 1967; Yalom, 1975). But much of this data has been presented in terms of anecdotal case studies, meta-psychological theorizing, or technical directives.

From the point of view of clinical psychology, this research attempts to add an empirical perspective on the importance of self-conscious ness to the understanding and treatment of psychopathology. REVIEW OF RELATED LITERATURE

This research applies self-awareness theory to several areas of clinical psychology. The literature review will be extensive.

It will begin with a conceptual overview which samples the theoretical perspectives on self-consciousness influencing current approaches to human relations and psychotherapy. This section provides the con text for the following one, which reviews self-awareness theory in some detail. Successive sections will cover the literature related to each of the clinical issues hypothesized to be related to self consciousness levels. In order, these include discussions of the

Ego Strength Scale (Barron, 1953) as an index of degree of pathology, variables related to premature termination from psychotherapy, and personality types in a psychopathological population. The chapter will conclude with a restatement of the central hypotheses tested in this dissertation.

Conceptual and Theoretical Overview

Socrates dictum, "Know Thyself", is oft-cited and time honored.

It serves not only as an epistemological tool and ethical directive; it also takes on the force of an ontological description of the human condition. The capacity to reflect on ones acts, attributes, and thought processes is uniquely connected to what we think of as human.

7 8

It is the defining feature conventionally used to distinguish man

from animal. It is intimately connected to the capacity for abstract

thought and evolution through culture.

With this general cultural attitude as background, it is not

surprising that the social sciences have treated self-awareness

reverently. Mead (1934) and Piaget (1954) make the ability to

"decenter" and examine oneself as an object a developmental hallmark.

For both, self-objectification is a necessary preliminary to effective

social relations. Some self-understanding is requisite for the funda

mental interpersonal skill of empathy, which in turn has been defined

as a basis of therapeutic interaction (Kohut, 1971, 1977; Rogers,

1961). And self-awareness is just as important for the client as

for the helper in the therapeutic dyad. Since Freud (Greenson, 1967)

a dominant thrust of dynamic theorizing has been on the importance

of insight to healing.

But there is a problematic side to self-awareness which is best captured in the word "self-consciousness". If the hyphen is

removed or exaggerated, the word shatters into two components, "self" and "consciousness", which jointly represent the crucial capacity for

self-focused attention. The connotation is positive. But as a single unit, "self-consciousness", the implication is very different. The associations are with anxiety, paralysis of action, exposure to harsh scrutiny, and shame. It is more than a semantic accident that the word can be defined as either "the state of being intensely aware of oneself" or "ill at ease" (Webster, 1977, p. 1048). There is a 9 fundamental ambivalence with which we approach the concept, belying the esteem held by the previously mentioned theorists.

Various writers have described the potential problems associated with self-focused attention. One perspective examines the advantages of unselfconscious action (Csikszentmihalyi, 1979;

Ornstein, 1972). The resulting "flow experience" (Furlong, 1976) is described glowingly as characterized by heightened concentration, merging of awareness with action, enhanced mastery, loss of doubts, and an ecstatic joy in action for its own sake. This description provides a counterpoint to the overcontrolled and unnatural busyness attributed to reflective states of consciousness (Bassos, 1976).

But the main critique of self-awareness is frontal and best epitomized in the writings of Sartre (1956). For him, the basic ontological structure of human beings is tripartite. First, as being-for-itself, we are the freedom and limitless possibilities of our consciousness. Second, as being-in-itself, we are paradoxically rooted in facticity by our bodies. Lastly, as being-for-others, we are objects and tools in other people's projects. As such, being- for-others represents the intolerable crystalization of the first two states which gives rise to the famous phrase "Hell is other people" (Sartre, 1946). The look of "the Other" reminds us that our freedom is limited:

I grasp the Other's look at the very center of my act as the solidification and alienation of my own possibilities. In fear or in anxious or prudent anticipation, I perceive that these possibilities which I am and which are the condition of my transcendence are given also to another, given as about to be transcended in turn by his own possibilities. The Other as a 10

look is only that -my transcendence transcended. (Cumming, 1965, p. 201)

The presence of the Other is not only the occasion for fear, but

for shame:

Shame .•• is shame of self; it is the recognition of the fact that I am indeed that object which the Other is looking at and judging. I can be ashamed only as my freedom escapes me in order to become a given object • • • The world flows out of the world and I flow outside myself. The Other's look makes me be beyond my being in this world and puts me in the midst of the world which is at once this world and beyond this world. (p. 199)

The Other's ability to reconstitute the world from an alien perspec-

tive changes the relation of self to itself:

Let us imagine that moved by jealousy, curiosity, or vice I have just glued my ear to the door and am looking through a keyhole. I am alone and on the level of non-thetic self consciousness. This means first of all that there is no self to inhabit my consciousness, nothing therefore to which I can refer my acts in order to qualify them. They are in no way known; I am my acts and hence they carry in themselves their whole justification ••• My consciousness clings to my acts, it is my acts; and my acts are commanded only by the ends to be attained and by the instruments to be employed . • . • Jealousy, as the possibility which I am, organizes this instru mental complex by transcending the complex towards itself. But I am this jealousy; I do not know it ••. Moreover, I cannot truly define myself as being in a situation; first because I am not a positional consciousness of myself; second because I am my own nothingness • • • •

But suddenly I hear footsteps in the hall. Someone is looking at me. What does this mean? It means that I am suddenly affected in my being and that essential modifications appear in my structure - modifications which I can apprehand and fix conceptually by means of the reflective cogito.

First of all, I now exist as myself for my unreflected con sciousness • • • • I see myself because someone sees me • • The unreflected consciousness does not apprehend the person directly as an object; the person is presented to consciousness insofar as the person is an object for the Other. (pp. 196-198) 11

In apprehending oneself as an object, the possibility of self-evaluation appears. So, while the Other's presence may be the occasion, the experience is internal and basic to a person who is at once a being and a consciousness. This embarassment is univer sal. At a party, people are dancing. Low lights buffer against self-consciousness. The music and the rhythm of the movement totally occupy each dancer's awareness. There is no self and so no self-judgment, only the kinesthetic awareness of flexing muscles and the sound of the music. A dancer may even close his eyes or defocus his gaze to enhance the awareness of these stimuli. Yet, suddenly, the gaze takes in an attentive bystander. Eyes focus.

Just as instantly, the dancer transposes himself into the perspec tive of the observer. Awareness of music and movement recede. And the dancer sees himself: the silly step, the harsh asymmetry of the movements, the awkwardness, the vacuous look. With this outside perspective, this "objective" and objectifying distance, opportunity for self-criticism abounds. The stimulus need not even be Sartre's

"Other". The reflection from a strategically placed mirror or the memory of other dances or occasions of criticism may be sufficient to begin the avalanche of self-evaluative internal dialogue. This is the anxiety and disruption that makes self-awareness problematic.

Psychotherapy presents a prototypical environment in which such harsh self-scrutiny is likely. The Other is ever-present and ever-attentive. Even more than in typical social interaction, the focus is kept on the patient. The distractions and pleasantries 12 that punctuate normal conversation are minimized. The therapist, to the extent that he or she approximates the analytic ideal of the

"blank screen", provides few clues as to course or satisfactory progress of the relationship. In the absense of clear signals from the attentive expert, the patient is even more likely to obsess about the adequacy of himself and his self-presentation. All of these conventional features of the therapy relationship compound the tendency toward self-consciousness that follows naturally from the discussion of sensitive and problematic areas in the patient's life. It is no wonder that patients demonstrate an amazingly crea tive and broad range of resistances to the therapeutic task, or that-- with great frequency-- the therapy ends with the early and unilateral termination of treatment by the patient.

This is, intentionally, a one-sided portrait of the therapy relationship. But its purpose is to highlight the likelihood that shame and anxiety will be important issues in the early phases of psychological treatment. To the extent that insight is a basic feature of verbal psychotherapy, self-consciousness, with attendant self-criticism and shame, will be present. Indeed, a wide range of therapeutic manuevers may be interpreted as vehicles to manage anxiety. Yalom (1975) stresses the curative power of discovering that a private pain is shared in common with other group members.

This experience of universality deflates shame by recasting a private problem as common. Likewise, an empathic response indicates that an experience is understandable and acceptable. There is the clear 13

implication that the problem expressed is neither so repugnant nor

so important as the patient's fantasy had likely made it seem. The therapeutic use of humor and anecdotes distract the patient and disrupt the chain of self-evaluation as much as they imply alterna tive perspectives for analyzing a problem. Perhaps most directly, paradox and reinterpretation suggest socially valued motivations that can replace the patient's self-critical explanations of behavior.

Diverse techniques, even from incompatible approaches to personality change, share in common their efficacy at disrupting self-scrutiny or neutralizing its troublesome consequences.

While therapists seldom address the issue of self-conscious ness, there is discussion of the ways in which focal self-awareness changes over time. Greenson (1967) pegs the development of a working relationship between patient and analyst to the capacity of the observing ego to split off from the experiencing ego. As this split develops, the observing function, called the reasonable ego, operates as the analyst's ally in making and applying interpretations. What

Greenson is describing is a gradual process by which the anxiety and shame attending self-scrutiny are neutralized and self-observation becomes a useful tool rather than a threat. But the development of the reasonable ego is not regarded as the culmination of successful treatment. Rogers (1958) notes that as client-centered therapy proceeds there is movement on the part of the client from initial avoidance of reference to the self, through a period of focal self awareness, to a final stage in which reflective consciousness becomes 14 less important as the organismic valuing process asserts itself without requiring focal attention.

A similar conception is offered by Bugenthal (1965), who posits a two-stage therapy process:

The analytic phase of our work is oriented to a conception of the person in which he is assumed to have both subject and object aspects. The ontologie phase calls for a recognition that this composite conception of the person is itself a part of the resistance to full authenticity of being. Instead, the ontologie phase insists on the necessity of freeing the I from its bondage to the Self in order to move toward true ontologie freedom. (P. 184)

Here the "I-process" is "pure subject" (p. 194) or "feelingful awareness" (p. 204). In contrast, "the Self is an object in the awareness of the I-process. Under the influence of cultural learnings, we often regard the Self as synonomous with the I-process, but it is not" (p. 203). Rather the Self serves as a cognitive map permitting consistency of behavior with minimal awareness. Both excessive consistency and minimization of awareness are threats to authentic living. In addition, the Self is constantly checked against an idealized version, promoting a tyranny of the ideal over the actual.

Therapy then becomes a reworking of the Self which allows maximum awareness and culminates in the liberation of the I-process from the

Self. To Bugenthal, peak experiences occur at "moments in which the

I-process breaks free from the Self" (p. 252).

Bugenthal and Rogers agree that therapy must include a phase in which the self-as-object is carefully examined, but that such examination cannot be the end point of treatment. Fully human living requires a more spontaneous responsiveness than can be 15 possible with a self-scrutinizing posture toward life. In adopting this stance, they are in line with those who question the absolute value of self-awareness. Yet, they clearly reject the Sartrean notion that self-apprehension, or the presence of other people which stimulate it, is undesirable or intolerable. There seem to be two reasons for their lack of emphasis on the shame and anxiety accom panying self-consciousness. One reason is legitimate, while the other is nat.

Sartre is wrong in contending that social relationships are inevitably fraught with shame and violated freedom. As Mead and others have demonstrated, all thought is fundamentally social

(Pfuetze, 1954). The internal dialogue of thinking is modeled on social dialogue. Our freedom is not contravened by the Other, but is made possible by him. Shame, then, is not an ontologically pri mary experience, but one among a number of potential responses to the presence of an observer. Similarly, self-evaluation is but one of a range of ways of responding to the objectified self. To return to a previous example, once again picture our dancer. Instead of disrup tive shame, the awareness of an observer may motivate greater effort and concentration on the dancer's part. Whether for vanity or in appreciation of imagined approbation, the dancer may excell himself.

Especially if he is experienced and competent, the dancer may be able to use the other's reactions as a type of mirror to inform his efforts.

Here the Other becomes a source of information or support for the enhancement of the dancer's project. This capacity to productively 16

use the presence of another, and the entire process of self-observa

tion, is what makes psychotherapy possible. Greenson (1967),

Bugenthal (1965), and Rogers (1958) are correct in stressing the

potentially beneficial consequences of self-examination. What

becomes important is to discriminate the conditions or situations in

which self-consciousness will support insight and change, in contrast

to those times when it will erupt into paralyzing shame.

Where all three therapists underestimate the importance of

shame is in the first few sessions of psychotherapy. Clearly with

Greenson, and to a lesser extent with Bugenthal and Rogers, the

emphasis is on the course of treatment once a therapeutic alliance

has developed. It is then that interpretation is most useful, that

the I-process can be nurtured, or that the organismic valuing process

becomes evident. But the working relationship develops slowly. It

is during the beginning sessions, when the patient makes the first

tentative attempts at self-disclosure, that the vulnerability to

anxiety is greatest. The developed transference, if positive, is a

bond which enables the patient to withstand periods of self-conscious

ness. Before it has arisen the patient can tolerate considerably less

self-scrutiny and the therapist must sensitively administer corrective

doses from his or her anxiety-management techniques. Premature

pressure for self-examination can easily result in the experience of

shame and coercion that Sartre so eloquently describes. It is during

this early phase in treatment, a phase with varying duration for different patients, that the Sartrean paradigm is most applicable. 17

Self-focused attention is a double-edged sword. That it can cut for good has been amply demonstrated. It is the vehicle for conscience and social cooperation. It is a necessary condition for the crucial internal dialogue between the "I" and the "me" (Mead,

1934). It is one--if not the--critical ingredient in verbal psycho therapy. It is this role that has received most attention by psycho logical theorists. But self-consciousness can have ill effects as well. Influential observers like Bugenthal and Rogers stress that treatment is not finished so long as the self-as-object is the primary focus. Perhaps even more important, according to the argument in this section, are the potentially disruptive effects of self-consciousness in the early stages of treatment before the therapeutic alliance is robust. Research bearing on these affects, both positive and negative, will be presented next.

Self-Awareness Theory

Duval and Wicklund's (1972) theory of objective self-awareness purports to have general application across the field of social psycho logy. The authors treat such classic subjects as conformity, social facilitation, attribution and self-esteem. This range is achieved because the authors conceptually link motivational consequences to the cognitive processes controlling direction of attention.

Their contention is that consciousness is intrinsically bidi rectional. During periods of action or sensory stimulation the person will be subjectively self-aware (SSA): attention will be directed 18 toward the external environment and there will be only a minimal, nonfocal self-awareness. However, any stimuli reminding the person of himself will prompt the uniquely human self-consciousness in which the person treats himself as an object. This latter state is what is called objective self-awareness (OSA). Consciousness is held to oscillate between these two points. While there are undoubtedly individual differences in the proportion of time spent in OSA, Duval and Wicklund contend that the direction of attention is primarily determined by environmental stimuli. They use such stimuli as mirrors, videotapes, and voice recordings in their experiments to manipulate attention toward the self.

It is the motivational consequences of attentional direction that gives the theory its wide range of application. Duval and

Wicklund theorize that people use internalized standards of action to regulate their behavior. It is while in the objective state, and only then, that actual behavior is compared with personal standards.

Typically there is an awareness of discrepancy between behavior and ideal. Although Wicklund (1975b) notes that a success may lead to a pleasurable discrepancy when behavior supersedes the relevant stan dard, the usual experience is one of failing to live up to the standard. In such cases objective self-awareness would lead to self derogation. The drop in self-esteem can be handled in various ways: attempts to avoid OSA may include withdrawal or distraction through involvement in other activities, or dissonance may be reduced by altering standards or changing behavior to bring it in line with ideals. 19

Whichever of the options is chosen, the motivational conse quences of objective self-awareness are clearly based on the desire to avoid the experience of discrepancy between behavior and standards.

Duval and Wicklund (1972) suggest that redirection of attention away from the self will be the initial response; it is the fastest means of avoiding awareness of discrepancy. So the embarrassed partygoer will try to escape from the social setting. Or, failing that, he will become engrossed in other activities (any other) such as mixing a drink or admiring the host's preparations. Anything that shifts the attentional focus away from the self restores OSA and provides immediate relief from self-derogation. Only when such escape is impossible would behavior or attitude change be expected. So persis tent confrontation from a friend may well prove effective, since the subject of the discrepancy cannot be avoided and the valued friendship will prevent fleeing the situation.

Such a formulation implies a preference for subjective self awareness, with its freedom from self-evaluation. This raises two crucial questions. First, what is to prevent a constant attentional focus on external events? Duval and Wicklund respond the stimuli to OSA are unavoidable. Anything that reminds the person of himself will elicit self-focused attention. Environmental stimuli such as mirrors, monitors, or personal writings raise the self to consciousness.

The very actions sought to avoid OSA may require the focus on one's hands or kinesthetic cues that trigger self-consciousness. But, the most obvious and unavoidable stimuli to OSA are social. Awareness 20

of self as a separate object arises in interaction with other people

(Mead, 1934; Piaget, 1966; Smith, 1979). The experience or anticipa

tion of evaluation by others remains the most common and surefire way

to elicit OSA. Given our constant exposure to a social network, and our dependence on it to meet so many of our goals, OSA is unavoidable.

The second crucial question might be: if OSA is unpleasant, why do people seem to seek it? People are constantly and voluntarily performing before audiences. They decorate their dwellings with mirrors and their offices with certificates enblazened with their names. They seek introspective activities, even to the extreme of paying for interminable courses of psychoanalysis. Here the answer is vaguer. Wicklund (1975b) suggests that the motivation for self improvement may overcome the tendency to avoid OSA. The research to be discussed below also suggests that people may rapidly habituate to the experience of self-focused attention and thus sustain it without adverse reactions.

There is a further explanation of why OSA is not always avoided. While Duval and Wicklund (1972) never explicitly discuss the possibility, it is clear that self-evaluation need not be the inevitable result of self-focused attention. The dancer described above may very attentively consider how he appears to his audience, and use this awareness to enhance his performance without engaging in self-criticism. Self-consciousness may be necessary for self evaluation without inevitably causing it.

Having attempted to outline the scope and central tenants of 21

Duval and Wicklund's theory of objective self-awareness, and having

just suggested a qualification to it, we will review a cross-section

of the research cited in its support. Because of the realization that

OSA need not be inevitably or constantly aversive, we will monitor

the magnitude and duration of the OSA effect whenever possible.

Demonstrations of objective self-awareness. Self-awareness

research manipulates the salience of stimuli reminding the subject

of the self. Most often the manipulation uses a mirror, a video

camera, or an audience. Davis and Brock (1975) demonstrated the validity of the camera in stimulating objective self-awareness (OSA).

The 48 subjects were told to translate pronouns from a passage in a foreign language, with the expectation that self-awareness would be evident in increased use of first person pronouns. Some subjects were given spurious feedback about their level of creativity to induce either positive or negative discrepancy from standards. But even when there was no feedback and no induction of self-evaluative posture, exposure to the camera led to a significant increase in the use of first person pronouns. Of a total of 49 pronouns, the control group translated 9.5 to the first person compared to the 14.2 used with exposure to the camera. Gellerand Shaver (1976) used a more projec tive approach, the Stroop color-naming technique. Previous research indicates a reliable tendency for it to take longer to name the color of a stimulus card if a word printed on the card has important personal meaning. Any OSA induction should increase the self-relevance of stimulus words, making the procedure a good indirect test of the 22

presence of OSA. Half of the 57 female undergraduates were admin

istered the procedure in the presence of a mirror and camera. Ad

ditionally, half the group was exposed to neutral words while the

other half saw self-evaluative and self-referrent words on the

stimulus cards. Table 1 shows the seconds required to name the

colors. There were significant main effects for type of word and

for OSA stimuli, with exposure to either causing greater latencies

in naming colors. In line with self-awareness theory, the results validate the OSA-inducing properties of mirrors and cameras, speci

fically, and of self-referrent stimuli in general.

Carver and Scheier (1978) validated the OSA-inducing proper

ties of a mirror or an observer. Female subjects filled out Exner's

(1970) Self-Focus Sentence Completion blank, which can be scored for reference to self, to the environment, to both, or to neither. For each subject a ratio was created with references to self or self plus environment divided by total responses. With exposure to a mirror, 41 undergraduates generated a ratio of .57, significnatly higher than the .50 for 38 students not exposed to the mirror.

Similar results were obtained when OSA was manipulated by the presence of an observer: a ratio of .58 for the 41 observed subjects differing to a statistically significant degree from the .51 obtained by stu dents who had no audience. In summary, a variety of approaches have shown that the self becomes more salient with exposure to any of the most commonly used stimuli to OSA: the camera, mirror, or observer.

Duval and Wicklund (1972) demonstrate that the responses to 23

Table 1

Response Latencies Under

High and Low Self-Awareness Conditions

Word type Mirror and camera No mirror or camera

Neutral 8.91 6.25

Self-referrent 13.36 8.57

Note. From Geller and Shaver, 1976. 24

their manipulations of self-awareness have implications in a number

of important areas. Examples will be drawn from studies on attribu

tion of responsibility, social facilitation, and self-esteem change.

Duval and Wicklund (1973) had female undergraduates read

five short passages in which a positive outcome could be plausibly

attributed to oneself or to another. The subjects were asked to

apportion responsibility. The self-awareness manipulation consisted

of exposure to the reflecting (OSA) or nonreflecting surface of the

mirror. As hypothesized, the OSA condition lead to greater self

attribution. With exposure to the mirror, the subjects ascribed

60% of the responsibility to self, while only 49.9% was self

attributed by subjects not exposed to the mirror. These results

cannot be explained as simple self-aggrandizement. Focus of attention was equally potent when outcomes were negative. Objective self

awareness again increased self-attribution, with an average of 60.2%

of the culpability assigned to the self by OSA subjects, as compared

to 51.1% for the no-mirror subjects.

Responsibility attributed to self can be diminished by stimuli

directing attention away from the self. Duval and Wicklund (1973)

used physical activity as a distraction. Under the guise of warming

up for a manual dexterity task, 14 undergraduate subjects operated a pursuit-rotor device while apportioning responsibility for negative

outcomes in 10 situations. Self-attribution was lower than that obtained from 16 control subjects in 8 of the 10 situations. Overall, distraction lead to a significant decrease in assigning responsibility 25 to self, from 57.63% for the controls to 48.65% for the OSA group.

Duval and Hensey (1976) and Duval and Wicklund (1972) detail similar results in experiments using other types of competing activities to increase subjective self-awareness.

These experiments indicate a consistent and statistically significant effect in which manipulations of self-awareness alter the degree of self-attribution. The fact that ratings of personal respon sibility were alternately increased and decreased in keeping with theoretical expectations and that these changes occurred for both positive and negative outcomes, suggest that the effect is pervasive.

A second application of self-awareness theory is in the explanation of social facilitation effects. Zajonc (1965) and Cottrell

(1968) observed that the presence of an audience could lead to greater efforts and increased productivity in certain types of activities.

The effect was attributed to a nonspecific drive aroused in the social context. Duval and Wicklund (1972) reinterpreted these experiments as evidence of self-awareness effects. They argued that an audience, particularly if it is attentive and expert in the area of performance, serves to elicit OSA. Attempts to approximate standards of excellence in turn lead to improved performance. By this analysis, "social facilitation" effects should be found even with nonsocial stimuli, so long as they lead to self-focused attention. Wicklund and Duval

(1971/1972) told 34 female undergraduates that they were participating in a study of the ease of copying unfamiliar passages. The OSA stim uli, a mirror, was justified with the explanation that a second phase 26

of the session involved mirror-writing. All subjects then copied

selections of German prose for successive, 5 minute sessions, with

the experimental group being exposed to the mirror during the second

interval. As predicted, the OSA group showed more improvement in num

ber of letters copied (43.33 letters more than during the first session)

than did the control group (17.65). The group difference was statis

tically significant at the .05 level. Again, self-awareness theory dem

onstrates its explanatory value in another area of social psychological

inquiry.

Further research on social facilitation qualifies the OSA ex

planation. In a partial replication of Wicklund and Duval's (1971) study, Pasternak (1978) had 80 undergraduates copy constant-vowel-con

stant syllables with low association value. Half of the subjects were aware of being videotaped during the experiment. Yet no group differ ences were found. The performance facilitation effect may not be espe cially robust. Others question the range of application of the self awareness account. Citing Wine's (1971) argument that test anxiety causes performance decrements because it interferes with attending to the task, Liebling and Shaver (1973) hypothesized that OSA may interfere with performance if it raises evaluative apprehension beyond an optimal point. They replicated Wicklund and Duval's (1971) experiment closely

(having subjects copy passages in Swedish) but added an additional group exposed to OSA under conditions intended to heighten evaluation.

This group of subjects were told that the task reflected reaction time, verbal fluency and language ability and therefore "might reflect their 27

intelligence." A significant interaction was found as predicted.

The replication condition resulted in more improvement with exposure to the mirror, but the high evaluation condition lead to less improvement (7.26%) with the mirror than without it (13.69%). Such findings caused Wicklund (1975b) to revise self-awareness theory to say that:

the relationship between self-focused attention and task per formance almost has to be curvilinear, and, further, the simpler the task • • . the more that attention can be turned to the self before reaching the point at which that attention comes to inter fere with performance. (p. 265)

Self-awareness theory has demonstrated its utility in predicting facilitation with simple and nonstressful activities. With more complex and self-evaluative tasks (self-exploration in psychotherapy for example), excessive OSA may prove counterproductive.

One last area in which self-awareness effects have been demonstrated is in the crucial area of self-esteem. Ickes, Wicklund & Ferris

(1973) reported two similar experiments in which subjects rated themselves as they ideally and as they actually were on continuums between pairs of traits. Half of the subjects completed one of the scales while OSA was created through the use of audio tape recordings of the subject's voice. In both studies, OSA lead to nonsignificantly greater discrepancies overall and to significantly greater self-ideal differences over the first block of five items. For the first block in study I, the mean discrepancy on the 20 point Lickert-type scale was

4.0 for the OSA condition as compared to 2.6 for the group listening to another student's voice. The differences were attributable to real-self ratings, as the ideals remained consistent across groups 28 despite exposure to OSA. So subjects appear to respond to OSA by becoming temporarily more self-critical. The authors hypothesized that the diminution of the initial effect can be attributed to habitua tion or to self-distraction from the stimulus to OSA.

Crandell (Note 1) performed a study which partially replicated

Ickes, Wicklund, & Ferris' (1973) research on self-ideal discrepancies.

Subjects selected by 26 student experimenters were told they were participating in a study of nonverbal behavior and nervous habits.

With this rationale, 51 subjects were observed closely while they indicated their actual and ideal standing on one set of 48 items drawn from the Gough Adjective Checklist (Gough & Heilbrum, 1965). A second set was completed with no observation, thus permitting each subject to serve as his or her own control. The State-Trait Anxiety Inventory

(Spielberger, Gorsuch, & Lushene, 1970) was also administered before debriefing. The manipulation had no effect on the discrepancies, which averaged 1.73 per item with OSA and 1.76 without it (on a 9-point scale).

Even in the first block of 12 items, the OSA group was only minimally and nonsignificantly more discrepant, averaging 1.82 instead of 1.80 per item. This study offered no support for self-awareness effects on self-ideal congruence. However, OSA did affect ratings of state anxiety. If the subject was observed while completing the first self concept scale, and accordingly had an intervening and unobserved ques tionnaire to complete before rating their anxiety, A-State averaged

34.90. When anxiety ratings were immediately preceded by observation,

A-State averaged 40.03. The difference is marginally significant, 29

! (1, 48) = 3.7~~< .10. There appears to have been OSA effects in

this experiment, although they were not evident in the self-ratings.

Perhaps the stimulus items (e.g., practical, adaptable, sociable) are

too consistent in the subject's life to be greatly affected by OSA.

In other words, trait ratings are less responsive than are state

ratings (such as the anxiety measure).

The link between OSA and self-concept or emotional states has

been found only occasionally. Collins (1976) found no relationship between OSA and self-esteem among high school students. Brehm (1974) failed to elicit self-derogation among 9th graders with self-focused attention. Carver and Scheier (1978) found no increase in negative affect in sentence completion tests administered in the presence of a mirror. And Wolfe (1975) failed to find any link between OSA and ratings of mood among 120 female college students.

However, Scheier and Carver (1977), alternately using mirror manipulated OSA and a measure of dispositional self-consciousness, reported consistent results across several experiments showing that

OSA heightens awareness of emotions. It need not be inevitably nega tive. OSA led to greater appreciation and responsiveness to slides of nudes. But it also caused significantly greater response to statements eliciting depression. Scheier and Carver's conclusion reasserted the relationship of OSA and emotions, suggesting that attention may shift to emotions as easily as to standards during periods of self-focused attention.

Studies that fail to replicate or discover expected results 30

suggest that OSA effects are elusive and seldom robust. Wicklund

(1975b) explains that such effects are temporary and may be easily missed if the dependent measure is not properly timed. The theory predicts the discomfort of self-consciousness will motivate attempts at discrepancy reduction or distraction, resulting in rapid disappear ance of the usual sequelae to OSA. There is a further comment needed on the self-esteem findings. In the self-ratings on the dependent variables, subjects were typically asked to rate central traits: enduring and often important aspects of themselves. It is not sur prising that they show little change. Crandell's (Note 1) research provided a comparison between state and trait variables and found little change in Trait Anxiety (40.90 immediately after OSA as com pared with 38.15 when there was an intervening task, a finding with a chance probability of .36) despite a marginally significant increase in state anxiety. Trait ratings are a stern test. That they show any change in response to attentional state indicates that OSA causes powerful, if temporary, alterations in self-concept.

The results obtained by Ickes, Wicklund, & Ferris (1973) and

Crandell (Note 1) confirm that OSA has its maximum effect at the time of initial stimulation and that its potency dissipates rapidly.

While OSA effects are of moderate magnitude and transient, they have been demonstrated in a wide range of circumstances. The studies reviewed suggest that self-awareness theory is broadly appli cable within personality psychology. Selected research has shown its contribution to the understanding of self-attribution, social 31 facilitation, and the self-concept. Further discussion could docu ment its applicability to such central areas of social psychology as conformity ~uval, 1972; Wicklund & Duval, 1971/1972), cognitive dis sonance ~nsko, Worchel, Songer, & Arnold, 1973; Wicklund & Duval, 1971/

1972), and accuracy of self-reports (Carver, 1975; Pryor, Gibbons,

Wicklund, Fazio, & Hood, 1977; Scheier, Buss, & Buss, 1978; Turner,

1978).

Importance of standards during objective self-awareness. It is the awareness of discrepancy from pre-existing standards which is assumed to cause the self-derogation, with the resulting attempts to escape objective self-awareness (OSA) or change. So it is worth digressing to examine our understanding of these standards. While they may be modeled on social norms, it is the individual's standards that are assumed to be the central referent for behavior during periods of OSA. Research by Carver (1975) supports this account. He selected subjects representing extremes in their belief in their use of punish ment to influence behavior. When instructed to use punishment in a

"pseudo-teaching" task, the two groups varied in the intensity of shock delivered to the erring "student", with the punitive subjects delivering significantly more, as expected. However, this occurred only when the subjects observed themselves in a mirror. When there was no stimulus to OSA, the groups delivered punishment of very similar intensity. This research suggests that it is, indeed, the personal standards that are activated by OSA.

Other researchers (Scheier, 1976; Scheier & Carver, 1977) 32

indicate that behavior may be guided by other aspects of the self than

internal standards. They argue that most research highlights the sa

lience of the standard through feedback on the dimension in question

or demand characteristics of the experiment itself. The salience

manipulation may be the crucial element. They have reported experi

ments in which the salient dimension was the subjects' emotional

state. Scheier (1976) examined aggression, in this case monitoring

the intensity of shock administered to a "confederate-student" who

first had taunted half of the subjects. There were no significant

differences between the angered and not angered groups except when

a mirror was introduced. With OSA, the angered subjects significantly

increased the intensity of shock, while other subjects decreased it

slightly. Scheier concluded: "when affect is strong, awareness is

directed to the affect, not the standard, and discrepancy reduction

does not occur" (pp. 637-638). Scheier and Carver (1977) showed that

OSA can heighten responsiveness to a variety of emotions, including attraction, elation, and depression. In keeping with their findings, a revision in self-awareness theory is warranted. Objective self awareness appears to heighten consciousness of whatever dimension of the self is salient. Neither self-evaluation, nor self-derogation have been shown to be certain concommitants of OSA. So escape behavior or discrepancy reduction are to be expected only when the ideal is more

salient than affective reactions.

Clinical implications of research on self-awareness. Some of

the effects suggest an important link with research areas and concepts 33

important to clinical psychology. The findings of the curvilinear relationship of OSA and self-evaluation with performance is highly similar to that found in research on Manifest Anxiety (Byrne, 1974).

Indeed, the disruptive effects accompanying OSA are reminiscent of those accompanying high evaluative apprehension (Mandler & Sarason,

1952; Wine, 1971). Objective self-awareness is a source of anxiety, and may prove a necessary condition for anxiety of high levels or long duration.

Self-ideal discrepancies, such as those elicited by Ickes et al. (1973) have been accepted in clinical research as an operationalization of self-esteem and emotional adjustment (Block & Thomas,

1955; Chase, 1957; Dymond, 1954). Self-ideal discrepancy reduction has been considered the primary criteria for improvement in client- centered psychotherapy (Butler & Haigh, 1954; Rogers & Dymond, 1954).

Now OSA has been shown to cause temporary increases in the self-ideal discrepancy. The implication of these lines of research are clear: self-focused attention may be a source of disruption in psychological functioning.

Psychotherapy aims to decrease self-ideal discrepancy. Yet, the process entails the type of intensive and extended self-scrutiny on the part of the client that assures considerable OSA. Temporary sequelae of self-focused attention, such as anxiety and a drop in self-esteem, may well interfere with therapy.

Avoidance of objective self-awareness. Another feature of self-awareness theory with important implications for psychotherapy

!, ~ .. 34

is the specification of responses expected after exposure to stimuli

heightening objective self-awareness (OSA). Behavior change is the

goal of psychotherapy. Yet, Duval and Wicklund (1972) suggest that

this response is less likely than attempts at OSA avoidance. Research

does show attempts to escape self-focused attention. Cummings (1976)

found that exposure to an observer and a video camera was a noxious

stimulus which could be used as a punisher to delay a contingent

behavior or condition an escape response. Duval, Wicklund, and Fine

(1971/1972) directly measured avoidance as a function of OSA versus non-OSA and high or low discrepancy from standards. Students were administered a personality test, which they were told was correlated with creativity and intelligence. The degree of discrepancy was manipulated by telling subjects their test results indicated they were

in the top or bottom 10% on the dimensions. Then, under the guise of

participating in a second experiment, they were led to another room.

For subjects in the OSA condition, the room contained a mirror and a camera. Subjects were asked to wait for about 5 minutes and then seek

the second researcher if he had not yet returned to administer the experiment. Table 2 shows the length of time before the subject left the room. Two findings are clear. First, OSA was tolerable for a shorter time if there was high rather than low intraself discrepancy,

£< .05. Second, subjects made a quicker exit when exposed to GSA inducing stimuli (significantly so for the high discrepancy condition).

Gibbons and Wicklund (reported in Wicklund, 1975b) observe the same pattern of results (with a significant overall interaction) when the 35

Table 2

Effect of Self-Awareness and Discrepancy

from Standards on Minutes of Participation

Discrepancy level

Self-awareness stimuli High Low

Mirror 6.39 (_g_=13) 7.80 (_g_=12)

No mirror 8.12 (_g_=12) 8. 20 (_g_=15)

Note. From Duval, Wicklund, and Fine (1971/1972). 36

dependent measure is the proportion of time spent listening to one's

own versus anothers voice during a 12 minute period following OSA and

discrepancy manipulation. These experiments indicate OSA will lead to

avoidance or escape behaviors.

There is less evidence on the relative priority given to

discrepancy reduction versus GSA-avoidance when both options are

available. Indirect evidence is provided in a report by McDonald

(1977). Subjects were asked to write responses to a visual cue after

having been given positive or negative feedback on their performance

on a "creativity test." Half of the 96 female undergraduates were told

their response was related to creativity. A mirror was used to create

OSA for half of the subjects. Avoidance of objective self-awareness might appear in shorter responses, while discrepancy reduction would

be expected to be evident in longer responses. In examining writing

time, there was no main effect related to self-awareness, and so no

support for an avoidance account. The author reported that among the

OSA subjects, however, those with negative feedback wrote more than those given a high rating on creativity. While no significance testing was reported for this interaction, the implication is that discrepancy reduction is the prepotent response.

A different result was obtained in a study by Liebling, Seiler and Shaver (1974). Sixteen subjects were observed during each of two

30 minute sessions of listening to music. During one period, OSA was induced through the presence of a mirror. All subjects were smokers, including 14 who had indicated an ideal would be to decrease the 37 frequency of cigarettes smoked. Despite this goal, the stimulus to osA led to significant increases in the number of cigarettes lit, number of puffs, amount of time holding a cigarette, and the number of flicks. The authors discussed their findings as evidence of the drive inducing effects of OSA. But Wicklund (1975a) pointed out that behaviors also served as a distraction, and hence a means of avoiding

OSA.

The two experiments reviewed offer apparently conflicting results when discrepancy reduction and OSA-avoidance lead to opposing behavior. However, the studies differed in the salience attached to the discrepancy in question. In one, the dimension (creativity) was directly manipulated while the other downplayed it (by embedding the query about ideal smoking among a number of questions about attitudes).

It may well be that the likelihood that OSA will be avoided is inversely related to the salience of the discrepancy. Avoidance may be more likely if the self versus ideal self dissonance is not focal.

Duval and Wicklund (1972) outline three conditions that increase the likelihood that an individual will attempt to escape from a situa tion. First, there must be discrepancies between behavior and rele vant standards. Second, there must be OSA. Third, discrepancy reducing behavior must be difficult or impossible. This last condition was met in the Duval, Wicklund and Fine (1971/1972) experiment by focusing on enduring abilities and restricting the environment so that there were no discrepancy-reducing options available.

The proposed experiment on premature termination from psycho- 38

therapy closely follows this outline, not because of experimental

manipulation, but because the client's phenomenal situation at the

beginning of therapy closely corresponds to that described above.

There is a real, important, and painful discrepancy between the clients

actual and ideal level of emotional adjustment. It is usually a

problem of chronic duration. Should the patient opt for a discrepancy

reducing tact--such as denying the severity of the problem, attributing

it to a temporary situation, or projecting responsibility ("if only my wife ••• ")--the therapist firmly confronts the patient with his

central and continuing involvement with the problem. Lastly, there is

intense and continuing self-focused attention. There is the expert observer, closely monitoring the patient's self-report and continuously

redirecting the patient's focus to his own statements. All of these

features of the beginning therapy relationship increase the likelihood that OSA will prove intolerably painful, and that the would be client will flee therapy.

Clearly, not all clients terminate prematurely. Nor would self awareness theory lead to such a blanket prediction. Implicit in

Wicklund's (1975a) comments on the salience of intra-self discrepancy, is the expectation that clear identification of an area of dissonance will lead to corrective action rather than flight. The research on performance facilitation effects of OSA suggests that some OSA (like some evaluation apprehension) may be productive rather than disruptive.

The literature reviewed also indicates that the disruptive effects of

OSA are likely moderate and transient. Clinicians have certainly 39 developed tactics for gradual exposure to OSA and for anxiety manage ment, all of which aid the novice patient in adjusting to a stressful role.

The current extrapolations from the social psychology of self awareness theory to hypotheses about behavior in therapy is possible because of similarity between the experimental design used to study

OSA avoidance and the situation facing the new client. Just as Duval,

Wicklund, and Fine's (1971/1972) subjects were able to tolerate discrep ancy and OSA for some time, although it was less than the other group withstood, so the effects of OSA are less disruptive to some new patients than to others. Self-awareness theory does not predict 100% dropout. But it does offer a testable explanation of much of the early termination that is found: if the beginning of psychotherapy fosters excessive self-consciousness and if discrepancies addressed are not easily resolved, flight from therapy is likely to occur.

Individual differences in self-consciousness. Duval and

Wicklund (1972) examined situational variables effecting OSA levels.

This choice was based, in part, on their belief that direction of attention is potently determined by environmental stimuli. It is also advantageous that these stimuli are easy to manipulate, allowing for the true experimentation so necessary for theory building. But the focus on situational variables also results from a hesitancy to tackle the problems associated with examining individual differences.

To request a self-report on self-consciousness level would almost inevitably alter the variable of interest, confounding the answer. 40

And there are other "potential ambiguities":

it is difficult to know whether the measures are relevant to actual differences in self-focused attention, differences in types of personal standards or styles of discrepancy reduction, differences in ability to avoid self-focusing stimuli, or even theoretically irrelevant differences. (Wicklund, 1975b, p. 268)

These are impediments to the precise research needed to develop a comprehensive theory. But none of this implies that individual differences in direction of consciousness are nonexistent or unimportant.

Indeed, when the focus shifts from theory to application, the molar level of self-consciousness is more critical than which of the options suggested above accounts for it.

Fenigstein et al. (1975) have developed a scale to measure dispositional self-awareness. In contrast to the situationally determined focus of attention, they investigated the trait of self- consciousness and found it to be composed of three distinct, though related, factors. The first subscale measures private self-consciousness (PR-SC): "attending to one's inner thoughts and feelings"

(p. 523). The conceptually and factorially distinct second scale measures public self-consciousness (PU-SC): "a general awareness of the self as a social object that has an effect on others" (p. 523).

A third subscale captures social anxiety (SA-SC), or "discomfort in the presence of others" (p. 523). The Self-Consciousness scale, with subscales identified, is found in Appendix A.

Fenigstein et al. (1975) report that the subscales are reliable and consistent in internal structure when replicated with varying groups of normal undergraduates. There are significant, but low, correlations 41

between PR-SC and PU-SC, and between PU-SC and SA-SC, while private

self-consciousness and social anxiety are unrelated. Table 3 sum

marizes the subscale means and intercorrelates reported in the pub

lished literature. They generally confirm the initial findings.

Other research also replicates the absence of sex differences noted

by Fenigstein et al. (1975). Findings are interpreted to indicate

that self-consciousness is dichotomous, with private and public as

pects. Social anxiety is a possible, but not inevitable, consequence

of awareness of the self as social object. It is possible, then, that a mirror may elicit a different type of self-awareness than an audience, and that the latter may more readily lead to the negative consequences associated with the colloquial usage of "self-consciousness."

Fenigstein et al. (1975) suggests two principle uses for the scale. Self-consciousness, as a dispositional expression of self awareness, can be expected to mediate accuracy of self-description; inaccurate self-reports may reflect a lack of attention to self as object. Second, the Self-Consciousness scale (SC) can supplement or replace the reactive manipulations used to create OSA. Rather than inducing OSA, among other types of arousal, by exposing subjects to cameras or audio feedback of their voices, it is possible to select subjects whose high score on PR-SC or PU-SC indicates a preexisting tendency to be attentive to themselves. The growing body of research on self-consciousness has tended to combine attempts to validate the scale with research in these two areas.

Dispositional self-consciousness has been shown to effect Table 3

Means ::md Jnter:corrclations of the Priv:.te (PR-SC), Public (PU-SC), and Social Anxiety (SA-SC) Subscales ------Means Correlations

PR-SC and PR-SC and PU-SC and Reference Population N PR-SC PU-SC SA-SC PU-SC SA-SC <;A-SC ·------.Buss & Seheier n976> normal, UGS, F .23a

C.<~ rver & Gtt;ss (19i6) normal, UGS, H 105 25.4 19.5 11.6

C:lfVC'l' ~~ Scheier (1978) normal, UGS, E 68 .33 -.05 -.02

I'E•ui~'>teln (l9i9) normal, UGS, F 92 .17 .24

Fe•tl~~·tcin, Scheler, normal, UGS, M 179 25.9 18.9 12.5 r, Buss (197 3) normal, UGS, F 253 26.6 19.3 12.8 normal, UGS 452 .23 .11 .21 n.ormal, UGS, H 152 .26 -.06 .20

Scl1eler ~.!976) normal, UG:J 5!0 28.0 .34b Scheier, Buss, & Buss (197 8) normal, ur_:s 100+ 24.2 Schet,,r & Carver (1977) normal, UGS 26.0

:,che if~r:, Carver, Schulz, normal, UGS 160 .27 Gl:-1ss, F. Kat~ (l97o) Turn.-:!r (l

TurnPr f. P!~t erson (1977) nurmal 45 .57

TurnPr, !'cht~ ter, Carv~r, normal, UGS 1395 • 31 • lit • 21 & J d

No:· c. UGS reprC'scnt!! unrlcrgrurlu<'tes. 1~11ere appropriate, sex is de'Jignated by F for females and ~~ for males. avxtrPme scores on PR-SC. hrer 91 s;:bJects with extreme scores on PR.

··~lNtrts anci corr('L

In a follow-up study, Turner (1978) examined the correlations between observed and reported dominance of 62 undergraduates. Subjects were observed for 10 minutes of interaction with a male and female confed erate during which they had to engage in a problem solving discussion, respond to a period of silence, handle a tangential monologue, and intervene in a disagreement. Their dominance was then rated by the confederates and two judges, whose highly consistent ratings were combined and correlated with the subjects' previous self-ratings.

Again, the correlations were higher for subjects above the mean in

PR-SC (.47 versus .24) or below the mean in PU-SC (.46, as opposed to the correlation of .28 found among subjects high in public self consciousness). Self-reports were comparably accurate for all groups except those low in PR-SC and high in PU-SC. Scheier, Buss, and Buss

(1978) compared validity of self-reports obtained from those represent ing the upper and lower thirds of the PR-SC distribution. All 63 44 undergraduates completed the overt aggression subscale of the Buss

Durkee Hostility Inventory. Several weeks later,measures of the intensity of shock delivered to a "learner-confederate" were obtained for each subject. Overall, the self-report correlated only .34 with observed aggression. Those scoring high in private self-consciousness averaged .66, reporting with significantly more accuracy than low

PR-SC subjects, whose correlation was only .09. There was no comparable effect when accuracy was tabulated for those above (.38) and below (.31) the mean in PU-SC. This last finding may have resulted because the lack of social interaction involved in the aggression experiment mini mized awareness of the self as social object.

The three experiments reported indicates a clear relationship between self-consciousness and accuracy of self-report. Whenever groups representing the extremes on PR-SC are compared, significant differences are found. Scheier, Buss, and Buss (1978) reported that those high in PR-SC give significantly more statements than low PR-SC subjects when asked to describe themselves. "These data, taken together with Turner's findings argue strongly for the use of private self consciousness as a moderator variable in the relationship between self reports and observed behavior" (p. 139). These experiments indicate that the Self-Consciousness scale is a valid instrument in tapping the degree to which normal subjects are aware of their feelings and response tendencies.

Other research correlates SC subscale scores with measures of related characteristics in an attempt to determine the convergent and 45

discriminant validity of the instrument (Campbell and Fiske, 1959).

Among the scales that should not correlate is social desirability

(Marlowe and Crowne, 1964). Turner, Scheier, Carver, and Ickes

(1978) report low to nonsignificant correlations for two undergraduate

groups as indicated in Table 4. Self-consciousness is also relatively

independent of intelligence, as measured by the Otis Quick-scoring

Mental Ability Test (Otis, 1954). Carver and Glass (1976) find an overall correlation of -.06 in a study of 105 male undergraduates.

Only SA-SC, for which£= -.21, ~ (.05, is related. The authors suggest that lower intelligence may lead to a lack of social competence and resulting anxiety. Alternatively, anxiety may have interfered with performance in the evaluative context of test-taking.

Surprisingly, self-consciousness is minimally related to test anxiety (Mandler and Sarasen, 1952). Table 5 shows the low inter correlations reported in two published studies. The previous discus sion shows that the effects of OSA approximate those of evaluative apprehension. Yet, the lack of substantial covariation with test anxiety "suggests the scale is specific to social anxiety" (Carver and Glass, 1976, p. 172).

Self-esteem and OSA are related (Ickes, Wicklund, & Ferris,

1973), so the Self-Consciousness scale would be expected to correlate with the Self-Esteem scale (Morse & Gergen, 1970). Turner et al.

(1978) report significant relationships of -.26, -.26, and -.35 between self-esteem and PR-SC, PU-SC, and SA-SC respectively. For the 505 subjects involved, self-consciousness was associated with 46

Table 4

Correlations between Social Desirability

and Subscales of the Self-Consciousness Scale

Scale Sample 1 Sample 2

Private Self-Consciousness .02 -.15

Public Self-Consciousness .06 .01

Social Anxiety -.03 -.23*

N 146 122

Note. From Turner, Scheier, Carver, and Ickes (1978).

*.E. (.05 47

Table 5 correlations between Test Anxiety and the Self-Consciousness Scale

Source

Carver and Glass Turner et al. Scale (1976) (1978)

Private Self-Consciousness -.11 -.02

Public Self-Consciousness -.01 .20*

Social Anxiety .14

Self-Consciousness Scale .00

N 105 258

*p<.01 48

lower self-esteem, again suggesting that self-awareness is a double

edge sword. Fenigstein (1979) failed to find a relationship between

PU-SC and three global self-esteem items. So it is inappropriate to

assume an inevitable covariation. But a drop in self-esteem is a

likely consequence of self-awareness induction, particularly if social

anxiety is elicited.

Other personality measures have been tested against the Self

Consciousness scale. Turner et al. (1978) report PU-SC and SA-SC

are moderately correlated with low masculinity and high femininity as

measured on Berns' (1974) Sex Role Inventory. Despite this, other

research (Fenigstein et al., 1975; Scheier & Carver, 1977; Turner et

al., 1978) reports no significant sex differences in scores on the SC

subscales. Achievement motivation might be hypothesized to underlie

self-awareness, particularly sensitivity to one's social impact. Yet,

no substantial relationship appears (Carver & Glass, 1976). Snyder's

(1974) Self-Monitoring Scale also assesses sensitivity to cues of

personal expressive behavior for social ends. Turner et al. (1978)

report results on a substantial sample that indicate significant but

low amounts of covariance. For 1094 undergraduates, self-monitoring

correlates .15 with PR-SC, .24 with PU-SC, and -.20 with SA-SC.

Fenigstein (1979) also reports a correlation of .24 for self-moni

toring with PU-SC. While expected correlations with these partially

related measures are obtained, the Self-Consciousness scale (SC)

proves to be distinct. The efforts at discriminant validation, reported above, indicate the SC cannot be readily reinterpretted to be a sub- 49

category of any of the other concepts investigated.

Convergence with related concepts also suggests the validity

of the SC subscales. Those high in private self-consciousness (PR-SC)

should create and use mental imagery and should be temperamentally

inclined toward reflectiveness and philosophical interests. Turner

et al. (1978) report that PR-SC correlated .30 with the Pavio (1971)

Imagery Scale and .48 with Guilford and Zimmerman's (1949) measure of

Thoughtfulness. Public self-consciousness (PU-SC) also correlates with Thoughtfulness (.22), but with more emphasis on social self awareness. Carver and Glass (1976) report low but significant cor relations with Buss and Plomin's (1975) measures of emotionality (.20) and sociability (.22). Turner et al. (1978) were unable to replicate the relationship with sociability; it was positive but nonsignificant.

Support for the validity of PU-SC is, accordingly, more tenuous. Of the subscales, Social Anxiety (SA-SC) is most strongly related to self-esteem (-.35) and test anxiety (.14 and .23). It is consistently associated with avoidance of social settings (~39 and -.46) and with inactivity (-.27). Such findings suggest an avoidant stance toward interaction in which anxiety interferes with optimal functioning (e.g., for IQ, £ = -.21). In sum, correlations with related measures offer at least moderate support for the validity of the subscales.

Stronger support for the distinctiveness and validity of each of the subscales is reported when they are used in the course of self awareness studies.

Of the subscales, PR-SC has been used most frequently. Buss 50 and Scheier (1976) replicated Duval and Wicklund's (1973) experiment in which the degree of self-attributed responsibility was assessed for scenarios with positive and negative outcomes. Subjects were selected from the highest and lowest thirds of the PR-SC continuum and randomly assigned to complete their ratings with a mirror either present or averted. The percent of self-attributed responsibility is presented in Table 6. A significant main effect was found for level of PR-SC and a marginally significant difference resulted from exposure to the mirror. The PR-SC performs in a manner similar to and yet more effec tive than GSA-induction. The PU-SC, which correlated .23 is PR-SC, did not prove significantly related to self-attribution, nor did its covar iance contribute to the power of PR-SC. Tangentially, it is worth not ing that the negativity bias reported with OSA also is evident in the ratings of high scorers on PR-SC.

Carver and Scheier (1978), in their study of OSA and the pro portion of sentence completion items referring to the self, also exam ined the effects of PR-SC. They found a significant correlation between SC and the ratio of self-referent to total statements, ~ (35)

= .29, £ <.05. For PU-SC, the correlation was .07. Table 7, again, indicates the power of dispositional relative to situational self awareness. It appears that PR-SC provides a ceiling such that OSA induction is effective only among those lacking the trait of self consciousness. Scheier and his associates (Scheier, 1976; Scheier and Carver, 1977; Scheier, Carver, Schutz, Glass, & Katz, 1978) also found that PR-SC can be used as an index of sensitivity to affect. Table 6

Percent of Self-Attributed Responsibility with

High and Low Self-Awareness and Private Self-Consciousness

Level of private self-consciousness

High Low

Self-awareness stimuli Positive Negative Positive Negative

Mirror 53 66 50 54

No mirror 56 55 46 53

Note. From Buss and Sceier (1976). 52

Table 7

Percent of Self-Reference with High and Low

Self-Awareness and Self-Consciousness

Stimuli to self-awareness

Private self-consciousness level Mirror No mirror

High .60 .61

Low .55 .44

Note. From Carver and Scheier (1978). 53

In parallel experiments, OSA induction or high private self-con-

sciousness proved effective in heightening ratings of anger, attrac-

tion, repulsion, and depression. They summarize the validity of

PR-SC as a measure of self-awareness:

The parallel between the mirror manipulation and private self consciousness is not limited to the present research. For example, it has been shown that exposure to a mirror increases the proportion of self-focused responses given on a projective test; private self-consciousness does the same. The mirror increases attributions to self; so does private self-conscious ness. The mirror facilitates angry aggression; private self consciousness does the same. And the mirror enhances the validity of self-reports, and likewise for private self-consciousness. (Scheier & Carver, 1977, p. 634)

There has been less research on Public Self-Consciousness

(PU-SC). Fenigstein (1979) notes that most experiments, even those ostensibly involving interpersonal behaviors such as aggression, are

structured so as to be impersonal. Accordingly, the absence of effects attributable to PU-SC is not surprising. Fenigstein's research directly exposed subjects to a group setting in which their comments were either responded to or ignored. Subjects then rated their attraction to the group, their desire to continue working with it, and the amount of self-attributed responsibility for the group's reaction to them. As expected, 40 female undergraduates scoring in the top third of the PU-SC continuum denigrated the group and personalized its behavior significantly more than 40 subjects in the lower third of the distribution. Private self-consciousness had a negligible effect on this pattern in an analysis of covariance, providing "further evidence of the discriminant validity of the two self-consciousness dimensions"

(p. 80). Fenigstein also included a direct query as to the amount of 54

time spent attending to oneself during the group session. There

were small (18.5% versus 15%) differences attributable to the accept-

ance versus rejection condition. But high PU-SC subjects reported

significantly more self-focused attention (24% of the time) than did

low scorers (9%). So the experiment validates the subscale leading

to the conclusion:

When we attend to ourselves in social situations, it is the social self that is focused ••.• In this respect, it has been shown that public self-consciousness (awareness of ourselves as social objects), rather than private self-consciousness (awareness of our internal selves) is the crucial determinant of how much the evaluation of others affect us. (Fenigstein, 1979, p. 85)

Results also paralleled simultaneously reported findings that mirror-

stimulated OSA lead to heightened reactiveness to positive or negative

evaluation in an interview.

Turner (1977) examined the contribution of self-consciousness differences to the finding that subjects will shift their attitudes in

the direction of an announced but not delivered persuasive communica-

tion by another. High PR-SC should limit the shift, because the subject

focuses on personal standards. But anticipatory belief changes were expected for subjects high in PU-SC, out of a desire to maintain a socially correct appearance, and SA-SC, to avoid attracting attention.

Those above and below the mean in PR-SC and PU-SC failed to differ significantly. The PU-SC dimension appears unrelated to experiencing

social attention as aversive. In contrast, those high in SA-SC showed the greatest change in the direction of the other's expected position.

This experiment represents a preliminary step toward validation of 55

SA-SC as an index of fear of social attention.

The preceding review leads to two clear conclusions. The

subscales of the Self-Consciousness scale are valid measures of dis

tinct components of self-awareness. While all are related to a cogni

zance of self, they represent distinct features of personality.

Using them, populations can be identified which are responsive in different situations and have characteristic expressions of their

self-awareness. Secondly, many of the phenomena surfaced using the

situational tact of stimulating OSA can be reproduced successfully by selecting individuals dispositionally inclined to particular types of self-consciousness. Those high in PR-SC parallel those exposed to mirrors or cameras in becoming attentive to internal affective states and cognitions. Those high in PU-SC appear to respond similarly to those in whom OSA is aroused by exposure to audiences or reflective stimuli in social situations. The explanation of results of research on SA-SC is similar to that offered for Duval and Wicklund's (1972) research on conformity. The theoretical explanation of this OSA research suggested that cognizance of discrepant views increased the likelihood of focal self-awareness and a desire to minimize disagree ment as a means of escaping aversive OSA. Similarly, those high in

SA-SC will alter their expressed views, at least temporarily, in an apparent response to anxiety aroused by social attention. The Self

Consciousness scale can be used as a relatively nonreactive alterna tive to situational manipulations of OSA in research on normal popula tions. 56

Clinical implications of self-consciousness. There has been inadequate research using clinical populations to understand the meaning of self-consciousness with these groups. The one published study (Turner & Keyson, 1978) does not provide information on norms for either the neurotic or schizophrenic groups sampled. Nor does it indicate the relationship of the subscales. It may well be that psycho pathological groups unite self-focus with anxiety. One purpose of the following research will be to examine the performance of the Self

Consciousness scale with a clinical population.

Turner and Keyson (1978) found that therapists could success fully predict the self-ratings of their patients for PR-SC and SA-SC, but not for PU-SC. The implication is that the first two subscales tap more clinically relevant phenomena that are therefore more focal in the course of therapy. This is consistent with the commonsense analysis which indicates that anxiety and awareness of internal exper iencing are core considerations in psychotherapy. Social anxiety, in particular, would be of concern to a therapist because of its disrup tive impact on the client's functioning. Yet, PU-SC may be an impor tant measure. The early stages of therapy represent the point at which self-consciousness before the therapist is at a maximum. It is at this time that some clients will experience the Sartrean entrap ment in the presence of the other that may cause them to flee therapy.

So, levels of PU-SC and SA-SC may both be important in determining length of stay.

All three subscales, and the total score on the Self-Conscious- 57 ness scale are hypothesized to be clinically relevant. All will be considered as predictors of premature termination from psychotherapy.

Additionally, different diagnostic groups are characterized by dif ferences in their propensity for and reaction to self-focused attention

(Kennedy, 1977; MacKinnon & Michels, 1971). The SC subscale profile may vary across diagnostic groups. Paranoid patients, whose defenses externalize responsibility, would be expected to be low scorers on all subscales. Obsessives and passive character types are hypothe sized to be generally high in self-consciousness. Obsessives should peak on PR-SC, but their SA-SC scores may be attentuated by the re liance on intellectualization. High PU-SC may be characteristic of hysterics and sociopathies, since the self-esteem of both groups is tied to their success in managing other's impressions of them.

Hypothesizing about the relationship between diagnosis and self consciousness remains speculative in the absence of additional re search on clinical groups using the SC scale. An additional goal of this dissertation project will be to provide empirical evidence on this subject.

But a preliminary step is clearly required. In the absence of previous published research on SC scores among psychopathological groups, norms and intercorrelations for the subscales need to be obtained. Specific hypotheses derived from self-awareness theory can be applied to clinical situations. But they will be useful only to the extent that the theory generalizes to a clinical population. The next section reviews the Ego Strength scale, which will be used in this 58

study of external validity.

~o Strength

In the previous section, the application of self-awareness

theory to clinical psychology was based on analogy. For example,

the situation that caused subjects to leave the experimental setting

in Duval, Wicklund, and Fine's (1971/1972) research is similar to the situation faced by a patient at the beginning of psychotherapy.

An empirical test of the generalizability of self-awareness theory is also possible. The central issue is the degree to which patients seeking treatment respond to the Self-Consciousness scale in a manner similar to that found for college students.

Barron's Ego Strength (ES) scale will be correlated with the

Self-Consciousness measure to examine the impact of degree of pathology or, conversely, the availability of ego resources, on the handling of self-focused attention. The Ego Strength scale (Barron, 1953) will be used as a global measure of the degree of ego impairment.

The following review argues that the scale, which has been used for multiple purposes over the last three decades, functions most ade quately for this specific purpose.

Ego Stregth scale and prediction of improvement in psycho therapy. Barron (1953) developed the ES scale to predict response to treatment with brief, psychoanalytically oriented psychotherapy. From the MMPI item pool he selected the 68 items most highly correlated with rated improvement of 33 neurotic outpatients. On cross-validation 59 samples, the scale correlated .42, .38, and .54 with ratings of improvement. But recent reviewers have questioned the adequacy of the scale for this purpose (Clayton & Graham, 1979; Clopton, 1979; Graham, 1977a,

1977b). Graham (1977b) concludes:

While the ES scale predicts response to psychotherapy when neurotic patients and individual, psychoanalytically oriented therapy are involved, it probably is not very useful for pre dicting responses to other kinds of treatment or for other kinds of patients. (pp. 35-36)

He apparently relies heavily on Barron's work as other research is inconsistent for all diagnostic types. There are some successes reported, but among inpatients. Dahlstrom, Welsh, and Dahlstrom (1975) report that ES has predicted clinical improvement of VA patients in one sample. Holmes (1967) reports modest correlations between the ES and speed of therapeutic response of 38 inpatients as rated by the treating psychiatrist. And Wirt (1955) reports significant but un- specified differences in ES between three groups of VA inpatients rated by therapists and supervisors as showing degrees of clinical movement. Other results are less clearcut. Distler, May, and Tuma

(1964) found, among male schizophrenics in their first inpatient treatment episode, that high ES scores related to greater improvement in symptom checklists but lower ratings by judges on overall health.

The relationships were reversed for females. Kidd (1968) reported a trend for female outpatients with ES scores above 30 to be rated in better condition at discharge. No such pattern was found for males.

The absence of expected correlations with improvement has been reported for students in personal counseling (Gallagher, 1954), for alcoholics 60

(Fowler, Teal, & Coyle, 1967), for chronic state hospital patients

(Hawkinsen, 1962), for VA inpatients (Levine & Cohen, 1962), and

for several samples of outpatients (Endicott & Endicott, 1964; Fiske,

cartwright, & Kirtner, 1964; Getter & Sundland, 1962).

For some samples, high ES indicates a poor prognosis. Gottesman

(1959) notes that delinquents are as consistently high on ES as they

are low in therapeutic change. Clayton and Graham (1979) indicate that

low scorers among a sample of 92 inpatients showed more improvement as measured by pre and post testing with the MMPI. Crumpton, Cantor, and

Batiste (1960) found that psychiatrist ratings of neurotic outpatients

correlated with ES scores to a significant degree, but inversely, ~

-.41. These latter results may be confounded by regression effects and patient defensiveness, respectively. But they illustrate the

unreliability of the Ego Strength scale as a predictor of improvement with therapy.

The ES scale is consistently poor in predicting perseverance

in therapy. Watson (1968) found it unrelated to hospital stay. It

fails to correlate with length of treatment for alcoholics (Fowler et al., 1967; Gertler, Raynes, & Harris, 1973; Trice & Roman, 1970).

And ES fails repeatedly in predicting outpatient stay (Getler &

Sundland, 1962; Lorr, McNair, Michaux, & Raskin, 1962; Rosenzweig &

Folman, 1973; Sullivan, Miller, & Smelser, 1958).

If Barron's measure fails to predict change in therapy, there is evidence that therapy leads to changes on the Ego Strength scale.

The ES scale increases as a patient's mental status clears (Dahlstrom 61 et al., 1975) and during the course of inpatient treatment (Lewinsohn,

1965), though not after electroconvulsive therapy (Dana, 1957). Higher

ES scores follow outpatient treatment for VA patients (Lorr et al.,

1962; McNair, Callahan, & Lorr, 1962), for obese women, rehabilitation patients, AA members, and narcotic addicts (Hollon & Mandell, 1979).

Barron & Leary (1955) found that eight months of treatment lead to comparable and significant increases for outpatients in individual and group therapy. Interestingly, similar increases were found for

27 untreated therapy candidates from the waiting list. It may be argued that ES changes mirror improvement in the clinical condition, whether or not therapy is the source of enhanced functioning.

To this point the review has cited studies which are, on the whole, critical of the performance of the ES scale in predicting out come of psychotherapy. Dahlstrom et al. remark that "as a predictor of response to relatively brief, psychoanalytically oriented psycho therapy, the ES scale has received little cross-validation beyond the original efforts of Barron" (1975, p. 169). They are technically accurate, if the scale is restricted to this very specific use. But the expansive tone of the original article, epitomized in the name given the scale, indicates a broader application. It fails empirically, in application to other populations and schools of therapy, as a predictor of outcome. Yet, the consistent relationship with therapy, evident in

ES changes over treatment, suggest the scale has valid meaning and satisfactory reliability (Barron, 1953; Gaines & Fretz, 1969; Silverman,

1963). 62

There are two basic and interacting features of the measure which likely account for its failure in predicting therapy outcome.

First, there is a dilemma posed by the definition of improvement.

Change is relative to a baseline. A person entering therapy with a high baseline level of functioning will have less room for improvement than a patient with more initial impairment. If the key element in predicting improvement is level of premorbid functioning, then a ceiling effect is likely. One conclusion of the critical literature is that the ES is more accurately construed as a measure of functioning and underlying personality resources than as a measure of change.

Second, the ES is a self-report scale subject to defensive distortion.

A person may achieve a high score because he denies problems. Yet, that same defensiveness will limit change in treatment. To anticipate the future argument, it is noteworthy that defensiveness need not imply greater pathology or, therefore, invalidate the measure as an index of ego strength.

Ends and Page (1957) correlated the ES of 63 hospitalized, male alcoholics with the change over treatment in MMPI clinical scales and a Q-sort measure. The expectation was ES would correlate significantly, as it did with decreases in the MAS, D, and Pt scales. It proved un related to the Q-sort measure. But, rather than accept the null hypo thesis, that the two measures are unrelated, the authors offer an alternative account: "If ES is measuring ego strength, one of course would not expect it to be closely related to personality change asso ciated with short term therapy" (p. 150). The irony of this statement 63 is that it interprets the use of the scale in a manner totally opposite to that offered by Barron in first introducing the scale as a predictor of success (presumably personality change) in brief treatment. This quote illustrates the lack of conceptual clarity which has complicated the interpretation of the ES. Clayton and Graham (1979) found that in patients low on ES showed greater improvement, as measured by change in average T-score of the MMPI clinical scales, than did those high on

ES. It would be plausible to conclude:

1) that low ES scores are the better therapy candidates, in

contrast to Barron's contention, or,

2) that the results are an artifactual consequence of regres

sion effects. Knowing that low ES is related to high T

scores on the clinical scales, one would expect more im

provement for the low ES group based simply on regression

to the mean. Or, conversely, if high ES patients show

relatively little elevation in their composite MMPI profile,

there is little room for change.

Future research should distinguish carefully between condition on dis charge, with which ES should be positively correlated, and change during treatment, with which ES may be inversely related.

A second reason often cited in explanation of inconsistent re sults is defensive test-taking attitudes on the part of some patients.

King and Schiller (1958) note that a maximum score requires 43 false to only 25 true responses. An acquiescent set leads to low scores, while a set to deny the symptoms listed leads to high ratings. For 64 one sample, King and Schiller report a higher correlation (.67) with the response set than withES scoring (.34). Block (1965) counters with evidence that an acquiescence controlled version of the scale is highly correlated with the original, and that acquiescence, therefore, is not an adequate explanation of the scale. This may be true in general, but certain defensive subjects are likely to distort their responses and present an artificially favorable picture of themselves.

Significant correlations are reported for ES with K (Gottesman, 1959;

Kleinmuntz, 1960; Von Evra & Rosenberg, 1963) and with the K-F ratio

(Distler et al., 1964). Crumpton et al. (1960) report that defensive patients achieve scores similar to normal students, with both groups higher on the ES than nondefensive patients. This interaction of defensiveness with ES leads Dahlstrom et al. to qualify the conventional interpretation of the prognostic utility of the scale:

If extremely high ES scorers in fact have standard MMPI profiles clear of any important elevation, yet are patients in a clinic or hospital, they are probably massively denying their personal or emotional difficulties. This extreme elevation on ES, then, is probably a poor prognostic sign for almost any psychological intervention. If, however, there is evidence from the rest of the record that they have some significant problem, and they also have scores on ES at or above the general mean, their dif ficulties may be of recent origin, their symptoms may be circum scribed, and the total interference in their life adjustment from these troubles is likely to be only moderate. For such cases, the prognosis for some form of psychotherapy is understandably favor able. (1975, pp. 169-170)

Along the same lines, Gottesman (1959) found that delinquents were significantly higher than a normal comparison group, and so suggested that ES be interpreted in conjunction with K and Pd. The Ego Strength scale, he concluded, "may be a sign of psychic energy per se, whether 65

it be thought of as coming from the ego or the id" (p. 345). Von Evra

and Rosenberg (1963) examined the ES scores of two groups of psycho

paths, 1ifferentiated into primary and neurotic subtypes according to

the MAS scores. Aa predicted, the nonanxious, primary psychopaths were significantly higher in ES (48.84 as opposed to 40.88) as well as

1 and K, reinforcing the interpretation of ES as a measure of defen

siveness and the ability to supply socially desirable answers. Primary

psychopaths were found to have lower scores on the neurotic triad and on an EPPS measure of conflict, which serves as a reminder that defen

siveness need not imply greater ego weakness. King and Schiller (1960) failed to find a significant association between ES and defensiveness among subjects whose driving licenses had been revoked. But those with higher ES scores were likely to use the relatively mature defense of rationalization and less reliant on the more primitive mechanisms of denial and projection. Among normal students, however, it was the lowest ES scorers that used rationalization, as part of a pattern of responding to a task with more defensiveness and less logical analysis.

The literature suggests that defensive subjects, at least among patient groups, tend to achieve higher ES scores. To the extent that defensiveness limits change in therapy, this further invalidates the use of the ES scale for its intended purpose of predicting progress in treatment. Yet, more refined analysis suggests that high ES scorers use more mature defenses and suffer less neurotic conflict. Intact de fenses enhance ego functioning. So the failure of the ES as a prog nostic instrument need not limit its utility as an index of available 66 personality resources.

Ego Strength scale versus ego strength concept. While exten sive use of the ES scale casts doubt on its utility in predicting response to therapy, an accumulation of data supports its validity as a measure of the somewhat nebulous concept, ego strength. The rela tionship was suggested initially by Barron (1953) on the basis of the content of the scale items. High scores result from endorsing items indicative of satisfactory physical functioning and stability; freedom from psychastheni~seclusiveness, phobias, various immature anxieties, religious dogmatism, and excessive moral rigidity; a secure sense of reality; and confidence in personal adequacy and ability to cope. To

Barron, this represents a "capacity for personal integration, or ego strength" (p. 329). Crumpton et al. (1960) obtained 14 factors from a sample of inpatients and medical students, prominently including absense of symptoms, religious dogmatism, psychopathic tendencies, and lack of anxiety and obsessions. A cluster analysis reported by

Stein and Chu (1967) substantially confirms these accounts. Three interrelated clusters tap physical, cognitive, and emotional aspects of well being. They involve freedom from symptoms of physical distress, intruding primary process thoughts, and anxiety and depression. Addi tional clusters include items tapping a "religious attitude of nonbelief and nonparticipation" (p. 155)--or denial of items with religious con tent--and "seeking heterosexual stimulation and escape from boredom"

(p. 155). Comparisons among various normal and clinical groups found lower cluster scores among the more pathological with the exception of 67

the religion items, on which normals obtained lower scaled scores than

either anxious or schizophrenic subjects. The failure of the religion

factor is in keeping with Crumpton, et al.'s (1960) finding that medical

students were more religiously dogmatic than patients. This subscale

of the ES detracts from the overall validity of the scale, which other-

wise has adequate construct validity.

Further attempts to validate the scale have included correla-

tional studies in which the ES is related to test scores and behavioral

indices of effective functioning and empirical comparisons between

groups presumably differing in ego strength. Results have been dis-

sappointing only with regard to the relationship of ES and ego strength measures obtained from projective tests. Otherwise there has been consistent support drawn from sufficiently diverse criteria and a large enough sampling of populations that the ES scale can be considered a valid index of ego strength.

Crumpton et al. (1960) summarize published results prior to

1960 when they write that:

Correlates of high ES scores are such variables as low MMPI clinical symptom scales, intelligence, and lack of ethnic pre judice; ratings of poise, drive, etc.; Q-sort items descriptive of personality functioning; independence of judgment in a group situational test {Asch-type]; ability to orient oneself correctly to the vertical plane in darkness [Witkin's field independencel; favorable aspects of functioning such as W and M on the Rorschach; denial of need for help; improvement in hospitalized patients. (pp. 283-384)

A diverse network of interrelationships link ES to effective ego functioning. It is consistently related to intelligence. Although Tamkin

(1957) found a nonsignificant relationship for one small sample, most 68

reports show correlations with the Wechsler scales ranging from .32

to .44 (Adams & Cooper, 1962; Barron, 1953; Tamkin & Klett, 1957;

Williams & Lawrence, 1954; Wirt, 1956). Other significant correlates include the Otis (Fowler et al., 1967), the Shipley-Hartford (Roos,

1962) and a measure of conceptual ability (Grace, 1960). Correlations with the CPI index of intellectual efficiency are even higher, ranging from .47 to .52 (Adams & Cooper, 1962; Barron, 1953). Not surprisingly,

ES is also related to education level (Clopton, 1979; Fowler et al.,

1967; Tamkin & Klett, 1957) and to social status (Fowler et al., 1967).

Modest and inverse correlations are typically obtained for age

(Getter & Sundland, 1962; Roos, 1962; Tamkin & Klett, 1957), although

Graham (1977b) indicates some inconsistencies have been reported.

Barron (1953) found that ES and ethnocentrism correlated -.47 for pa tients, -.46 for graduate students, and -.23 for air force officers.

Not only is it related to tolerance of others, but also toward oneself.

David (1968) factor analyzed patient's self and ideal-self descrip tions and found ES loading on a factor with lenient self-evaluation, large self-ideal discrepencies, an ideal distinct from that of the parents, and low psychopathic tendencies. The composite suggests a more individuated and self-accepting person. Similarly, Fiske et al.

(1964) found that ES loaded highly on a self-evaluation factor which included Q-sort adjustment.

The validity of ES as an ego strength measure is supported not only by its relationship to self-acceptance, but by its associa tion with constructive response to stress. Grace (1960), using the 69

McKinney Test, found inpatients high in ES more able to tolerate

stress and to recover from it. Dahlstrom et al. (1975) reviewed

studies using a psychogalvanometer which show that high ES subjects

are less responsive at rest and more responsive during stress. The

interpretation made was that high ES scorers spend less energy on

homeostasis and so have more reserves available in crisis. This

research is particularly compelling because it is not subject to the

same distortions attending self-report measures.

Studies correlating the ES with other derived and clinical

scales from the MMPI routinely show less anxiety and symptoms among

high scorers. Distler et al. (1964) obtained a correlation of -.71

with MAS for schizophrenic patients, while Ends and Page (1957) found

ES inversely related to pre and post hospitalization MAS for alcoholics.

Factor analyzing responses to a mental status examination from female

schizophrenics, Spitzer, Fleiss, Endicott, and Cohen (1967) obtained

a main anxiety factor, to which ES correlated -.52. Similarly, Levine

and Cohen (1962) found ES associated with absense of anxiety and psycho

tic symptoms for a sample of VA schizophrenics. The Ego Strength scale

correlates -.66 with the MMPI critical items (Tamkin, 1957; Tamkin &

Klett, 1957) and is similarly related to the other clinical scales

(Barron, 1953; Dahlstrom et al., 1975). While the degree of associa

tion may be exaggerated by the inclusion of overlapping items, it is

clearly considerable. Barron (1953) concludes that the ES "is picking

up a general factor of psychopathology in the MMPI" (p. 330). It

also correlates .47 with Meeker's LH-4 Scale, an independent MMPI- 70 derived ego strength scale (Greenfield, Roessler, & Crosley, 1959).

Factor analyses of MMPI scales typically generate a main factor often labelled ego strength. The ES scale is usually found to have a high and negative loading, again showing the denial of symptoms

(Adams & Cooper, 1962; Corotto & Cornut, 1962; Kassebaum, Couch, &

Slater, 1959). Lewinsohn (1965) found a similar structure when examining residual change scores (discharge less predictions from admission) of treated psychiatric inpatients. The ES correlated -.84 with the main anxiety/ego weakness factor. Williams and Lawrence

(1954) also found ES highly loaded on their ego strength factor which included verbal IQ and the Rorschach W, CF, and K scores. The Ego

Strength scale is clearly a good index of degree of pathology as measured by the MMPI.

Despite the naming of the above-mentioned factors, it would be premature to call ES a measure of ego strength inasmuch as it shares only the most minimal common variance with similarly intended scales derived from other instruments. Intercorrelations include:

.09 with the Rorschach Genetic Level (Levine & Cohen, 1962); .08 with the Goldberg Scale (Kidd, 1968); .02 and -.12, respectively, with the Bender Gestalt Z-score (Corotto & Cornutt, 1962; Roos, 1962);

.12 with Cartwright's revision of the Rorschach Prognostic Rating Scale

(Adams & Cooper, 1962); and .13, .22, .12, and .13 for different samples using Klopfer's Rorschach Prognostic Rating Scale (Adams &

Cooper, 1962; Endicott & Endicott, 1964; Herron, Gvedo, & Kanter,

1965). After finding only 2 of 36 significant intercorrelations 71

among different ego strength measures, Herron et al. (1965) thorough

ly understated the situation when they concluded "there seems to be

little evidence for the construct of objectively measurable ego

strength" (p. 404). The low correlations may be attributed to dif

ferences between verbal-cognitive and visual-motor tests (Corotto

& Cornutt, 1962), but it is clear that ego strength is not a unitary

concept. In the absense of convergent validation on this point, the

ES must be evaluated in terms of its correlation with other measures.

The literature surveyed has been remarkably consistent in showing high ES scorers to have more cognitive resources, greater self acceptance, greater resilience, and relative freedom from anxiety and other pathological symptoms.

Further validation is available when comparisons are made between high and low ES scorers on a variety of dimensions. Most globally, Hunter and Goodstein (1967) found that ES scores matched judge's ratings of ego strength as rated on the basis of defensive ness expressed in transcripts obtained from undergraduates explaining their poor performance on a symbolic reasoning task. In his original validation study on 40 graduate students, Barron (1953) found that high scorers were rated as functioning better, directing their energy more effectively, having broader cultural backgrounds, being more physically adequate, and more at ease socially. For students above, but not for those below, 21, Crites (1960) reported more developed career interests among those higher in ES, even when intelligence dif ferences were partialed out. Teter and Dana (1964) found no difference 72 in adjustment, as questionably operationalized as persistence at an unsolvable maze. Yet, Kormun (1960) found psychiatric inpatients with high ES scores outperformed lower scorers in a sensory discrimination task. Students self-referred for therapy were found to have higher

ES scores than those referred by others (Himelstein, 1964). But student counselees were lower in ES than those not requesting coun seling (Reschke, 1967). Welkowitz (1960) used the Bales rating system to evaluate the behavior of neurotic and psychotic patients in group therapy. Those higher in ES showed a more positive attitude toward therapy, were more sociable, and interacted more during the session. Jones (1969) studied the ES profiles of physically handi capped adults in a rehabilitation program. Those who were unsuccess ful were equally represented at all ES levels. But, for those who improved, ES was related to degree of success. Another interesting nonpsychiatric application is reported by Greenfield and his asso ciates (1959). Reasoning that ego strength should be expressed in the speed of recovery from illness, he compared ES scores of 18 under graduates with rapid recuperation and those of 20 students who took longer than average to recover from mononeucleosis. The quick recovery group was, as expected, significantly higher in ES. Though the difference in means was small, 50.39 versus 48.00, it was felt to be meaningful in an otherwise homogeneous group. To summarize this disparate group of studies, it is again apparent that hypothesized relationships between ES and ego strength are generally supported des pite a variety of perspectives, multiple methods, and diverse criteria. 73

consistent results in the face of such varied applications is power

ful support for the general application of ES as an index of ego re-

sources.

The greatest research effort assessing the adequacy of ES as a

measure of ego strength has been directed at the comparison of groups

varying in degree of psychopathology. The general conclusion of re

viewers, supported here, is that the scale functions well in making

gross distinctions--as between normal and psychopathological groups-

but is inconsistent with more refined discriminations (Dahlstrom et

al., 1975; Graham, 1977b). In comparisons between normals and patients,

only one of nine studies failed to report significant differences.

Winters and Stortroen (1963) examined ES scores from 25 normals, 25

general medical patients, and 25 schizophrenics. The top third of

the distribution of ES scorers was found to contain 47% schizophrenics,

when only 33% would have been expected by chance and even less were

hypothesized. While unexpected, these results may be demonstrating a

ceiling effect. The lower part of the ES range may be more discrim

inating. Normals were found to be significantly higher than a

variety of psychopathological groups in other studies (Crumpton et

al., 1960; Gottesman, 1959; Himelstein, 1964; Quay, 1955; Silverman,

1963; Spiegel, 1969; Stein & Chu, 1967; Taft, 1957). Among these,

Silverman (1963) repeated the comparison between general medical

patients (in this case, those with tuberculosis) and schizophrenics.

He found in two separate administrations, once with ES embedded in

the full MMPI and once alone, that the normals achieved significantly 74

higher ES scores. Spiegel (1969), for males and females separately,

not only found significant differences between normal and pathological

individuals, but subdivided each population into subgroups according

to rated health and found consistent and progressively increasing

ES scores as healthiness increased. Similarly, Gottesman (1959),

found educationally and occupationally superior adults significantly

higher than normals, who were, in turn, higher than most disturbed

samples. The ES, then, can dichotomize normal from abnormal samples, and may be able to differentiate those with exceptional resources.

Barron's scale is less effective, although not totally un satisfactory, for discrin1inating between diagnostic groups. Gottesman's

(1959) study showed mixed results. Delinquents were higher than emotionally disturbed adolescents, indeed, they were higher than normal adults. But no significant differences were found between psychotics and two groups of neurotics, those with peak MMPI scores on Hs and Hy or D and Pt. In his research, Tamkin reports no dif ferences between mixed neurotic groups and psychotics (Tamkin & Klett,

1957) or neurotics and character disorders (Tamkin, 1957). Hawkinson

(1962) was unable to discriminate neurotic~ from manic-depressives, from schizophrenics with the ES. He questioned the utility of the scale noting that self-esteem, sense of reality, and ability to delay grati fication are separate ego functions with differing implications for each of the diagnostic groups sampled. Others have extended this argu ment, which is the most compelling against the utility of the ES. It is not a unitary scale. Crumpton et al. (1960) extracted 14 factors, 75 of which 13 significantly discriminated medical students from VA in

patients. But patients obtained the higher score on four of these.

Significant differences on one subscale were offset by opposing dif ferences on another. Stein and Chu (1967) found that only the core subscale tapping sense of well being satisfactorily discriminated normals from neurotics, who, in turn, had higher scores than schizo phrenics. But within this factor, the cluster of items relating to physical symptoms failed to differentiate. And on the separate religiosity factor, schizophrenics achieved higher ES subscale scores than either normals or anxiety neurotics. There is an apparent double bind in using the ES: it must contain separate subscales to have ade quate construct validity as a measure of such a global construct as ego strength. Yet, this internal complexity tends to obscure dif ferences between diagnostic groups.

There have been successes, even when fine discriminations have been made. As previously discussed, Von Evra and Rosenberg (1963) found that primary psychopaths have higher ES scores than neurotic psychopaths. Roos (1962) reported groups with character disorders or neurotic diagnoses exceeded schizophrenics. Rosen (1963) com pared five internally homogeneous diagnostic groups. While some individual ~-tests were nonsignificant, the ES scores did success fully distinguish neurotics from psychotics. More specifically, neurotics with physical symptoms were significantly higher than neurotics with anxiety or depressive features, who were higher than paranoid schizophrenics. In his discussion. Roos criticized Tamkin 76

and Klett (1957) for attributing nonsignificant differences to the ES when the problem may well have been in failing to create homogeneous

comparison groups. In support of such an account, Figetakis (1964) was able to demonstrate that reactive schizophrenics achieve higher

ES scores than do process schizophrenics.

Graham's (1977b) review concludes that the ES scale generally works as a measure of adjustment. Valid discriminations can usually be made between normals and psychiatric patients, and between neurot

ics and psychotics. This review concurs with two specific qualifica

tions. As previously discussed, the scale is multidimensional. So the general accuracy of the scale in measuring adjustment will be dis rupted when one of the dimensions is particularly pertinent to a specific comparison group, as with religiosity among schizophrenics

(Stein & Chu, 1967) or students from a fundamentalist medical school

(Crumpton et al., 1960). Second, as a self-report instrument, it will underestimate the pathology of those who deny or ignore their pathology.

People with ego syntonic disorders will score higher on ES than those whose symptoms are ego-alien, as is demonstrated in the case of socio pathic types (Gottesman, 1959; Roos, 1962; Von Evra & Rosenberg, 1963).

Apart from these limitations, the ES is often capable of fine dis criminations among diagnostic groups and certainly adequate at the molar level as an index of maladjustment.

The ES scale, in summary, has been validated with a wide range of designs and measurement techniques, for a wide cross-section of the population. It has been shown to be a multidimensional index of 77

ego strength related to cognitive and emotional resources, to self acceptance, and to freedom from symptoms. While it is unreliable as

a vehicle for predicting change with psychotherapy, it is an adequate measure of the general degree of adjustment demonstrated by the indi vidual at a given point in time.

Ego Strength scale and objective self-awareness. One thrust of this research project will be to assess the validity of the Self

Consciousness scale, and the whole concept of the handling of reflec tive self-awareness, among clinical populations. The ES provides a means of relating SC to the level of psychopathology. The ES will be presented as an independent 68-item scale. While most of the research presented abstracts ES from the total MMPI, evidence suggests the separately administered scale performs similarly (Gravits, 1970).

While it has been previously argued that different diagnostic groups will present with discrepant patterns and overall levels of SC, no major common variance is expected when the ES and SC are correlated.

Individual subscales may well be linked. Private self-consciousness, because of the cognitive abilities required for the reflective process, may be positively related to ES. Social anxiety, in contrast, would be expected to peak among the less functional, and so to vary inversely with ES. But the validity of the SC with a clinical population will be optimized if the common variance with ES is limited. To the extent that the scales are independent, the SC should function for patient groups in a manner comparable to that found with college students.

Should the scales prove to be correlated, the different pattern in 78

the use of the SC by disturbed groups may undermine the hypothesized

link between Self-Consciousness and behavior in therapy to be dis

cussed in the next section.

Premature Termination From Psychotherapy

Premature termination from psychotherapy has been the subject

of a great deal of study over the past three decades. There are two main reasons for this interest. First, it was quickly recognized that a high percent of patients leave treatment early on and without con currence from their therapist. Brandt (1965) indicates that up to 35% fail to begin agreed upon treatment. Baekeland and Lundwall (1975) estimate 20 to 57% leave after one session, and that a minimum of 31% have dropped out by the fourth session. Malzer (1980) suggests 65% ultimately end treatment unilaterally. (Table 8 summarizes the per cent of termination and the criteria used for the populations sampled in this review). Premature termination is a major problem in the alloca tion of limited professional resources. Second, length of treatment is typically assumed to be related to improvement. As a conveniently available and quantitative index, it is often used as an outcome measure in therapy research.

Methodological issues. The generally shared assumption that length of therapy is an index of successful outcome (Lorr, McNair,

Michaux, & Raskin, 1962; Rosenthal & Frank, 1958; Saltzmen et al.,

1975; Strickland & Crowne, 1962; Tolman & Meyer, 1957) has received at least some support. Table 8

Pe;:cent of Patients Terminating Prematurely in Stt!dies Reviewed

Percent Reference Population or center type Comparison group cri.teri.:t N terminating

Affle<:k & Hednick (1959) VA, non-OBS 5- vs. 20+ sessions 50 Auld & Eron (1953) CHile, mixed 9- vs. 10+ sessions 33 36 Brown & Kosterlit:;: (19611) Uuiversity cl !nil',, mixed 4- vs. 5+ sessions 76 59 Carncena (1965) University cmmseling center, mixed 5- vs. 6+ sessions 60 35 Cartwr:!.t:ht ( 1955) University counseling center, mixed 5- vs. 6+ sessions 78 20 Conr acl (19 Sl1) VA, mixed 6- vs. 7+ sessions 100 4 7 Dcl.oac:l (l '177) University clinie, neurotic students 1- vs. 1+ month 62 D<'dd (l 970) Hospltal clinic, mixed 3- vs. 4+ sessions 169 4 7 Fi,.,;~<>r et "'. (1974} CMIIC, mix::.d 2-, 2+ but not ~m, ~IB 391 J6 (2-) 35 (2+) r'ralll<. et ai.. (i.957) llni•;ersity clinic, nonpsychotic 3- vs. 4+ sessions 91 30 FreeJman "t al. (!958) n·mc, schizophrenic 9- vs. 10+ sessions 54 '•6 GatfidJ et aJ.. (1963) 0•Jtpatient clinic, nonpsychotic 6- vs. 7+ sessions 24 50 Garfield & Kn~z (1952) VA, mixed 5- vs. 6+ sessions 560 '• 3 f!0rtlcr et al. (1973) Outpatient clinic. alcohol!~ 12- vs. 12+ mont.hs 97 79 Gihbey et al. (1953) VA, e1ixed 0, 1 to 5 sessions, q9 6+ months Gibbey et al. (1954) VA, mixed 5-, 6 to 19, 20+ sessions 269 40 (5-) Gandlach & Gelle::- (1958) An:tlyt:h: training center, mixed 5- vs. 6+ sessions 6 Hei!hrun (1961) University counseli.ng center, mixed 5- vs. 6+ sessions 73 55 !leine & T>:osman (1960) Outpatient clink, mixed 5- vs. 6+ sessions 46 48 H!Jer (1958}, UJler (1959) VA, mb:cd 5- vs. 20+ sessions 216 Horenstein (J 975), llormtstein Unlversity cllnic, mixed students 3- vs. 4+ orientation J r,t, 1! & Houston 0976) sessions !lorton [, Kri.mtcf.unas (1970) Chl.ld 1\Uidllnre cl 'nf.c, mixed MR- vs. HI\ 60 adolescents Imber et al. (1955) lhi.verslty c! i.nic, nonpsychot i.e 4- vs, 5+ sessions 60 )3 Imber ct al. (l956) llnivcn~ity clinic, nonpsyrhotic 3- vs. 4+ sessions 57 37 Katz & Solomon (1958) !~spital clinic, mixed 5- vs. 6+ sessiuns 351 '•'• l~olkov (195R) VA, neurotjc 7- vs. 8+ sessions 52 Kotkov & "c~dow (1953) VA, neurotic 3- vs. 9+ sessions 52

Note. Use of "-" indicates less than or eq,.utl to the crltP.rion indicated. ThE" or.ly C'xception is for "Nil-", l,h!t~h t!esiP,twte'l l0ss lhnn JI"lxb'"'" benifit. Use of "+'' represents p,reater th3n or equal to the crttPrion. Unreported d~ta is represented with a dash. Table 8 (Continued)

Percent Reference Population or center type Comparison group criteria N terminating

Lorr at al (1958} VA, m:l.xed 7- sessions, 26+ weeks 230 torr et al (1962) VI\, ncn-OBS 133 McNair et al. (1963) VA, non-em; 15- vs. 16+ sessions 282 38 Mc·ndelr.::.hn (1%6) University coun3elin::; center, student 111 Hensh & Golden (!951) VA, !l'ixed 4- vs. 5+ sesd.ons 638 65 r;ash et :tl (1957} University clinic, non;lsychot:!.c J- vs. 13+ sessions 48 14 2 'N'?t·aJ l f., Aron,cn (1963) llonpital cU.nic,, nonpsychoti.c 1- vs. 2+ sessiom; t,o '•3 R:1porort ( 1976) VA, mixed 0, 1 to 6, 7+ session~ 90 RogC'rS ct :tL (l'iSi) VA, mixed t, ·- vs. 5+ ~essi0ns IO

00 0 81

Cartwright (1955),in her examination of length of stay and outcome for 75 cases seen at the University of Chicago Counseling

Center, found a generally positive and linear relationship that was interrupted between the 13th and 21st sessions. Cases terminating in this "failure zone", were rated no more improved than those dropping out after three sessions. Taylor (1956) replicated this pattern of a disrupted linear relationship, with outcomes rated poorer during the teens. A plausible interpretation is that this is the stage in therapy at which relationship factors become paramount. Dropouts leave therapy unimproved because they are unable to tolerate a depen dent relationship with the therapist (Fulkerson & Barry, 1961).

Cartwright (1955) suggests that two different functions, both essen tially linear, may be superimposed in her study. Short-term therapy, with a goal of symptom relief, may achieve maximum benefit and end by the 12th session. Long term therapy, with its focus on personality change, begins to show results, once the relationship is developed, after 20 sessions. The implications are two-fold. First, number of sessions are related to therapy outcome, justifying an underlying pre mise of dropout research. But second, the relationship is not simple and so the criterion number of sessions should be selected carefully.

Either the criterion number of interviews should be kept low (about five), or the terminator should be compared with long term remainers having more than 20 sessions. A split comparing patients above and below, say, nine sessions runs the risk of comparing short term and long term patients and failing to make the assumed comparison between 82 terminators and remainers.

Baekeland and Lundwall (1975) add that:

patients with acute situational problems will derive little benefit from extended treatment and may resist it by dropping out after they have gotten from it what they wanted in the first place • • • • Having realized their goals, which are not the same as those of their therapist, they may nonetheless get considerable benefit from treatment. (p. 744)

Meltzoff and Komreich (1970) concur. Fiester and Rudestam (1975) found four factors characterizing treatment dropouts from a state hospital clinic (none replicable for a community mental health center cross- validation sample). Two of the four factors (patient seeking and getting direct advice, and achieving some problem resolution with a close and well-adjusted therapist) suggest the patient achieved desired objectives before termination. Rosenthal and Frank (1958) reported

32.5% of their improved patients had only five sessions. To set a higher number of sessions as criteria increases the risk that people who accomplished their treatment objectives, and so should be thought of as remainers, will be mistakenly considered dropouts. For these reasons, the present research will use five sessions as its criteria.

Those having five or less sessions will be designated as terminators, unless the chart indicates a mutually agreed upon ending or a referral on discharge, while those continuing for six or more sessions will be defined to be remainers.

Despite the volume of research, there is a considerable lack of clarity about the determinants of premature termination. A considera- tiou of the possible domain of relevant variables would minimally 83

include patient traits, therapist dispositions and techniques, state variables resulting from their interaction, clinic policy and setting, and the family and cultural context in which treatment takes place.

Research began with client characteristics and has paid less attention to each successive level. As a result, there is inadequate data of clinic and contextual features.

The lack of clarity also results from methodological complex ities peculiar to the research area. In the following paragraphs, some of the most important will be briefly reviewed. Among them:

1) lack of agreement in the definition of premature termina

tion leads to differing and contradictory criteria of

dropping out,

2) even when premature termination is defined in terms of a

set number of sessions, the number chosen to delimit the

terminator group varies widely,

3) there is concern that aggregate data may obscure different

profiles if, as is likely, there is more than one type of

dropout, and

4) there is some disagreement whether or not those who fail

to begin treatment are sufficiently similar to those who

begin therapy and then quit unilaterally that the two

groups can be combined for comparison with those who

remain in treatment.

These issues will be considered before this review proceeds to consid er the results of the research literature. 84

There have been three approaches to the operationalization of premature termination. Some have suggested that time in therapy is more relevant to outcome than is the number of sessions (Lorr et al.,

1962). Comparison groups are constructed based on duration of treat ment. There are obvious practical problems with such an approach, such as how to handle the patient who delays therapy because of a vacation or illness. A second approach, equally quantifiable, uses a set number of sessions as the criteria for remaining in treatment.

"Since the number of visits directly measures the patient's exposure to treatment, it is to be preferred to time in treatment as a measure of dropping out" (Baekeland & Lundwall, 1975, p. 741). Perhaps the conceptually purest measure is one which defines premature termina tion as any unilateral move by the patient to end treatment (Fiester,

Mahrer, Giambra, & Ormiston, 1974). With this definition, a patient could still be considered a terminator after hundreds of hours of treatment. It is questionable whether this individual should be defined as belonging in the same group as the patient quitting after a single interview. Since, as noted above, unilateral termination may express either dissatisfaction with therapy or the attainment of desired goals, Fiester's approach is misguided. Moreover, the quality of the archives used in most of these studies is inadequate to reliably differentiate unilateral from mutual decisions to end treatment. This study will use the second approach, setting six sessions as the mini mal criteria for being considered a "continuer". This means that the results will not be directly comparable to studies using either alter- 85 native approach to operationalizing premature termination.

Although most studies use the number of sessions for the criterion, there is still considerable disagreement about where to set the cutting point. As indicated in Table 8, McNair, Lorr, and

Callahan (1963) called all of those terminating before the 16th session dropouts, even though the typical course of treatment is only eight sessions in the majority of clinics studied (Garfield, 1978).

Again, a much lower number of sessions is used to minimize the inclu sion of patients leaving with major treatment goals satisfied.

Inconsistent results can be expected since the evidence sug gests that premature termination is a complex and overdetermined phenomenon. Aggregate data may obscure as well as elucidate contri buting factors if too many distinct subgroups are pooled in composing the terminator sample. Yalom (1966) is the most clear in presenting this argument. Citing nine different factors leading to defection from group therapy, he stressed that few would have proven significant in any analysis of aggregate data. A variable might be applicable to only one or two respondents, yet its importance could be so paramount in these cases that it is the determining element. Evidence of such subgroup differences appears repeatedly. Saltzman et al. (1976) sampled patient and therapist reactions after the first, third, and fifth sessions. In analyzing the results from subsequent dropouts, termination after one session was attributed to low anxiety while dissatisfaction with the relationship was evident for those leaving later. So there is some indication that there are subgroups based on 86 length of stay even when the criterion number of sessions is very limited. Heilburn (1961) compared terminators and remainers on eight subscales of the Gough Adjective Checklist. For six of them there were significant interactions by sex indicating that different criteria are needed for males and females. But later research on the Counseling

Readiness subscale (Heilbrun, 1962) failed to replicate sex differences.

A more consistent observation is that dropouts can be discrim inated into subgroups according to whether they later seek help or not. Tolman and Meyer (1957) examined records of 354 VA patients and found that only 6% of those terminating within four sessions return, while 12% of the continuers have subsequent contact with the Mental

Hygiene Clinic. No information was provided on use of other mental health resources. Chameides and Yamanoto (1973) observed that 78% of their failed referrals got help elsewhere, as did 63% of the sample studied by Brandt (1964). Fiester et al. (1974) found that their sample of community mental health center patients who remained until mutually agreed upon termination were more likely than dropouts to have had previous clinic contact. Some portion of the terminators return for successful treatment later while others reject mental health in terventions entirely. Although no differences are reported, presumably these two groups are distinct. The usual practice of combining them is likely to obscure distinctive patterns characterizing the subtypes.

However, in the absense of specified and empirically validated pre dictive variables, it is impossible to separate subgroups. Aggregate data must continue to be used while clinical impressions and followup 87 research attempt to clarify the means of discriminating distinct subsamples within the larger population of terminators. At the same time, retrospective interviews with dropouts are suggested as useful in attempts to more fully appreciate the multidimensional deter minants of premature termination from psychotherapy.

There has been some discussion of differences between those who dropout of treatment after one or more sessions and those who refuse treatment or fail before it begins. Here the distinction is easily objectifiable: no treatment versus one or more sessions.

Several studies compare these groups. Garfield and Kurz (1952) report that for a VA sample of 768 men found suitable for treatment,

27% refused treatment (compared to 31% dropping out through the fifth session). Garfield (1978) described rejectors as having low income and socioeconomic status, vague problems, and poor motivation.

Such descriptions are typical for dropouts as well, as will be seen below. Gibby, Stotsky, Miller, and Hiler (1953) found rejectors and terminators similar in their constricted Rorschach records and somatic complaints, features that distinguish them from remainers. Williams and Pollack (1964) found no demographic distinction between families of refusers and terminators at a child guidance clinic. Brandt's (1965) review suggests that differences that have been reported are not con sistent, have not been replicated, and are not clinically useful for prediction. His conclusion, consonant with the evidence from this re view, is that reported differences are sample specific and artifactual.

However, the sample used in this study will be examined on a post hoc 88 basis, on the assumption that anxiety related to objective self-aware ness will be aroused more among those who begin treatment.

One final source of inconsistency, and one most typically encountered in literature reviews, is the reality that studies examine different populations armed with a wide variety of measurement devices.

Even a strong and consistent effect would likely be missed as often as found. Conflicting results are seen as evidence of multiple perspec tives being taken on a complex phenomenon. Fortunately, enough con sistency emerges that crucial determinants can be identified.

Psychological tests. The first research used psychological tests to measure patient characteristics related to premature termina tion. The results have been generally discouraging. Initial suc cesses have failed on replication or have been reinterpretted as evi dence of the impact of confounding variables. This pattern holds for the course of research on the Rorschach, the MMPI, and the specially designed Terminator-Remainer Battery. The studies to be reviewed below indicate that standard psychological tests have little to contri bute to the prediction of psychotherapy dropouts.

Research with the Rorschach began inauspiciously when Rogers,

Knaus, and Hammond (1951) found that none of 99 selected formal ele ments significantly discriminated VA patients terminating before the fifth session. Taulbee (1958) found significant differences for R, C,

V, An, A%, and number of rejections. But his use of 13 interviews as the criterion for remainers makes his results suspect, for the reasons described above. When Kotkov and Meadow (1953) attempted to replicate 89 a formula based on R, FC-CF, and D%, they found a nonsignificant trend for D% to be higher for continuers, which contradicted their earlier results with group therapy patients. Continuers offered significantly more R and FC-CF responses. Auld and Eron (1953) w2re able to replicate differences only for R. When they partialed out the effects of verbal IQ, the correlation of R with the number of sessions dropped from .55 to .07. Affleck and Mednick (1959) found significant differences in R, M, and marginally in H when 25 veterans dropping out in 4 or less sessions were compared with 25 continuing 20 or more interviews. The importance of IQ was challenged since re mainers, although higher, didn't differ significantly from terminators.

Gibby et al. (1953) characterized terminators as producing constricted records, with few movement and color responses, and reliance on easy shapes (F+%). The same researchers (Gibby, Stotsky, Hiler, & Miller,

1954) later developed a formula using R, K, and M which was 67% accu rate on cross validation. However, R alone correctly predicted 69%.

With IQ partialed out, its correlation with continuation dropped from

.38 to .32. Much of this was probably related to social class stand ing, which itself predicted perseverance in psychotherapy with 63% accuracy.

This discussion leads to the same conclusion reported by other reviewers (Fulkerson & Barry, 1961; Garfield, 1978): R is the only

Rorschach variable that reliably discriminates between those perse vering in therapy from early terminators. Continuers are not only more intelligent, but also more ego-involved and cooperative. The 90 lower productivity of the terminator is equated with depression, pas sivity, and incapacitating anxiety (Hiler, 1958; Meltzoff & Kornreich,

1970). Interestingly, Hiler (1958) and his associates (Gibby et al.,

1954) found that some therapists are able to retain patients with low

R. So patient productivity interacts with therapist characteristics instead of being a main effect in some research. In all, the Rorschach is a rather cumbersome screening device for predicting continuation in therapy.

Results with the MMPI have been inconclusive. Taulbee (1958) found continuers (13 or more sessions) higher on D, Pa, Pt, and Sc.

Sullivan et al. (1958) dichotomized his veterans' sample at nine ses sions and found terminators higher on Pa and lower on K. Their first cross-validation sample showed remainers higher on Pt, but no dif ferences on other MMPI scales, including initially significant derived measures of social status, intellectual efficiency, ego strength, anxiety, and repression. All MMPI scales failed to discriminate on a second cross-validation sample. Dodd's (1970) study of outpatients at a university hospital clinic also failed to find MMPI differences in average T-score or number of elevated clinical scales. Horton and Kriavciunas (1970) found adolescent continuers compliant, as indicated by lower F, Pa, Sc, and Ma. And students continuing at a counseling center were typified by moderate Hy and Pd scores (DeLoach,

1977). Their Ego Strength scale scores were no different from termina tors, as was the case with the veterans studied by McNair, Callahan, and

Lorr (1962). Lastly, McAdoo and Roeske (1973) found no differences in 91

the MMPI profiles of continuers versus terminators among parents at

a child guidance center. In summary, neither clinical or derived

scales from the MMPI have yet proven effective predictors of perse verance in therapy.

Lorr and his associates (Lorr, Katz, & Rubenstein, 1958;

Rubenstein & Lorr, 1956) derived a Terminator-Remainer battery (TR)

from the most discriminating items of a self-ideal discrepancy rating

scale, the F-scale, Taylor's Manifest Anxiety scale, and a behavior

disturbance scale (Applezweig, Dibner, & Osbourne, 1958). The TR

Battery outperformed the parent scales and significantly discrim

inated dropouts from continuers, with predictive accuracy above 60%.

McNair et al. (1963) cross-validated the derived formula and was able to exceed base rate accuracy. But Garfield (1978) reports that

the multiple R obtained was only .44. The implications of McNair's study are further confused by considering all patients with 15 or less sessions to be terminators. Despite this, terminators were described as impulsive psychopaths or as authoritarians who deny anxiety or self-dissatisfaction. Stern, Moore, and Gross (1975) begin their critique by noting that such a description has often been applied to lower class clients. In a completely crossed design, with 17 nonpsychotic outpatient clinic clients per cell, middle and lower class terminators and remainers were compared. The TR battery did predict premature termination (with five or fewer sessions), but only across social class. Within class the TR accuracy was only

50%, identical to chance expectancy. The failure of the TR to 92

reappear in the literature since Stern's report suggests that it was

persuasively rejected by this study. The TR battery is like other

P'~rsonality tests in having failed to demonstrate patient personality

characteristics that are related to premature termination from psycho

therapy.

Social class. The variable that has most consistently pre

dicted therapy dropouts is social class. Whether measured by a compo

site index or component elements such as income, education, occupational

status, or housing type, length of stay in psychotherapy is shorter and

dropout percent higher for the lower class client (Auld & Myers, 1954;

Frank et al., 1957; Imber, Nash, & Stone, 1955; Imber, Frank, Gliedman,

Nash, & Stone, 1956; Karmin & Coughlin, 1963; Katz & Solomon, 1958;

McNair et al., 1963; Myers & Schaffer, 1954; Rosenthal & Frank, 1958;

Rubenstein & Lorr, 1956; Sandler, 1975; Schaffer & Myers, 1954; Stern et al., 1975; Weiss & Schaie, 1958; Winder & Hersko, 1955).

This finding is not universal, since some studies find no relationship (Brill & Storrow, 1960; Brown & Kosterlitz, 1964; Lorr et al., 1962; Strickland & Crowne, 1963; Taulbee, 1958) or mixed results across cross-validation samples (Dodd, 1970; Fiester et al.,

1974; Lorr et al., 1958; Sullivan et al., 1958). But the supportive evidence is great enough that lower class patients must be considered likely to terminate prematurely.

In part, this may be attributed to referral and selection pro cedures. Clinics routinely put patients through a "creaming" process so that the most desirable patients are retained. Middle class patients 93

are among the selected (Brill & Storrow, 1960; Brown & Kosterlitz, 1964;

Rosenthal & Frank, 1958). Those with a lower class background are more

likely to request help with somatic problems and be judged as unsuit

able for insight oriented therapy (Schaffer & Myers, 1954). When they

are retained on the client roles, lower class clients get group, cus

todial, or organic therapies instead of analytically-oriented psycho

therapy (Bailey, Warshaw, & Eichler, 1959; Kamin & Coughline, 1963;

Robinson, Redlich, & Myers, 1954; Winder & Hersko, 1955). They may

be assigned to less experienced and less competent therapists (Brill

& Storrow, 1960; Myers & Schaffer, 1954). So clinics may complicate

preexisting difficulties in retaining lower class clients by offering

them the least preferable services. But the preexisting difficulties

are exemplified by one clinic's finding that 48% of lower class, com

pared with 12% of middle class, drop out before therapy begins (Myers

& Schaffer, 1954). Selection practices reflect rather than create the

communication obstacles that prevent the retention of the socio

economically deprived.

Other variables, including the psychological test indices re viewed above, have been reinterpretted in terms of social class.

Auld and Myers (1954) argue that requisite therapy skills (psycholo gical mindedness), rewards (such as empathy and insight), and punish ments (fees) are all mediated by socio-economic status. Noting the

.403 correlation of class and number of therapy sessions, they suggest that what are called personality factors are really socio-demographic ones. But the syllogism can be reversed. Freeman (1961) found that 94 the relatives of patients with higher amounts of education used psycho- genic explanations of illness, responded favorably to hospital treatment, expected recovery, and were less likely to blame the patient. He suggested that social class is a screen for the more crucial education and intelligence differences that determine attitudes. Reviewers seem to concur. Garfield (1978) believes that "educational level may be only one component of a larger factor or complex of factors that may include verbal ability, sophistication about psychotherapy, income, interest in receiving psychotherapy, and similar components" (p. 198).

Meltzoff and Kornreich continue: "socioeconomic variables do not seem to be of great importance by themselves. What would seem to be significant are the psychological implications of the social position of the patient: his learned behaviors, roles, attitudes, expectancies, and traits" (1970, p. 362). Lower class patients modally are less inclined to monitor or express thoughts and feelings, to interpret anxiety of behavior disturbance psychologically, to accept prolonged treatment from a relatively passive therapist, or to find acceptance for his status as a psychotherapy patient. Note how this early comment by Myers and Schaffer (1954) anticipates the current interest in client/therapist similarity and expectancy match:

psychotherapy involves intimate communicative interaction between the patient and therapist. Therefore, it may be facili tated if a certain similarity in culturally determined symbols and learned drives exist in both patient and therapist. Dif ferences in value systems and patterns of communication on the other hand may hamper the establishment of the therapeutic rela tionship. At present, it appears possible that lower-class patients need to acquire new symbols and values to participate in expressive psychotherapy. Since this is a difficult process, 95

many of them may be considered unpromising candidates for suc cessful treatment. (p. 310)

Other patient characteristics. Therapist and expectancy variables will be discussed in detail below. But first to be considered are personal characteristics of the patient that have been linked to premature termination from psychotherapy. As a starting point, Figure 1 summarizes the findings of previous reviews. The constellation of personality characteristics related to social class typically includes lower intelligence, referral from social agency or courts, low TR scores, somatic or situational presenting problems, lack of psychological mindedness, and poor motivation for treatment

(items 2, 5, 9, 10, 11, 18, and 19, respectively). While the results are not entirely consistent, the bulk of the evidence suggest patients compelled to seek treatment drop out (Frank et al., 1957), while continuers are self-referred or coming from a psychiatric source

(Rosenthal & Frank, 1958; Straker, Davanloo, & Moll, 1967; Weiss &

Schaie, 1958). Four of the six studies reviewed concur that initial complaints that are somatic or nonpsychological are more frequent from patients who drop out early in treatment (Gibbey et al., 1953;

Hiler, 1959; Katz & Solomon, 1958; Kotkov, 1958). Complaints empha- sizing depression, anxiety or other neurotic features not only typified those who continue in treatment, but those with a middle class background. Both psychological mindedness and motivation for therapy are the types of global and abstract concepts that are seldom objectively measured but appear in summary statements, particularly after projec- Reviewers

Factor Fulkerson and Brandt (1965) Dodd (1970) Me] tzo ff and Baekeland and Garfield (1978) Type ll,rry (1961) Kornreich (1970) J,undwall (19 7 5)

Patient Characteristics: 1. Low Rorschach R 2. !...ow intclligencl! Low intellige-nce 3. Low suggestibility t,. Dtagnosis other Diagnosis other Diagnosis without niagnosis differ Low an.dety and D:f.agnosis differ than neurosin than neurosis anxiety ences denied depression ences denied

5. Lo~! socl~l clasH Low social class Low social class Low social class Low social class

6. No previous ther~py 7. Race other than caucasian 8. Sex differences Sex differences Female S•'Y. d l f f erl'nces denied denied denietl 9. Agenr.y referral 10. Low TR battery score I I. So mat lc prescr.t ing pcohlem 12. Paranoi.d signs t3. Sociopathy u•. l). Nomadic ln. Aggressive or passive-aggressive !7. Social isolation 18. Not psychologically minded

-··-----· ---·------

Fi)•,qrp I. RC'VI!'\~erG' r('sponHPS to variabl!'s at'lsoelate

19. Poor motivation 20. High social desirability

21. Age difference"' Age differences Younger Age difference~ denied deni<'d denied Therapist Characteristics: 22. Inexperience Inexperience Inexperience z3. Lack11 warmth 24. Rejects dependency Rejects dependency 25. Dislikes patient Rates patient's attractiveness low n. Male 2 7. Rates vrognosh; roor

[nt:er:-~ctl.nn o( !'nti!'nt and Therapist Characteristics: 28. Dissimilar Dissimilar backgro•.:nd background /9. Dissimilar llissjmi l a r. expectations e>xpect~tions

Cl inlr Charncteristlcs:

]0. Lon~ waiting lists

Note. Each f:u·tor liRted is specifically mentioned by one or Hore reviewers as n cnuse of pre1Mtu1:e terrnlnation. Also l !sted Hre f ac.tors rejected by eac.b review. In these· cases, "denied" is added to the 1:1 sting.

Flgure I. (Contlnued). 98 tive testing (Gibby et al., 1954; McNair et al., 1963). Frank et al.

(1957) rated motivation from patients' retrospective self-reports.

They found the variable significant only for comparisons with those who failed to come for even a single session. Once contact was initiated other factors assumed preeminence. Whether psychological mindedness or motivation adds anything beyond that explained by con cepts such as expectancy match remains to be proven.

Suggestibility (item 3) has been investigated by the Johns

Hopkins group. It has been measured using the sway test, in which the patient is blindfolded and given the repeated suggestion that he or she is falling backwards. When degree of motion is coded, the "swayer" can be identified with 67% accuracy as a continuer (more than three sessions) and the nonswayer is likely to drop out in a similar percent age of cases (Imber et al., 1956). A replication found marginally significant differences in the expected direction (Frank et al., 1957).

Presumably the treatment is enhanced by the suggestible subject's dependence on the therapist. Yet, judges' ratings of patient depen dence were not significantly related to perseverance (Garfield et al.,

1963) and therapist reaction to dependent verbalizations by the patient was found to be unrelated to continuing (Caracena, 1965). Baekeland and Lundwall (1975) indicate that extremes of dependence or counter dependence are bad signs, and counterdependence is a hallmark of dropouts from alcoholism treatment (Mozdzierz, Macchitelli, Conway,

& Krauss, 1973). Yet, the verdict is still out on general psychiatric outpatients. It is likely that the importance of suggestibility and 99 dependence are found in interaction with specific situations rather than on a trait level. Winder, Ahmad, Bandura, and Roo (1962) com pared therapists responses to dependent verbalizations (item 24) by patients leaving by the lOth session or remaining past the 20th.

While there were no significant differences in overall dependent behavior or its acceptance, almost 100% accurate discrimination was achieved for this small sample when therapist's response to patient's verbalization of dependence on the therapist was measured. Dis couraging responses led to premature termination in each case. More research is clearly needed, although existing data is promising.

Cartwright's (1955) explanation of the "failure zone" in terms of difficulties in negotiating a satisfactory relationship offers a con vergent theoretical perspective that could be used in research.

The relationship of diagnosis (item 4) to duration in therapy is controversial, as indicated by differences among the major reviews.

In part, this is a function of the way in which the term is used: the wider the scope, the more likely that some element will be shown to differentiate. In the literature reviewed, the weight of the nega tive evidence is heavy. No differences among diagnostic groupings were reported by Brown and Kosterlitz (1964), Fiester et al. (1974),

Katz and Solomon (1958), Lorr et al. (1958), or Mensh and Golden

(1951). Number of symptoms, (Lorr et al., 1958), degree of pathology

(Conrad, 1954), and severity and duration of illness (Katz & Solomon,

1958) were also unrelated. However, Frank et al. (1957) found pa tients stayed in treatment if they had been ill longer or if symptoms 100

fluctuated. These researchers also found those complaining of anxiety

or depression remained. Other experiments indicate that specific symp

toms correlate with premature termination: absense of anxiety

(Freedman, Englehardt, Hankoff, Glick, Kaye, Buckwald, & Stark, 1958;

Kotkov, 1958); absence of depression (Straker et al., 1967); or alcoholic, sociopathic, or paranoid features (Sherry, 1977; Straker et al., 1967). But conflicting results appear when individual symptoms are assessed. In contrast with Sherry's results, Freedman and his associates found less suspiciousness or hostility among the ambula tory schizophrenics leaving treatment early. Apparently they were functioning well enough that treatment was discretionary. Similarly,

Heilbrun's (1962) dropouts describe themselves as better adjusted then did continuers. People seek therapy for various reasons: because they want to change or because their pathology is so evident that they are pressured to seek help. Any of a variety of symptoms

(including items 12, 13, 14 & 16 from Figure 1) may lead a person to treatment. Because of this diversity, it is unlikely that diagnosis, severity of impairment, or presence of specific target symptoms will consistently discriminate terminators from remainers. Kotkov (1958) found neurotics and personality disorders overrepresented among con tinuers. Dodd (1970) found more psychotic reactions or neurotics.

To summarize these observations in the context of earlier reviews, little more specificity can be expected than that remainers are likely to be in more acute distress, and therefore, to be experiencing more problems with anxiety or depression than will terminators. 101

Previous psychotherapy (item 6) characterizes those who remain in therapy. Since Brandt (1965) noted this finding in three of five studies reviewed, similar results have been reported by Fiester et al. (1974) and Sherry (1977). In part, this may represent the develop ment of a chronic population who depend on clinics but never attain maximum benefit. More optimistically, initial therapy experiences may educate the consumer as to process of treatment so that later ther apy is begun with more realistic expectations.

Demographic variables have generally failed to differentiate dropouts from continuers. Race (item 7) is the most problematic.

Three studies reviewed (Fiester et al., 1974; Mensh & Golden, 1951;

Overall & Aronson, 1963) find no differences, while three others re port whites are likely to stay longer (Dodd, 1970; Lorr et al., 1958;

Rosenthal & Frank, 1958). Garfield (1978) argues that race differences are usually an indirect expression of social class. But two of the experiments finding differences evaluated socioeconomic status and found minimal effects. Further studies in which race is assessed with class controlled are needed to resolve the issue. Males are found to remain in treatment longer than women in three studies reviewed (Brown & Kosterlitz, 1964; Rosenthal & Frank, 1958; Weiss

& Schaie, 1958). Heilbrun (1973) suggests that traditionally femin ine clients will have difficulties with a male therapist's passive or detached role. Five other studies found no sex differences (item 8).

Nine experiments reported no age differences (item 21). When age proved important, Katz and Solomon (1958) found old terminators, Sherry 102

(1977) lost young patients, and Brown and Kosterlitz (1964) found disproportionate termination at both ends of the age spectrum. The inconsistency suggests an artifactual or sample specific basis for significant differences. Marital status discriminated in only two of eight studies. Weiss & Schaie (1958) found single patients to be con tinuers, while Katz and Solomon (1958) found both single and married clients remained longer than those with broken relationships. On balance, the evidence from this review leads to the same conclusions reached by Dodd (1970) and Garfield (1978) in denying the importance of age, sex, and marital status.

Dodd found differences based on religion in two of four studies. This review also found two samples with differences. Lorr et al. (1958) had relatively few Jewish terminators. Dodd (1970) had Protestant continuers, although this pattern failed on cross validation. A lack of significant effects attributable to religion characterized four other reports (Fiester et al., 1974; Katz &

Solomon, 1958; Rubenstein & Lorr, 1956; Weiss & Schaie, 1958). Demo graphic variables have generated conflicting results, but no clinically significant pattern has emerged and no theoretical basis has been developed to justify their continued inclusion. Most likely demo graphic variables are reported so frequently because they are easily quantifiable and readily available in archival sources.

A cluster of variables relating to the patients social skills has appeared in the literature. Baekeland and Lundwall (1975) listed both an unstable, nomadic lifestyle and social isolation (items 15 and 103

17) among the correlates of premature termination from therapy. Both characteristics indicate a lack of enduring interpersonal attachments that should be reflected in a shallow or unstable relationship in therapy. But experimental studies are contradictory. The Johns

Hopkins group found more social ineffectiveness among dropouts in one study (Nash et al., 1957) but no difference in another (Frank et al.,

1957). Dropouts in this latter study, however, presented a history of fewer group memberships or relationships. McNair and his associates

(1963) found that it was the continuers who were more retiring, as rated on the Gilford-Zimmerman Sociability Scale. But a previous study by the same group (Lorr et al., 1962) had failed to find dif ferences on the Sociability or Friendliness scales or on the Leary

Interpersonal Checklist. Freedman et al. (1958) found, in keeping with their expectation that dropouts were healthier, that these patients were more active and involved with others. Once again, the inconsis tency of the evidence makes it impossible to endorse Baekeland and

Lundwall's enumeration of patient characteristics discriminating drop outs from therapy continuers.

No other patient characteristic reviewed, including the per cent of disability compensation in studies of veterans (Lorr et al.,

1958; Mensh & Golden, 1951; Rubenstein & Lorr, 1956), was found to be significant in one study and replicated in another. Such results have led some researchers (Fiester & Rudestam, 1975; Frank et al.,

1957; Garfield et al., 1963) to comment on their surprise at the lack of potency of patient differences and to turn their scrutiny to state 104 variables or therapist character traits. However, this review has

confirmed the importance of some factors. The terminator is likely to be characterized by some combination of the following: low socio economic status, with associated impoverishment in intelligence, ver bal skills, cooperation, or productivity in therapy (Rorschach R); somatic concerns; low distress, with less anxiety or depression in evidence; less experience with psychotherapy; and difficulties entering a dependent relationship. One other abstract characteristic deserves consideration because it recurs in the work of several different re searchers. Usually this comment appears in the summary of discussion almost as an intuitive footnote. Those who persevere in therapy seem to bring all projects begun to closure. Rubenstein and Lorr (1956) talk about the combination of impulse-control and frustration tolerance which results in a certain stick-to-it-iveness. Garfield et al. (1963) call them achievers. Frank et al. (1957) find evidence of this per severance in continuing group activities, and in finishing a stage of education (at whatever level) rather than quitting midway. These authors felt this to be the only true personality trait represented among several related to the interaction of different features of the therapeutic situation, predicting continuation in therapy. It has not been measured in a standard or specific way. But perhaps there is some cogency in calling such patients the ones who persevere in treatment.

Therapist characteristics. In an interpersonal activity such as therapy, it is clear that characteristics of the therapist also 105 contribute significantly to the patient's experience. Confirmation of this notion is found in the long training of mental health pro

fessionals and in the frequent discussion in the applied literature of how the therapist uses his self to further the treatment. Some studies have examined the early dropout phenomenon with an eye to the

therapist's contribution, and, in a few cases, to the interaction or match of the patient and counselor. This review will begin with gen

eral/demographic characteristics, move on to more treatment-specific qualities, and end by discussing interacting features. Because it has generated such a volume of studies, expectancy match, a special case of interaction, will be considered in a separate section of this re view.

Despite Baekeland and Lundwall's (1975) contention to the con trary, there is no direct relationship between therapist sex and treatment dropouts (Lorr et al., 1958; McNair et al., 1963; Sullivan et al., 1958). Nor is there an interaction between sex of therapist and sex of patient (Mendelsohn, 1966; Saltzman et al., 1976). Hiler

(1958) did find differences attributable to therapist sex in studying the interaction with patient productivity. It will be recalled that patients producing few responses on the Rorschach are likely to ter minate in the first sessions. Hiler found that female therapists lose fewer of their nonproductive patients. But this was offset across all clients because they tended to lose more of the patients with high

R. For the predominantly male samples in the studies reviewed, thera pist sex (item 26 on Figure 1) does not seem to be a relevant variable 106

in dropout rate. Studies on the overlapping variables of race and

socioeconomic background of the therapist similarly find little direct

effect· Yamamoto, James, Bloombaun, and Hatten (1967) found that

therapists scoring high on ethnocentrism lost proportionally more

blacks than whites by the sixth ses'sion than did their less ethno

centric counterparts. This finding leads reviewers to conclude:

"Research does suggest that the attitudes and biases of the therapist -

for example, in regard to blacks, lower-class people, and women may

be more important than the demographic issues themselves" (Parloff,

Waskow, & Wolfe, 1978). In session behavior is more predictive than

the therapist's demographic background.

There is clearly a therapist contribution to the dropout question. Gibby et al. (1954) found that some therapists were able

to retain the unproductive patient. Hiler (1958) found the same pattern and suggested that therapist warmth and sex were critical mediator variables. McNair et al. (1963) also found a group of therapists who seemed to select and retain a disproportionate number of predicted quitters. But they were unable to discriminate this

group by sex, profession, experience, amount of personal therapy, A-B

type, judged competence, or liking for the patient.

The source of these therapist effects remain elusive. Studies

report failure to discriminate based on professional discipline (Lorr et al., 1958; McNair et al., 1963; Sullivan et al., 1958). Reviewers see experience as important (Dodd, 1970; Meltzoff & Kornreich, 1970), with Baekeland and Lundwall (1975) finding six of seven articles 107 indicating that experience is important. Of the studies reviewed here, only two considered experience significant, and no comparisons were presented in one (Fiester & Rudestam, 1975) while the other compared medical students to residents (Dodd, 1976). No difference was found in length of stay as a function of therapist experience in a number of standard studies from the literature (Caracena, 1965; Lorr et al., 1958; McNair et al., 1963; Saltzman et al., 1976; Sullivan et al., 1958) which were available to the reviewers. This review con cludes that faith in the central importance of experience has blinded others to the conflicting evidence on the subject. Further research, including this study, must evaluate this variable. There is more support for the related concept of therapist competence. McNair et al. (1963) found no differences and Garfield et al. (1963) found only a trend for the competent to retain more clients. But Hiler (1958) found productive patients remained for those rated most competent by judges (55% as compared to 29% kept by less competent therapists), and

Saltzman et al. (1976) found a significant relationship between con tinuing and patients' ratings of therapist commitment and competence after the third session.

Studies of therapist personality traits offer mixed results.

The often-examined A-B type offers little to the explanation of the dropout phenomenon (McNair et al., 1962, 1963). Reviewers have sug gested that therapists must permit dependency (Baekeland & Lundwall,

1975; Meltzoff & Kornreich, 1970). Caracena (1965) and Winder et al.

(1962) found no significant differences, but these results may be 108 attributed to the high acceptance of dependence in their therapist sample. When patients expressed direct dependence on them personally

(item 24), therapists did respond differently. And here the accep tance versus avoidance of dependence during the first two sessions was almost 100% accurate in predicting length of treatment (Winder et al.,

1962). Conversely, the detached therapist may discourage patients remaining in treatment. This may be particularly important to the traditionally oriented woman who expects to be able to be dependent on an active male therapist (Heilbrun, 1973). Kamin and Coughlin

(1963) interviewed patients after the conclusion of treatment. Most criticisms from both sexes focused on the quality of the relationship, specifically, on the perception of therapists as being silent and aloof.

Hiler (1958) failed to find the expected disproportionate attrition of unproductive patients of passive therapists. But Saltzman et al.

(1976) found that those who ultimately dropped out before the tenth session rated their therapists significantly lower on concern, parti cipation, and commitment through the first five sessions. By the third session these patients also felt they had experienced less respect and understanding from their therapist. Along the same lines,

Lesser (1961) found low therapist empathy related to early termination.

The importance of the counselor's warmth (item 23) is stressed in

Dodd's (1970) review and in Hiler's (1958) report that it removed the preexisting tendency for the unproductive patient to leave early.

But Freedman et al. (1958) failed to find a main effect for therapist warmth with their schizophrenic population. Instead, there was a 109 significant interaction in which there were more remainers when a warm therapist worked with a patient who acknowledged his illness or a detached therapist was matched with a denier. This last finding re emphasizes the power of the interaction during the session as a deter minant of premature termination of treatment. On the balance, however, patients will be more likely to continue if their therapists are able to accept dependency; be active and involved; and offer the facilita tive conditions of respect, empathy, and warmth. It is likely that it is through these variables that any contribution of therapist experi ence or competence is expressed.

The importance of interaction is also evident in the predictive value of therapist liking for the patient (item 25) and interest in the presenting problem. Despite the failure of McNair et al. (1963) to find a significant relationship, the tendency to lose the disliked patient is stressed by reviewers (Baekeland & Lundwall, 1975; Garfield,

1978; and Meltzoff & Kornreich, 1970) and empirically supported by

Caracena (1965) and Strickland and Crowne (1963). Moreover, the ef fect is mutual. Kamin and Coughlin (1963) found that 70% of patients with positive feelings toward their therapist rated themselves as improved after treatment, compared to 0% improvement for those who disliked the counselor. Interest in the patient's problem was found to relate to longer continuation in research by McNair et al. (1963).

Similarly, Garfield et al. (1963) found willingenss to accept the patient into therapy associated with the patient's length of treatment.

A good deal of outcome research has focused on the match of llO the patient and therapist in terms of personality or background. The field is complex, with researchers variously championing the value of similarity, complementarity, or positing a curvilinear relationship between degree of therapeutic success and similarity (Parloff et al.,

1978). A powerful argument can be made that pairs from a similar back ground will have the advantage in ease of communication, shared values, and common expectancies. McNair et al. (1962) adopt such an approach in suggesting type B therapists, those with interests in skilled labor and technical occupations, may have more in common with the predom inantly lower class client seen at a Veterans Administration Clinic.

In the absence of contradicting evidence, this review accepts the contention that similarity in background (item 28) encourages perse verance in treatment (Baekeland & Lundwall, 1975; Meltzoff & Kornreich,

1970). Mendelsohn (1966) compared staff and clients on the Meyers

Briggs Type Indicator to assess the importance of similarity in cogni tive style. The relevance of the results is attenuated since the study used students reporting for one to six sessions of educational or vocational counseling. It is questionable whether such results can be generalized to the psychopathological veterans considered in the present study. But Mendelsohn found cognitive dissimilarity led to early termination, while the similar pairs were variable in duration of treatment. He concluded that "similarity is something of a necessary condition for continuation but it is certainly not a sufficient condi tion as well" (p. 231).

Expectancy match. A great deal of research has been focused 111 on expectancies. Based on the model of placebo effects in medicine, it has been argued that therapeutic results depend on the mobiliza tion of the client's hopes more than on specific techniques (Frank,

1968). Investigations have included attempts to manipulate patient expectancies. So true experiments have been done, in addition to the naturalistic studies relied upon so exclusively in the studies re viewed to this point. To avoid a common confusion, expectancy for improvement, which includes prognosis, will be distinguished from match of role expectations, which is a variant on the similarity issue addressed above.

Expectations for improvement is a translation of Frank's measure of hope. It may be operationalized as a single global measure

(as in therapists' prognostic rating), as a questionnaire, or as the difference between current and anticipated self-descriptions. Scores on one of these measures are crosstabulated with an outcome measure.

Wilkins (1973) has criticized the frequent use of global ratings of improvement or self-report, because of the possibility of self-serving bias. Martin and his associates (Martin, Moore, Stern & McNairy, 1977;

Martin, Stern, Moore, & Lindsey, 1977), administering the MMPI at intake and discharge as a measure of change, were able to demonstrate significant correlations of improvement with expectation as rated by therapists and by patients undergoing their first course of inpatient treatment. From both perspectives, expectation was more highly cor related with the initial MMPI than the one at discharge. If expec tancies were causative, the opposite would be expected, namely, there 112

should be more impact on the final measure of adjustment. Martin

et al. concluded that both patients and therapists are able to read

preexisting clues in making their judgments about probably outcomes.

Expectations are predictions.

Outpatients may be less successful in predicting outcome, or

at least their expectations are not clearly related to likelihood of

continuing in psychotherapy. Heine and Trosman (1960) found no dif

ference in expectations for improvement among those terminating before

or after the fifth session. Garfield et al. (1963) found role and

outcome expectancies unrelated to length of stay or change. Goldstein

(1960) found duration of treatment and subjective ratings of change

were unrelated to initial expectations for improvement of 15 students

treated at a university clinic.

Therapist expectations for improvement may be more important.

Goldstein's (1960) therapists' predictions were significantly related

to duration of treatment and degree of improvement. Sherry (1977)

examined 71 patient/therapist pairs at a Veterans Administration clin

ic and found that a negative prognosis was more likely for those who eventually terminated. But Garfield et al. (1963) cautioned against

assuming a simple main effect for therapist outcome expectations.

Match of expectations is important. A disproportionate number of ter minators were found to have therapists expecting more change than anti

cipated by the patient. Such cases may well be examples of disagree ment about therapeutic goals rather than failed therapy (Baekeland &

Lundwall, 1975). 113

A number of studies examine the importance of expectancy match in regard to treatment goals, duration, and roles (item 29 of Figure 1).

Most report them to be important. Heine and Trosman (1960) found significantly more remainers if patients and therapists agreed on treatment aims (symptom relief through medication and information or behavior change through talk and advice). However, Gliedman, Stone,

Frank, Nash, and Imber (1957) found terminators only slightly and non significantly more discrepant with conventional treatment goals than were continuers. Sandler (1975) and Sherry (1977) found that degree of agreement in predictions of duration was significantly related to the likelihood of continuing in treatment.

Congruence in role expectations is significantly related to remaining in therapy according to reports by Gulas (1974), Heine and

Trosman (1960), Overall and Aronson (1963), Rapaport (1976), Sandler

(1975), and Straker et al. (1967). Some dissenting reports have been found. Garfield et al. (1963) found judges ratings of congruence unrelated to therapy duration. Sherry (1977) obtained mixed results.

Terminators were more discrepant in regard to the types of techniques used. But remainers were less likely to agree about expectations for therapist detachment. Horenstein and Houston (1976) argue that too much similarity may be counterproductive, as it minimizes the need to change. They found a quadratic trend. Remainers were moderately dis crepant. Those that were extremely congruent, like those patients who had inaccurate expectations, were more likely to terminate during the course of a three session orientation group. 114

Expectancy disconfirmation is a part of all therapy. Gliedman et al (1957) found that 71% of the dropouts and 63% of the continuers had goals for therapy that were at variance with those conventionally associated with insight oriented psychotherapy. Overall and Aronson

(1963) found that their model patient found the treatment to be less active, supportive, and medically oriented than expected. Such dis crepancies were especially characteristic in the treatment of lower class patients and may be the basis of the social class effects dis cussed previously (Garfield, 1978; Heitler, 1976; Malzer, 1980).

Some of the more recent studies reviewed (Rapaport, 1976; Sherry,

1976) comment that even lower class patients have more accurate ex pectations than expected. It may be that pop psychologies and the greater visibility of the helping professions during the last decade have contributed to a greater public awareness of the therapy process.

If a course of therapy continues, role expectancies converge (Gulas,

1974; Sandler, 1975). The critical issue is not the fact of discrepant expectations but the way in which they are managed. Either differences of great magnitude or therapist insensitivity to their existence is likely to result in a brief and unsatisfactory course of treatment.

Direct education has been used to correct mistaken assumptions.

Hoehn-Saric, Frank, Imber, Nash, Stone, and Battle (1964) used a role induction interview to explain expected behaviors, realistic outcomes and when they should become apparent, and the nature and handling of superficial resistance. Using a true experimental design, with thera pists blind to the manipulation, patients given the education were 115 rated significantly better than controls on a variety of process and outcome variables. Relevant to the current study is the observation that experimental subjects attended an average of two more sessions during the four months of treatment evaluated. Later research by

Sloane, Cristal, Pepernick, and Stapler (1970) clarified that it is the role socialization rather than hope induction that accounts for improvement. Heitler's (1976) review stresses that reduced attrition is a consequence of such efforts.

Wilkins (1973) criticized the research of expectancies and questioned the spate of hopeful reports it has generated. Signifi cant results were attributed to reliance on global measures subject to demand characteristics. While this contention may be applicable to some, it is clearly not an explanation for all of the results re viewed. And these results have been remarkably consistent in showing that common role expectations (including estimates of duration and, possibly, of goals) are crucial for developing a sustained therapeutic relationship. Wilkins is correct, however, when he points out that expectancies are hypothetical mediating variables without direct or observable causal impact on length of treatment. People do not quit therapy becuase their understanding of the patient role is wrong.

They quit because they feel frustrated with the lack of desired change.

Expectancy is a summary label for a range of correlated beliefs and behaviors. But this name is preferable to other summary labels current in the field, particularly social class, because it implies something that is dynamic. Where social class evokes a sense of immutability 116

that can only be accommodated (perhaps by offering less ambitious

forms of therapy), expectancies can be modified and altered. Direct education becomes feasible, whether formally through therapy sociali

zation training or informally during the discussions in an intake in

terview. The concept of expectancies is useful because of its explana tory power and because of the activist stance it offers for the clinician.

Other factors. While patient and therapist characteristics and the interaction between them are of central importance in deter mining continuation in treatment, the therapy operates within a wider clinic and community context. This context has been barely considered despite its evident importance. The general cultural attitude toward mental health and the patient role influence the likelihood of seeking out and persisting in treatment. As the mediator of these values, the family clearly plays an important role. Dengrove and Kutash (1950) in an impressionistic report suggest the lack of family support, par ticularly the expectation to "handle things on your own", is important.

Freeman (1961) interviewed family members shortly after the return of a psychiatric inpatient. He found that those with low levels of in telligence and education were less positive about the treatment offered or the possibility of recovery, more likely to blame the expatient, and less likely to conceptualize the illness in the psychological terms that support further treatment. These attitudes are similar to those typi cal of lower class patie·ilts so often cited as the reasons for premature termination. Most discouraging is Freeman's conclusion that these 117

beliefs are not easily changed. Frank et al. (1957) obtained retro

spective ratings from patients in which dropouts, compared with re

mainers, said there were less advantages in continued treatment, less

support, and more opposition from relatives. The theoretical impor

tance of the family context and the limited empirical support available

ratify Baekeland and Lundwall's (1975) call for more study of this area.

The clinic environment, whether in terms of physical facilities or policy, also deserve more study. Feldman, Lorr, and Russell (1958) report that large VA clinics keep patients longer. Dengrove and Kutash

(1950) list short sessions, therapist absenses, changes in treatment personnel, pretherapy hurdles, and doubts about confidentiality among

the factors that discourage patients. The limited research available minimizes the importance of session length (Lorr et al., 1958) or frequency (Lorr et al., 1962). Baekeland and Lundwall (1975) suggest that retention rates will increase if waiting lists are shortened, significant others are involved, and a range of therapeutic modalities are made available. Bailey et al. (1959) and Dodd (1970) found longer continuation if patients request and are offered medication, while

Kotkov (1958) found that initial willingness to accept psychotherapy predicted persistence. So patients offered their choice of therapy type remain in treatment. Although Lorr et al. (1958) found no signi ficant differences across modalities, two studies from the John

Hopkins team found higher termination rates among group therapy mem bers. Many drop out before the first session, suggesting initial 118 apprehensions (Frank et al., 1958). But even those attending a ses sion experience more attack and less support than those in individual therapy (Nash, Frank, Gliedman, Imber, & Stone, 1957). Although the

termination rate may be higher from groups, "in many respects the psychological characteristics of the group therapy dropout parallel those of the dropout from individual psychotherapy" (Baekeland &

Lundwall, 1975). For example, Yalom (1966) identified features such as intimacy problems, lack of insight and psychological sophistication, and lower socioeconomic status and education as characteristic of ter minators from groups. To summarize, the type of service offered, and its match to patient preferences, are important variables. But there is no need to look for separate profiles of terminators from different modalities.

One last contextual issue to be considered is the contribution of so-called "reality factors". Therapy becomes impossible if it interferes with critical work or family tasks, if transportation ob stacles cannot be managed, or, most clearly, if the patient moves out of the area. Such difficulties are commonplace, especially among lower class clients. Garfield (1963) found 6 of 11 dropouts cited reality factors, as did at least 6 of 37 interviewed by Brandt (1964), 6 of

21 inpatients leaving against medical advice (Scheer & Barton, 1974), and 5 of 35 group therapy droupouts interviewed by Yalom (1966).

Such cases should not be included with the terminators in research attempting to identify psychological patterns of the patient or thera pist effecting length of treatment. Yet a majority of the studies 119

reviewed failed to specify how such cases are handled. This is a methodological problem. The current research project will exclude

from consideration, among either terminator or remainer groups, those who quit therapy because of reality factors or are referred for treat ment elsewhere.

This concludes the review of research on extrapatient factors

involved in premature termination from psychotherapy. With the signif

icant exception of research on patient expectancies, it is even more

sketchy and unsystematized than that on patient factors. Accordingly,

conclusions must be presented cautiously. Among the therapist vari ables contributing to early dropout are a lack of similarity or famil

iarity with the patient's background; low competence, as evident in an inability to tolerate dependence, a detached or passive stance, and low levels of facilitative conditions; dislike of the patient or dis interest in the presenting problem; and an inability to bridge dis crepent expectations about the therapy process and outcome. A summary of all factors judged to have been empirically substantiated is pre sented in Figure 2.

Self-consciousness and the treatment dropout phenomenon.

It is clear that there has been no direct examination of self-con sciousness levels of terminators as compared to remainers. Theoreti cally, terminators would be expected to have extreme scores on the

Self-Consciousness scale: high ratings corresponding to flight from

therapy because of the anxiety elicited, low scores signifying lack of requisite skills and motivation to use insight oriented therapy. In 120

Patient Factors:

Low socioeconomic status Low intelligence Lack of verbal skills Lack of cooperation Somatic presentation Low anxiety or low depression No previous therapy Counterdependence, inability to tolerate intimate relationships History of failure to complete projects

Therapist Factors: Low competence Inability to tolerate dependence Detached and passive Low levels of respect, warmth, and empathy Dislike for patient Disinterest in presenting problems

Interaction Factors: Dissimilar background or cognitive style Unresolved dissimilar role expectations

Contextual Factors: Lack of family support Group therapy Failure to provide desired treatment modality Other reality factors

Figure 2. Summary of variables associated with premature termination from psychotherapy. 121

the absense of direct measurement, the task is to reevaluate variables

that may be conceptually related to self-consciousness.

The patient characteristics reviewed relate termination to the

lack of requisite self-monitoring skills. The closest conceptual parallel to SC is psychological mindedness. Whether this is opera

tionalized as the lack of verbal skills or as focus on somatic inter pretations of the presenting problem, the likely terminator is portray ed as one who is unused to examining behaviors or affective states and comparing them to internal standards. The observation that low levels of anxiety or depression are typical of terminators also suggests low levels of self-consciousness. It appears that the dropout enters therapy without the habit of self-monitoring, and, presumably, without

the skills to handle it productively.

The demands of therapy, with the heavy emphasis on self-scruti ny, then propel the patient into an area for which he or she is ill prepared. If high levels of self-consciousness lead to premature ter mination from therapy, as extrapolated from the findings that objective self-awareness (OSA) is avoided, then features of the therapist and context contributing to dropping out should be those likely to force self-consciousness on the patient. This seems to be the case with at least some of the variables found to predict premature termination.

Group therapy produces more dropouts. It also should dramatically heighten social self-consciousness, and attendent anxiety. That this seems to be the case is evident in patient descriptions of experiencing the group as threatening and attacking. 122

The therapist's role, from the perspective of self-awareness theory, is to carefully manage the patient's self-monitoring so that internal standards can be accessed, to motivate desired behavior change, while debilitating anxiety is avoided. To this sensitive task, the therapist brings a number of tools. The active or directive therapist can distract the patient from excessive self-scrutiny. The supportive and empathic therapist can, in effect, desensitize the per son through the offering of facilitative conditions so that problematic areas of the patient's life can be explored without too much self criticism. Warmth, respect, and interest all communicate an acceptance that the patient can use as a model. Empathy indicates that an ex perience is understandable to another, while the nonjudgmental tone serves as a cue that self-chastisement is unnecessary. So therapist characteristics would be expected to be important mediators of length of stay. The silent or detached therapist is easily perceived as the judgmental other, heightening both self-scrutiny and anxiety, and so placing the therapist in the role of the passive accomplice in flight from treatment. In contrast, the competent therapist helps to manage self-consciousness by alternating among support, the implicit protec tion of encouraging dependence, the desensitization offered in com municating warmth and empathy, the modeling of expressions of attrac tion or interest in internal experiencing, and the distraction from or restructuring of internal standards that go with being directive. To use these skills with the deftness required for a successful outcome, the therapist must be very attuned to the patient's experience. Dis- 123

similarity in expectation, background, or cognitive style contribute

to premature termination because they block the "third ear" and so

render inoperative the array of tools for managing self-consciousness.

Self-awareness theory, as applied to the early stages of ther

apy, would suggest that a patient will leave because of insufficient

anxiety to motivate change or because of excessive anxiety resulting

from an immoderate dose of self-scrutiny. Most of the patient charac

teristics delineated as correlates of premature termination in Figure

2 are hypothetically associated with the absense of self-scrutiny.

Most of the therapist and interaction factors relate to excessive self

consciousness developing in the patient. In this sense, OSA theory can

draw together a number of disparate and seemingly unrelated sources of

flight from therapy. The theory has the unique advantage of being able

to explain both sources of early termination cited by Saltzman et al.

(1976): lack of anxiety after the first session and dissatisfaction with the therapy relationship during the third to fifth visits.

Personality Typology

This research tests specific applications of self-awareness

theory to two areas of clinical psychology. The last section relates

to the first of them: the prediction of premature termination from

therapy. This section will address the second, which is the distinction

among personality types in the level and handling of self-awareness.

A psychodynamic personality typology. The prevailing diag

nostic nomenclature is inadequate for the purposes of this research. 124

The number of categories, even when restricted to functional disorders

of adults, totaled a prohibitive 82 under DSMII and has risen with

DSMIII (American Psychiatric Association, 1968, 1980). Distinctions

between these diagnoses are notoriously unreliable (Spitzer & Fleiss,

1974). Even when syndromes are discriminable, they may obscure rather

than highlight important personality differences: "Although all the

individuals described as having the same mental disorder show at

least the defining features of the disorder, they may well differ in

other important ways that may affect clinical management and outcome

(American Psychiatric Association, 1980, p. 6). This is especially

true in diagnosing disorders of habit; a person may be classified

because of behavior in a restricted area without reference to overall

personality style and adaptation. These drawbacks more than offset

the advantage of using the clinician's language system.

A condensed and revised typology was developed for this study.

It collapses a number of the dimensions that lead to the extensive

and confusing scope of DSMII. First of all, it focuses on the pa

tient's pervasive personality style or life style. Circumscribed

behavior patterns (e.g., substance misuse or paraphilia) are not

regarded as a substantial enough sampling of the individual's manner

of relating to the world. Second, it collapses the DSMII distinctions between psychoticism, neuroticism, and personality disorders. Research

suggests that degree of pathology is better represented on a continuum

than as discrete categories (Eysenck, 1960; Luborsky, 1962). Clinical distinctions are also based on degree of anxiety and affective tone. 125

While fine discriminations in these two dimensions account for many of the diagnoses (in DSMII there were, for instance, five diagnoses in which depression is a defining element), they are seen as being symptomatic and more or less transient rather than basic to the pa tient's personality.

When these dimensions are collapsed, the residual categories are personality types which function much like the Axis II diagnoses of DSMIII. That is, they describe the basic and sustained styles which give rise to and yet persist beyond the symptoms that lead pa tients to seek treatment. In this sense they resemble, and borrow from, typologies developed in psychoanalytic theory. The styles are described in terms of characteristic defenses, likely symptoms or complaints, typical manner of relating, and quality of the transference.

There is empirical as well as theoretical support for such a manner of conceptualizing core personality differences. Miller and

Magaro (1977) successfully used cluster analysis to identify four distinct types. As with this research, they defined their types as applicable across the entire spectrum of adjustment, and they defined personality style in terms of "a specific combination of psychological defenses, cognitive and affective styles, belief and value systems, moral development, etc." (p. 460). Profiles were developed for hys teric, compulsive, depressive, and character disorder types. These were used to predict performance on a variety of brief personality measures (Rotter's Locus of Control Scale, Byrne's Repression-Sen- 126

sitization Scale, etc.). The experimental hypotheses for each group

specified which tests should be associated and the direction of devia

tion from test means. Protocols were then factor analyzed, with re

sults evaluated in terms of the number of predictions verified. When

a sample of 107 Introductory Psychology students were used, the hys

terical and character disorder types were validated. However, the

compulsive and depressive styles overlapped. A replication with 130

juniors and seniors with different majors, found increased discrimina

tion between types. Five clusters emerged and 18 of 23 scale predic

tions were verified. Hysterical (seven of eight predictions) and

character disorder types (four of five predictions) were crossvali

dated. The depressive cluster conformed to expectations (five of six

scale predictions) and explained the most variance. However, the sec

ond most important cluster represented an overlapping of hysterical

and depressive features. The compulsive style was only "moderately

supported", with two of four scale predictions confirmed. Scales

loading for the four predicted types are listed in Figure 3.

Miller and Magaro's (1977) findings offer general support for

the validity of a brief typology and do much to clarify its content.

Character disorder and hysterical types will be included in the current

research scale. Depression will not be. Despite its important contri bution to the variance, Miller and Magaro's results support my convic

tion that depression overlaps too many other clinical conditions to be considered a meaningfully discriminable personality style.

The five personality types and descriptions, as given to 127

Hysteric: Empirically: altruistic, repressor, field dependent, extraverted, lowa neuroticism, low dogmatism, high sensation seeking; Descriptively: reliance on repression, global cognitive and affective style, suggestible, naive, dependent.

Character Disorder: Empirically: high sensation seeking, Machievellian, extraverted, low altruism; Descriptively: egocentric, manipulative, highly developed social skills, lack of affect.

Depressive: Empirically: low sensation seeking, sensitizer, field dependent, introverted, high neuroticism; Descriptively: cognitively rigid, lacks personal self-worth, obedient to authority, reduced satisfactions, prejudice toward outgroups.

Compulsive: Empirically: field independent, introverted.

Note. From Miller and Magaro (1977). aThe adjectives "low" and "high" indicate a score deviating more than .5 standard deviations from the grand mean.

Figure 3. Characteristics of different personality styles. 128 clinicians, are listed in Figure 4. They were developed with the in tent of including the major personality styles while still preserving the economy of the scale and the clarity of distinction between styles.

The choice of styles, and their description is influenced by psychoanalytic theorizing. The distinction between obsessive and hys terical styles, for instance, was made early on by Freud in his attempts to explain patients presenting with ideational versus physical symptoms

(Freedman, Kaplan, & Sadock, 1972). At the first level, the scale draws on the tripartite model. With the consolidation of the superego, a second regulatory mechanism is available to assist the ego in the regulation of drives. Defects in superego development will lead to a personality distinct from those attributable to problems in ego devel opment. The sociopathic category is intended to represent those prob lems derived from the failure of the superego to operate effectively.

Drives are acted out instead of being blocked from discharge (Pervin,

1970). The references to substance abuse and conflict with authority capture this inability to delay impulse gratification. The likelihood of external compulsion to enter treatment is further testament to the absense of self-regulation. The sociopathic category is essentially equivalent to Miller and Magaro's (1977) empirically validated Charac ter Disorder type. Because it is traced to superego defects it is conceptually and clinically distinct from other typologies which de scribe different styles of ego functioning.

One crucial aspect of ego functioning is management of defense mechanisms. Because individuals develop habitual defensive styles, the 129

Hysterical: overuse of repression; likely presents with somatic concerns or the feeling of being used in personal relation ships; exhibitionistic, dramatic; relates so that the therapist has a sense of engagement with the patient despite a focus on external events.

Paranoid: projection prominent; presents with concerns about being isolated or deliberately abused by other people; relates vaguely and guardedly; therapist feels a strong pull to confirm the patient's perspective or else be confronted with the patient's underlying anger.

Obsessive: isolation and intellectualization prominent; problems related to feared loss of control or, if loss of control has been experienced, feelings of depression, anger, and guilt; high expectations for personal productivity and morality; insight into thoughts and motives without affect; patient is controlling, and the interview may become so wordy as to become boring.

Passive: schizoid and avoidant; may present with multiple fears and feelings of helplessness; inward focus on fantasy; history suggests lack of social skills and contacts; therapist may still feel detached after the interview; include: passive dependent, passive aggressive, and narcissistic char acters.

Sociopathic: history of acting out tensions; mlnlmlzes problems, and likely attends the interview because of pressure from family, courts, or other outside source; may be problems with alcohol or drugs; appears socially skilled and com fortable, and yet the contact remains superficial and the relationship feels tenuous.

Figure 4. Personality types and descriptions. 130

analysis of primary defenses is an important aspect of dynamically

oriented assessment (Dewald, 1971), and a convenient means of dis

tinguishing personality styles. Again, hysterical and obsessive

styles are easiest to separate. Repression limits access to conscious

ness of problematic contents, resulting in the apparently naive and

global cognitive style of the hysteric. In contrast, the ideational

defenses of the obsessive split off the affective component while

permitting verbal awareness of the content (Doyle, 1968). There

sult is a characteristically detached insight, with emotions restricted

or appearing convulsively (Kennedy, 1977). The tension in both of

these defensive styles is internal: conscious versus unconscious,

affect versus cognition. This contrasts with the remaining character

styles, in which the conflict is between self and other, or, more accurately, between internal representations of the self and the ex

ternal object. This is clearest in paranoids, whose "capacity for delusions represents a confusion in their sense of identity between what is part of their personality and what is separate from it"

(Kennedy, 1977, p. 302). To some degree all use of projection and

introjection harkens back to their developmental precursors, the con

fusion of the self-boundary that occurs prior to the cognitive dif ferentiation of self and other (Gedo & Goldberg, 1973). In combina

tion with underlying anger, this explains why the paranoid's world is experienced as dangerous. To the passive type, for whom the world is also experienced as threatening, the defense is characteristically avoidance. This may include withdrawal into a dependent relationship, 131 if available. The upshot is a retrenchment that leaves the narcis sistic character with impoverished object relations or restricts the accessible world of the phobic.

There is a development flavor to much of the preceding descrip tion of defensive styles. It is not accidental. Defenses become available in a developmental sequence (Fisher & Greenberg, 1977;

Gedo & Goldberg, 1973). Repression, isolation, projection, and with drawal are successively more primary and primitive (DeWald, 1971).

Indeed, each of the personality styles can be presented as represen tative of a step in the developmental sequence posited by psycho analytic theorists. This will be done in the succeeding paragraph.

But a clarification is required first. Because of the strong etio logical focus of analytic theorizing, most psychic phenomena are related to development events. This includes degree of psychopathology.

Freud explained progressively more severe disturbance through refer ence to earlier developmental arrest or regression to more primitive stages. The implication would be that typologies representing earlier psychosexual stages would also include more disturbed patients. But this would be an oversimplification. Experiences at later develop mental stages may result in the secondary automony of certain core ego functions (Gedo & Goldberg, 1973). Even a regression from the oedipal to the oral stage may not lead to as much evident pathology as develop mental arrest at, say, the phallic stage. So while the linking of personality type to developmental level may help elucidate the primary defense, the likely presenting problem, and the basic interpersonal 132 stance, it does not define the degree of psychopathology. Type re mains independent of degree of disturbance.

Each of the four types of ego-functioning can be roughly identified with a different developmental stage. Passive types are most clearly linked with oral characteristics. As hypothesized by

Abraham and empirically confirmed since oral traits are most pertinent to the dimensions of dependent/independent, passive/active, and close to others/distant from others (Fisher & Greenberg, 1977). The passive and avoidant emphasis of the typology, with its phobic and pessimistic stance, captures the likely response to oral frustration. At this stage, the possibilities of interaction is limited to the excessive dependence of self-object union or to the absense of contact of the autistically self-preoccupied. In the passive character these quali ties are apparent in the therapist's sense of detachment or in the dependent pull underlying the passive dependent or passive-aggressive style of relating (Freedman, Kaplan, & Sadock, 1972). The paranoid type is associated with frustration of oral-aggressive drives. The lasting insecurity and ambivalence about trusting relationships is evident in the patient's guarded style and suspicions of others. The underlying rage communicated in therapy is an attempt to avoid depen dence as well as an expression of aggressive drives (Kennedy, 1977).

The obsessive has traditionally been associated with the anal period

(Fisher & Greenberg, 1977). A central and continuing theme of the anal character is the tension between defiance and the feared loss of love if the rage surfaces (MacKinnon & Michels, 1971). So the rigid 133

moral standards, the feared loss of control, the isolation of emotion,

and the controlling style of relating can be seen as sequellae to the

attempt to control anger and obstinance (Kennedy, 1977). The link

between the phallic stage and the hysterical style is evident in that

the exhibitionism and demand for attention of the hysteric closely

parallels the child's assertiveness and beginning sexual identity

awareness during the phallic period (Pervin, 1970).

The analogy between four of the personality types and the psy

chosexual stages has three purposes. First, it should accent the

central issues and relationship styles of each style. Second, it

indicates the distinctness of each type, and so justifies their in clusion in the typology. Third, and finally, it is intended to con vey the sense that this list of character types is reasonably compre hensive. While there might be some quibbling about descriptive phrases for one or another specific type, the descriptions are generally con ventional and noncontroversial. The disagreement that might be antic ipated relates more to the number and choice of character types. This study has attempted to make use of basic psychodynamic categories and to justify their use by indicating how they collectively cover the range of personality types suggested by the tripartite and genetic models from psychoanalysis. The presence of empirical evidence of the importance of at least two of the categories adds further support for the assumed validity of the scales.

Personality type and level of self-awareness. Among the characteristics that differ across personality types, the level and 134 means of handling objective self-awareness can be expected to vary.

While I know of no studies within abnormal psychology or self-aware ness research that bear directly on this issue, there is some indirect experimental work and good theoretical grounds for expecting dif ferences.

Eysenk (1960) has attempted to redefine conventional diagnos tic categories in terms of his factor analytic schema. In the course of doing so, he presents replicated and consistent findings that neurotic groups, at least, vary along the dimension of introversion extraversion. Obsessives are highly introverted, or attuned to their internal, subjective world. Hysterics and psychopaths are on the op posite, extraverted pole. Obsessives can be expected to be the most self-conscious of the three groups overall. At the subscale level, obsessives would have the higher scores on private self-conscious ness (PR-SC), because of their attention to at least their internal ideational status. But hysterics and psychopaths, with the extra verts' greater general social skills, may well have higher scores on public self-consciousness (PU-SC).

Table 9 summarizes the expected rankings of the five person ality types on the Self-Consciousness scale. In deriving these hypoth eses, the empirical evidence discussed above is supplemented by ex pectations based on the typical defenses and interpersonal style of each group. Passive types are expected to be the most self-conscious overall. Their self-preoccupation should be evident in high scores on

PR-SC, while their social anxiety (SA-SC) is expected to be highest 135

Table 9

Expected Rank Order of Personality Types

on the Self-Consciousness Scale

Self-consciousness subscale Total self Personality type Private Public Social Anxiety consciousness

Passive 1 2 1 1

Obsessive 1 2 2 2

Hysterical 2 1 3 3

Sociopathic 3 1 3 4

Paranoid 3 3 3 5

Note. The group mean will be expected to differ significantly from that of any group with a different number. An identical ranking has been given whenever there is no basis for assuming significant dif ferences. 136

since their phobic/avoidant style entails experiencing anxiety

regularly and consciously. Passive types are regarded as being

intermediate in their sensitivity to the social impact; less of their

self-image is involved with impression management than would be the

case for hysterics and sociopaths. Obsessives will be highly self

aware, as evident in their capacity to report internal events. However,

the isolation of affect should limit their cognizance of the anxiety

that arises from their self-scrutiny.

Hysterics are renowned for their naivete. Despite their

ability to express superficial emotionality, there is an absense of

appreciation of internal states and motivation that will lower their

subscale scores on PR-SC. Even the characteristic daydreaming centers

on recognition in an interpersonal context. Accordingly, their highest

ranking will come in the area of public self-consciousness, which

emphasizes appearance and social impression. The hysteric tends to

experience less anxiety than other personality types (MacKinnon &

Michels, 1971). The fluid defenses, the reliance on repression, and

quick ventilation through emotional expression all are expected to

contribute to lower scores on SA-SC. As a consequence,hysterics should

have lower overall Self-Consciousness scores than passive types or

obsessives. A somewhat similar profile should apply for sociopaths.

They will have their highest subscale ranking on public self-conscious ness, since sensitivity to the impact of their interpersonal style is basic to obtaining their gratifications through other people. But

this external, goal-oriented focus will curtail internal attention and 137 lead to low scores on PR-SC. Their response on the SA-SC is problematic. Conventionally, sociopaths are regarded as unable to experience guilt or anxiety. But this has been interpretted differently:

They exhibit little if any anxiety, but this is a result, according to some, of their low tolerance for anxiety. Many of the dynamic mechanisms employed by the antisocial personality react very early to fend off, defend against, or snuff out the slightest bit of anxiety • • • . What seems like an indifferent exterior may be the outer evidence of a massive psychological effort to hold anxiety at bay. (Kennedy, 1977, pp. 283-284)

Presumably, the quick denial of anxiety will prevent their reporting concerns about social evaluation on the SC scale. Combined with low scores on private self-consciousness, they should be ranked low on the scale as a whole.

The lowest SC scores, overall and on each subscale, are expected from the paranoid patient. His whole dynamics force a vigilant focus on what is happening "out there" that must severely limit attention to internal ideation and affect. Even the scale measuring anxiety in social settings, a characteristic experience for the paranoid, will have a low score because the items are phrased in the first person.

The paranoid sees the world in terms of "they do" rather than "I feel".

Related to this personality style is the guarded response to all re- quests for information, including the research questionnaire. The patient's response bias will compound his phenomenological lack of self- consciousness and lead to very low SC scores. So the paranoid is hypothesized to be lowest overall and not significantly higher than any other group on any subscale.

In the absense of previous experimentation in this area, the 138 hypothesized relationship between personality type and level of self consciousness must be regarded as tentative. This is particularly true in regard to predictions of subscale rankings. At the most con servative, the total SC mean for the passive and obsessives would be expected to be significantly higher than that from the other three types. Additionally, the personality types will have identifiably different levels of self-consciousness. This assumes, as this chapter has argued, that the five styles can be reliably classified, are mean ingfully distinct from one another, and that self-awareness is an im portant contributer to their distinctiveness. The detailed hypotheses provide a predictive standard in terms of which to evaluate the via bility of these assumptions.

Summary of Hypotheses

This research constitutes an introductory exploration of the clinical implications of self-focused attention. Different aspects evaluate the relationship of self-consciousness to degree of pathology, premature termination from therapy, and personality type. For the sake of clarity, results will be presented separately for each of three experiments. The specific hypotheses are presented below:

Experiment 1: Self-Consciousness Levels of Psychiatric Patients:

1. Response to self-focused attention is less differentiated

among patients than in the standardization samples. The

three SC subscales will be more highly correlated than

among normals. 139

2. Patients are more self-conscious, or experience more dis

ruption than do normals in response to self-focused atten

tion. Social Anxiety (SA-SC) will have a higher mean than

previously reported. Total SC scores will vary signifi

cantly from that found among normal samples.

3. There will be greater variability in subscale and total

SC scores among patients.

4. The SC scale will be valid as an index of propensity for

self-focused attention among clinical populations. It

will be generally independent of degree of pathology.

There will be minimal common variance between the Self

Consciousness scale and the Ego Strength scale (ES).

5. Pathology is related to specific facets of self-conscious

ness. Private self-consciousness and ES will be positively

correlated. But, SA-SC and ES will be inversely related.

Experiment 2: Self-Consciousness and Premature Termination from

Psychotherapy:

1. Excessively high levels of self-consciousness lead to

flight from therapy. Patients dropping out of treatment

between sessions one and five will be overrepresented

among the top SC scorers.

2. Patients drop out or fail to begin treatment because

they lack sufficient insight or anxiety to motivate change.

Dropouts and treatment decliners will be overrepresented

among those scoring lowest on SC. 140

3. There will be a curvilinear (inverted U) relationship

between total SC and the proportion of patients con

tinuing in therapy for six or more sessions.

4. SC means will be significantly different for continuers,

terminators, and those who decline to begin treatment.

5. Group therapy provides excessive stimulus to awareness of

the social self and so will be intolerable to those dis

positionally sensitive to their public presentation.

Terminators will be higher in PU-SC than those who continue

in group therapy.

6. Patients with previous therapy will have developed self

consciousness management skills. Those returning for a

repeat course of treatment will be less likely to drop out.

7. Self-consciousness is an important determinant of pre

mature termination. A discriminant function separating

remainers from terminators will include SC scores.

8. Since therapists contribute a variety of anxiety manage

ment techniques during the beginning of therapy, SC will

be significantly related to ratings of therapist experience.

Experiment 3: Self-Consciousness and Personality Type:

1. Different personality types vary in the level and handling

of self-consciousness. Obsessive and Passive types will

be greater than Hysterics and Sociopaths, who, in turn,

will exceed Paranoid patients in mean SC.

2. These differences result from predictions of differentiated 141

elevation on subscales. On PR-SC, Obsessives will be

significantly higher than Hysterics and Sociopaths. The

opposite results are expected for PU-SC. Passive patients

should exceed others on SA-SC.

3. Each personality type will have a distinctive profile

across the three SC subscales. The average value of an

item will vary from one subscale to another. For Socio

paths, PU-SC should be greater than PR-SC. For Passive

types, PR-SC and SA-SC will be approximately equal and

greater than PU-SC. For Obsessives, PR-SC is predicted

to exceed SA-SC. While for Hysterics,PU-SC should be

greater than either PR-SC or SA-SC. No specific subscale

profile is predicted for Paranoid patients. METHOD

Subjects

The research was conducted at a large, urban, Veterans Admin istration Hospital between May, 1979 and July, 1980. All subjects were male veterans, with the rare exception of those who were still in ser vice. As the setting is metropolitan and the treatment is free of charge, subjects were diverse in terms of ethnicity and economic back ground. The sample was bimodally distributed by age, with Vietnam era and World War II veterans predominating. Two groups were used in this study: 94 inpatients and 136 outpatients. The characteristics of these groups will be described, and data will be presented on how rep resentative they are of their respective populations.

The 94 inpatients were tested near the end of their hospital ization. The treatment philosophy of the facility stresses short term stays sufficient to stabilize the veteran. Those in need of chronic or custodial care were generally excluded. Participants were identified by nursing staff as likely candidates for referral to the

Mental Hygiene Clinic. During the research period, 43.1% of those eligible to participate did so. The main constraint on including sub jects was the restricted availability of the researcher to only one or two set sessions per week. Only those veterans available during the

142 143 group testing sessions participated. A minority of patients invited refused to participate or turned in unusable records (14.5%). But the presence of the researcher permitted collecting responses even 1 from those unable to read or attend to the materials. Because it is relatively inclusive, the sample is probably representative of the population. But because of the initial selection by nursing staff, it is possible that cooperative patients and those with clearcut discharge plans are overrepresented.

The initial sample consisted of 106 cases. Experimental attrition resulted from: four patients withdrawing consent during testing, one failing to sign the consent form, two failing to turn in all test forms, two records having random responses, eliminating two records from patients retested on successive admissions, and eliminating the record of the one female patient.

The inpatient sample was selected as probable referrals to the outpatient mental hygiene clinic. Of the 94 subjects, 77 had intake appointments scheduled, 50 were seen, and 29 began treatment. Ten of those seen were rated by the intake interviewer. In the main, the group sampled does represent potential outpatient clinic clients. Al- though less likely to be diagnosed as schizophrenic or neurotic than veterans included in the outpatient sample, they were comparable in

1 Sessions were held on Monday and Thursday afternoons between May and July, 1979. From August, 1979 through July, 1980 testing was done on Thursday mornings. 144

global diagnosis (which included psychotic, neurotic, personality dis

order, organic brain syndrome, and others as categories) and in per

sonality type. They were comparable in such other group descriptors as age, incidence of service connected disability, and in the distance between their residence and the clinic. Although the ten inpatients receiving intake ratings are too few for sensitive statistical com parisons, no gross differences were evident in their prognosis or in

their motivation for treatment. It seems justifiable to combine in patient and outpatient samples in evaluating the clinical impact of self-focused attention.

However, as a group, inpatients were handled differently when they entered treatment at the Mental Hygiene Clinic. To begin with, they were more likely to have had previous contact with the clinic.

Of the inpatients, 45% had previous referrals and 20% had had treatment there. For outpatients, the respective indices of familiarity with the clinic drop to 23% and 12%. Perhaps because of this prior contact, inpatients were seen marginally sooner for intake (9.6 days after re ferral as opposed to 14.3 days), had significantly fewer evaluation sessions (1.2 versus 1.7), and accordingly began therapy significantly sooner (17.1 versus 26.7 days). Because of the potential importance of these variables in determining length of stay in treatment, the distinction between inpatients and outpatients will be pursued in the phase of the research pertaining to premature termination from psycho therapy. Despite these differences in the way in which they were pro cessed at the clinic, the two samples were similar demographically. 145

The 135 outpatients were sampled from those seen for intake

at the Mental Hygiene Clinic. These patients can be divided into

three subgroups. For 113, the scale was administered at the time of

the intake interview, and was accompanied by an Intake Rating Scale.

An additional 12 completed the form at intake but were not rated by a

therapist. The remaining 11 completed the Self-Consciousness scale while awaiting a screening session with a staff psychiatrist. As such,

they returned at a later date for intake interviews. Each man had

previous contact with the hospital and was referred through the Admit

ting Office, a general medical floor or clinic, or Inpatient Psychia

try. Participants included both recurrent and first time users. Those

sampled constituted 17.1% of those having intake appointments. Because

the research documents were self-administered, the sample was poten

tially unrepresentative of those entering the clinic. Those whose

insufficient education or extensive psychological impairment prevented

them from completing the materials were excluded, as were those who neglected to return the forms to the receptionist, and those few (16

total) who refused consent. The most impaired and negativistic were likely underrepresented in this sample. Materials from 12 additional cases were excluded. These included two females and the ten veterans previously sampled as inpatients. According to the specific hypothesis being evaluated, other cases were excluded. Such, for instance, was

the case in the premature termination study for cases closed during the first five sessions because of hospitalization or mutually agreed upon cessation of treatment. In addition, incomplete records were available 146 for some cases. Figure 5 shows how the sample was composed for each phase of the research. Reading vertically, one can see the number of subjects having each combination of research materials.

The 116 outpatients whose records were used in the personality type study had each been rated by their intake interviewer on a series of questions related to diagnosis and suitability for treatment. In terviewers had also rated 96 other cases who did not complete the

Self-Consciousness scale. A comparison between these two groups per mits an assessment of the degree to which research subjects are repre sentative of the outpatient population. Participants were not statistically different from the 96 rated patients in diagnosis, global diagnosis, prognosis, motivation for treatment, referral, or recommended therapy. The two groups did differ in their expressed reaction 2 to recommended treatment, x (2)=11.54, .E_ ( .01, with those who completed the SC scale disproportionately represented among those neutral in their response. This finding runs contrary to what would be expected if participants were unrepresentatively compliant or enthusiastic. The evidence suggests that those sampled are representative of the psy chiatric outpatients attending the VA clinic.

Inpatients are demographically and clinically similar to outpatient participants, who, in turn, are not significantly different than a sample of outpatient nonparticipants. This network of rela tionships suggests that the sampling procedures were at least grossly successful in creating a representative sample. The demographic pro file of the combined inpatient and outpatient groups is presented in

Table 10. The modal veteran is middle aged, self-referred, and not Total n for each Type of data available Combinations phase of the research

Self-Consciousness scale X X X X X X X X 230

Ego Strength and Self-Consciousness X X X X 94 (Inpatients)

Termination and Self-Consciousness X X X X 133

Personality type and Self-Consciousness X X X X 116

Total n with each combination of data 24 40 6 29 44 4 77 6

Figure 5. Number of subjects participating in each of the component studies. 148

Table 10

Demographic and Clinical Management Characteristics

of Research Participants

Category Percent of sample

Degree of service connected disability (n=132) 0% 86.4 10% 3.8 20% .8 30% 2.3 40% .8 50% 3.0 60% .8 80% .8 100% 1.5

Global diagnosis (~=198) Psychotic 33.5 Neurotic 31.6 Personality disorder 13.1 Organic brain syndrome 1.9 Other 19.9

Rated prognosis (~=193) Poor 9.3 Guarded 21.8 Fair 34.2 Good 33.7 Excellent 1.0

Rated motivation for treatment (~=194) Poor 17.5 Fair 38.7 Good 38.7 Excellent 5.2

Patient reaction to recommended treatment (~=207) Negative 6.3 No reaction or neutral 19.3 Positive 74.4 149

Table 10 (Continued)

Category Percent of sample

Disposition at registration (n=108) Accepted for: Individual therapy 38.9 Group therapy 20.4 Family therapy 19.4 Hypnotherapy 3.7 Medication only 7.4 Referred 8.3 Discharged without referral 1.9

Previous therapy at the Mental Hygiene Clinic (~=140) None, never referred 68.7 None, previously referred 16.4 All episodes terminated within five sessions 1.5 One or more episodes with at least six sessions 13.4

Referral source (n=198) Admitting Clinic 61.1 Inpatient Psychiatry 22.7 Inpatient General Medicine 7.6 Outpatient General Medicine 8.6

Sessions before beginning therapy (~=90) One 52.2 Two 38.9 Three 6.7 Four 2.2

Category Average

Days between referral and scheduled intake (~=139) 16.94 days

Days between referral and beginning treatment (~=87) 32.36 days

Age (~=134) 40.72 years 150 service-connected. Despite considerable variability in diagnosis, his prognosis is moderately favorable. His first contact with the clinic is approached with a neutral attitude (evident in midscale modes on motivation) but an acceptant (74% positive) reaction to the recom mendation of individual therapy.

Materials

Copies of the materials used in the research are located in

Appendix A. Primary among them is the Self-Consciousness scale

(Fenigstein et al., 1975). All inpatient and outpatient participants completed this 23-item survey, which was modified for this study.

The changes were minor, consisting of definitions or synonyms of words found in pretesting to be too sophisticated for some veterans. Com prehension was enhanced because of these parenthetical additions.

None of the original scale was deleted. Because the changes were clar ifications rather than alterations, the scale used is felt to be fully comparable to the original. The other standardized test used was

Barron's (1953) Ego Strength scale. All inpatients completed this protocol as a free-standing instrument, apart from the rest of the

MMPI.

Two additional forms were especially designed for this study.

The premature termination research necessitated a review of each par ticipants medical record and the clinic appointment calendar. Data recorded from this review included such potential covariants of length of stay as age, degree of service connected disability, referral source I5I to the Mental Hygiene Clinic, elapsed time between referral and ini tial appointment, time between referral and beginning treatment, num ber of appointments prior to initiating therapy, number of sessions attended, type of treatment provided, and discharge status. As in dicated in Figure 5, this data was completed for I33 cases. Addi tional information on variables related to termination and to person ality type was collected from intake interviewers with the "Intake

Rating Scale". This protocol (available in Appendix A) included DSM-II diagnosis, recommended disposition and modality of treatment, a rating of the patient's reaction to the suggested treatment, and therapist ratings of patient prognosis and motivation for treatment. The in take interviewer also used this scale to identify the veteran's personality type, using the definitions provided in Figure 4. The combination of forms administered tapped the therapist, patient, and the objective/archival information sources.

Reliability of personality type. Because the third experiment is so dependent on the adequacy of the personality type ratings, a special reliability study was done for this item. Over the course of five weeks, therapists of the Mental Hygiene Clinic used the typology to rate each patient attending the Diagnostic Staffing group. In this manner II therapists rated 20 patients. Despite the name, the purpose of the group is to decide on the clinical management of problematic cases. Beyond asking the patient for a brief perspective on the nature of his problem and the services desired, there is no attempt to solicit diagnostic information or to make a decision as to probable diagnosis. 152

Nor is there discussion between staff as to classification. The interviews are short and, since done in a group context, intentionally superficial. So judgments of personality type were abstractions based on manner of relating in the group setting rather than inferences sup ported by a review of the symptoms. Lastly, to the extent that the cases which are difficult to manage are also likely to present diagnos tic dilemmas, these cases are unrepresentatively difficult to classify.

As such, the reliability estimate obtained approximates the minimum range of reliability available with the scale.

Of a possible 220 ratings of the 20 patients by 11 therapists,

159 (or 72.3%) were made. (See Appendix B for raw scores and tables presenting the degree of inter-rater agreement for each case and type).

Of the 20 cases, three were modally rated hysterical, four paranoid, four obsessive, five passive, and four sociopathic. The first analysis examined inter-rater consistency in two ways: as percent agreement with the modal diagnosis and as percent agreement as to the appropri ateness of a given typology for the case. The latter statistic ex presses findings on a relatively less challenging task, since it is often easier to decide which types are inapplicable than to choose a specific diagnosis. However, since the task involved five choices per patient (decisions as to the suitability of each of the five scales) a larger pool of observations was available and a more refined estimate of agreement could be made. Table 11 presents these statistics as a summary for the total sample and for each personality type separately.

The average agreement with the modal diagnosis is 68.6%. But 153

Table 11

Agreement between Clinicians

in Classifying Personality Types

Percent agreement about Percent agreement appropriateness of Personality type with modal type each type

Hysterical 65.2 89.3

Paranoid 66.7 89.3

Obsessive 57.7 86.2

Passive 63.6 84.9

Sociopathic 87.9 91.2

Total sample 68.6 88.2 154

there is considerable variation in the accuracy with which different

types are identified. Most readily recognized were the sociopathic

patients, about whom clinicians agreed 87.9% of the time. Sur

prisingly, the most difficult discrimination was of the obsessive type,

for which there was only 57.7% agreement. Apparently the features of

this style of relating are less flagrant than others. Clinicians no

ticed them, but considered them secondary. If a secondary rating was accepted in evaluating percent agreement, and such ratings were per mitted in the instructions for the rating scale, then the consistency

in identifying the obsessive style rose to a more robust 73.1%. When asked to judge the applicability of each type for a given patient, the percent agreement rose, as anticipated, into the upper SO's. Such

findings are comparable to what have been reported in other studies of diagnostic consistency (Kreitman, 1961; McGuire, 1973; Zubin, 1967).

In view of the criticisms of diagnostic accuracy, this is, however, not entirely reassuring.

A more rigorous test of reliability is Cohen's Kappa, which corrects for chance agreement (Cohen, 1960). The weighted average

Kappa was calculated by combining the reliabilities obtained in pair wise comparisons. The resulting reliability estimate is .374 for pri mary types, and .473 if secondary types are included. These levels are disappointingly low. Although therapists agreed at better than chance levels, there was little consensus about type. While the

Diagnostic Staffing group provides considerably less information about the patient than would be obtained in the intake interview, the low 155

reliability will weaken the research. Effects attributable to per

sonality types will be diluted because of errors in classification.

Any of the hypothesized effects that do appear, despite the flaws in

the scale, are likely to be quite robust.

Procedures

Experiment 1. The purposes of the first phase of the research were to gather normative data on self-consciousness among psychiatric patients and to assess the validity of the Self-Consciousness scale

(SC) for this population. For this study, all inpatient and outpatient

records were used, providing a total sample of 230 cases. For those veterans arriving at the Mental Hygiene Clinic, the scale was ad ministered before the beginning of therapy. The typical intake pro

cess includes a brief screening by a psychiatrist, usually occurring on the same visit as a diagnostic assessment by clinic therapists.

The patient may be asked to return for a group interview at a staff meeting, if there is some question as to the appropriate treatment

recommendation, or he may be directly assigned to begin therapy. A full range of dynamically-oriented individual, group, and family ther

apy options are offered by staff and students with degrees in psychol ogy, social work, or nursing. Veterans are seldom followed for medica tion only.

Research materials were administered at the earliest point of contact possible. The procedure followed at the outset had the pa

tient complete the Self-Consciousness scale under the supervision of 156

the intake interviewer, which permitted prompt response to the pa

tient's questions and the opportunity for a clinical decision to can

cel a patient's participation should the protocol or accompanying con

sent (see Appendix A) appear disturbing to the subject. The scale was

administered before or after the intake interview proper at the dis

cretion of the therapist. In those instances when a patient's psy

chiatric screening occurred on a different day than the intake inter

view, the scale was handed out by the receptionist and completed while

the patient waited for the psychiatrist. After the first month of

data collection it became clear that participation in the research was

seldom stressful to the veteran but often an imposition on the thera

pist. The procedure was altered so that the scale was always self

administered. The research documents were handed out by the recep

tionist on the patient's arrival and completed in the waiting room

prior to the intake interview. Completed forms were given to the

therapist or to the receptionist.

During the course of this study, patients filled out the SC

scale under any of three conditions: before the session, at the begin

ning of the intake session, or at the end of the interview. An item

on the Intake Rating Scale identified the patient in terms of which procedure was followed. For the 102 outpatients for whom data was available on this variable, there were no significant differences in

total SC or on any subscale as a function of the time the question naire was completed. Similarly, the variation over the 10 months of

the data collection was neither consistent nor significant. 157

The procedure differed for inpatients. Their greater acces sibility permitted group testing under the supervision of the re searcher. The day prior to the research session, head nurses from each of the four inpatient units were asked to identify all patients who were likely to be discharged in the coming week, who would have referrals for outpatient therapy at the Mental Hygiene Clinic, and whose mental status was sufficiently clear that they could tolerate participation in the study. Each identified patient was subsequently asked to participate by the researcher, who assembled the subjects in a quiet room off the unit and administered the measures. Testing groups ranged in size from one to eight veterans. Introductory com ments were minimized so that the principal explanation of the study was found in the consent form, which was identical to the one used with outpatient subjects. During testing, the researcher provided minimal responses to patient's questions or concerns and remained otherwise occupied with reading. For occassional subjects, greater structure was provided. In three instances the items were read and responses recorded by the experimenter. In addition to the Self-Con sciousness scale, inpatients were also administered the Barron Ego

Strength scale. The scales were counterbalanced in order of presenta tion to control for order effects.

The order of presentation did effect SC scores. Table 12 pre sents subscale means as a function of order of presentation and shows a significant impact on Private Self-Consciousness. Apparently, con vening the group and presenting instructions elicited more attentiveness Table 12

Self-Consciousness Scale Scores

when Administered Before or After the Ego Strength Scale

Mean self-consciousness

Order of presentation N Private Public Social Anxiety Total

Self-Consciousness 25.55* Scale first 47 19.81 11.56 56.98

Self-Consciousness 47 22.83* 19.95 12.04 55.26 Scale second

*.E. < . OS, when means are compared. 159 to internal states than did the items from the ES scale. However, the impact on the total score was minor and nonsignificant. It is not felt to invalidate comparisons with the outpatient group or other samples.

Means, variances, and subscale correlations were calculated for the outpatient and inpatient samples separately. A factor analysis reexamined the internal structure of the scale. For inpatients addi tional correlations were obtained to analyze the relationship of Ego

Strength to the Self-Consciousness scale and its subscales.

Experiment 2. Self-consciousness was evaluated as a con tributer to premature termination from therapy. Again, the outpatient sample was supplemented by that portion of the inpatient sample who did, in fact, have post-discharge intakes at the Mental Hygiene Clinic.

Scores on the Self-Consciousness scales were obtained using the pro cedure described in Experiment 1. In addition, the Intake Rating Scale was obtained from the diagnostic interviewer for 116 of the veterans attending an intake session. If the therapist indicated that the case disposition was a referral for treatment elsewhere, the subject was excluded from this phase of the research. At the time of the archival review, 133 cases were examined and a determination was made as to the number of subsequent clinical appointments kept by each participant.

Those who attended six or more sessions were considered to be con tinuers. Terminators were defined as those dropping out of treatment unilaterally or dropped from the clinic rolls for consecutive missed appointments prior to the sixth session.

A subcategory of the terminators included those failing to 160

return after intake or staffing and those inpatients failing to attend

even the scheduled intake. This subsample, occasionally identified as

'decliners' or 'refusers' in the literature, was compared to other drop-

outs in terms of self-consciousness levels. While no differences were

expected, based on the dominant findings in published reports, a planned

comparison evaluated the significance of any differences between these 2 subgroups. A series of X analyses and ANOVAs were completed to eval-

uate the role of self-consciousness in the unilateral termination of

psychotherapy.

Self-consciousness is not the only variable likely related to perseverance in treatment. Variables drawn from the archival review and from the Intake Rating Scale were also evaluated as alternative

explanations of the differences between decliners, other terminators, and continuers. With SC scores, they were included in a discriminant analysis. This procedure makes it possible to evaluate the relative

importance of self-consciousness as a determinant of premature termination.

Experiment 3. The third phase of this research examines self- consciousness as a function of personality type. In this study the data was obtained for 116 outpatient subjects and included their SC scores and the clinician's rating of diagnosis and personality types on the Intake Rating Scale. A series of planned comparisons permitted evaluation of hypothesized profile differences within and across personality types. RESULTS AND DISCUSSION

Results from each experiment will be presented and briefly discussed in the order outlined in the Methods. This section will conclude with a more general discussion of the implications of this research for self-awareness theory and for the clinical handling of self-consciousness.

Self-Consciousness Levels of Psychiatric Patients

Comparison of patient groups. The means, standard deviations, and subscale intercorrelations of the Self-Consciousness scale (SC) were calculated for the patient samples separately, and are presented in Tables 13 through 15, respectively. A series of ANOVAs examined the significance of the differences between means. Because the "Other

Outpatient" sample - those not otherwise evaluated beyond completing the SC questionnaire - are neither hypothetically nor empirically dis crepant from the remaining outpatients, they are incorporated into one group for the purposes of the following inter-sample comparisons. For

Private Self-Consciousness (PR-SC), K(2,226)=4.544, ~( .05, and for the total SC scores, K(2,226)=3.945, ~< .05, the groups varied signifi cantly. For Public Self-Consciousness (PU-SC), K(2,226)=2.025, ~=.134, and for Social Anxiety (SA-SC), K(2,226)=2.165, ~=.117, there is a non-

161 Table 13

Mean Self-Consciousness Scale Scores for Patient Groups

Mean self-consciousness

Groups N Private Public Social anxiety Total

Inpatients 94 24.191 19.883 11.798 56.117

Outpatients 113 24.053 18.920 12.478 55.690

Screenings 11 30.000 21.545 15.545 67.091

Other outpatients 12 24.000 18.833 11.917 54.750

All patients 230 24.391 19.435 12.317 56.361 Table 14

Standard Deviations of Self-Consciousness Scale Scores for Patient Groups

Self-consciousness scale standard deviations

Groups N Private Public Social anxiety Total

Inpatients 94 6.569 4.817 6.035 12.793

Outpatients 113 6.138 5.304 5.177 12.847

Screenings 11 7.707 6.962 7. 725 20.384

Other outpatients 12 4.991 4.428 5.791 11.331

All patients 230 6.429 5.166 5. 723 13.320 Table 15

Intercorrelations of Self-Consciousness Subscale Scores for Patient Groups

Self-consciousness subscale correlations

Private and Private and Public and Groups N public social anxiety social anxiety

Inpatients 94 .458** .168 .343**

Outpatients 113 .415** .263* .524**

Screenings 11 • 718* .759* .749 *

Other outpatients 12 • 728* .025 .326

All patients 230 .469** .264** .455**

*p < .01 **E.< .001 165 significant but similar pattern in which the higher scores of those seen for psychiatric screening stand out against the essentially sim ilar scores of the inpatient and outpatient samples.

The interpretation of these differences rests on an apprecia tion of the hospital procedures. Inpatients were evaluated toward the end of their residence, and outpatients completed the self-rating an average of 14.3 days after being referred for treatment. In contrast, screenings are done on two weekday mornings when Mental Hygiene Clinic psychiatrists evaluate veterans applying for treatment at the Admitting

Clinic, the central intake unit for the Medical Center. Veterans com pleting the questionnaire prior to screening are making their first contact with the system. They are likely in crisis, at the point of maximum discomfort which precipates seeking help. But this is not a total explanation, as it is the private self-awareness rather than the anxiety subscale which is the most elevated relative to comparison groups. Additionally, the self-report device is used as an instrument to further the patients project of getting prompt treatment. This is not the same as faking bad--the scores are more elevated but in the same direction as is typical. But the high SC scores do express a self-preoccupation and heightened attentiveness to internal states preparatory to the patient's first attempt to explain his condition to an expert. What is important in understanding the scale, regardless of whether crisis or the necessity of self-presentation is invoked in explaining the elevated SC scores of screening patients, is that it appears to measure the patients' states in addition to enduring 166 dispositions.

It might be countered that those seen for screening are different dispositionally from other patients coming to the Mental Hygiene

Clinic. But the available data counters that view. Patients coming to the clinic via Admitting are common (16.9% of the current outpatient sample). It is not plausible to assume that those seeking treatment on the two days when screening takes place at the clinic are different from those arriving at admitting on the remaining days. When those seen first at screening are compared with others referred from the

Admitting Clinic, there are no significant differences in age, diagnosis, personality type, prognosis, motivation, distance traveled to the clinic, or history of prior treatment at the Mental Hygiene Clinic.

While a greater proportion of those seen for screening have service connected disabilities (28.6% versus 5.0%), the difference is not 2 statistically significant, X (1)=1.763, £> .10. More to the point, when those originally seen for screening complete the SC questionnaire at the time of their intake interview they do not differ from other

Admitting Clinic referrals in total or subscale scores. It is not dispositional differences that account for the significant elevation of the scores from screening patients. With the obvious rival interpretations rejected, the role of situational factors is more persuasive as an interpretation. The finding for screening patients is reinferred by the previously discussed effects of test order for inpatients. At least for psychiatric patients, the Self-Consciousness scale seems to be sensitive to transient variations in self-focused attention as well 167 as to the intended characterologically based self-awareness.

Self-Consciousness scale scores of patients compared with nor mative samples. Table 16 allows comparison of the SC scores for the combined groups in the present study with those obtained in previous research. The first hypothesis was that self-consciousness would be a more unitary dimension for patients than for undergraduates. This would be evident in higher intercorrelations among subscales. The data supports such a conclusion. Where normative samples find signifi cant correlations between PR-SC and PU-SC, and occasionally between

PU-SC and SA-SC, the present sample finds all three coefficients of sufficient magnitude to be statistically significant. Psychiatric patients tend, more than normals, to link self-focused attention and anxiety. However, this is a matter of degree. The common variance ranges between 6.9% (for PR-SC and SA-SC) and 22.0% (for PR-SC and

PU-SC). The scales remain substantially independent. Practically, then, introspection may be encouraged during the course of therapy without automatically triggering disruptive anxiety. This observation is mediated by the degree of crisis. As Table 15 shows clearly, intercorrelations are highest for those seen for screening. Here PR

SC and SA-SC have a common variance in excess of 57.6%. For inpa tients, this figure drops to 2.8%. Turner's (Note 2) data also shows pathological samples with intercorrelation elevated only modestly above the levels found in college students. In conclusion, patients show a great deal of variation and yet, on the whole, respond to self-focused attention in a more unitary manner than has been found for normal sub- Table 16

q~lf-Consciousn~ss Scale Means, Standard Deviations, and Intercorrel&tions for Patients and Normals

Self-Consciousness scale

Heans Stanclartl Dev:J.atlons Intercorrelations

Ref~rence and P~:ivate Private Public population and anti and studied Social Social public social social Private PubHc anxiety Total Private Public anxiety Total anxiety anxiety

Present res.-arch, 21,. 39 19.4] 12.32 56.36 6.43 5.17 5.72 13.32 .47** .26** .45 '~* LjO pc>t f.ents

-~------

Turner. (Note 2). 24.0 18.8 13.6 56.4 5.3 4.5 5.5 .39** .lJ .10 !, 7 neurollcs

~I psychot icR 22.1 18.5 13.3 53.9 7.4 6.7 4.9 .46** . t 7 .7.5

·-·------·-·----- ·------

Fe>n:lgste.~n, Scheier. & tluss (1975), 25.9 18.9 12.5 57.3 5.0 4.0 4.1 9.2 179 nm.le und~r- graduates

152 undergraduates .26** -.06

Carver & Glass (1976), 25.4 19.5 ll.S 56.5 h.7 4.2 3.8 8.0 .33** -.05 -.02 iO'j tr;>J.c under grathiates

Turner. Scheier, C;,rver, & Ickes 0978), . 31 ** .11• ** .21** 1395 undergra()u~tes

·------1-' 0\ *p (.05 00 **.r <.ot 169

jects.

Contrary to expectation, however, patients do not show greater self-consciousness or self-evaluative anxiety than previous normal samples. Hypothesis 2 stated the SA-SC scores would be higher and SC total scores would vary significantly from those previously reported.

Neither expectation is supported by the data gathered. Subscale and total scores are remarkably consistent with previously obtained re sults from normal subjects. This conclusion is reinforced, in that it replicates Turner's finding that patients generate unremarkable scores. It is further supported by a more microscopic analysis which reveals no significant variance, !(3,114)=1.157, £> .30, among any of the diagnostic groupings (psychotics, neurotics, personality disorders, or others). If prognosis is used as an index of severity of illness, then the absense of any relationship with self-consciousness, !(4,103)

=.103, ~> .90, further suggests that self-consciousness levels do not vary as a function of intensity of psychopathology. Theories that sug gest that psychopathology or behavior in therapy can be broadly ex plained in terms of either excesses of deficits in the skills of self monitoring receive no support from this research.

Again, this observation should not be overgeneralized. Pa tients are a diverse group, and level of self-focused attention may be a valid issue in individual cases, though not in general. Consistent with Hypothesis 3, this data does show greater variability in subscale and total SC scores than has been reported for normal samples. Table

14 shows that the standard deviations are higher than those obtained 170 previously. In comparison with the samples from Fenigstein et al.

(1975) and Carver and Glass (1976), tests for homogeneity of variance for the total SC and subscale scores are always significant beyond

~ =.01. The fact that Turner also observed a comparably increased level of variance, lends further support to the conclusion that dif ferences observed are more than error of estimate. (Tests of homo geneity of variance show nonsignificant differences between the cur rent sample and those surveyed by Turner, with the exception of PR-SC for neurotics and PU-SC for psychotics, both different at the .05 level.) Patients are more variable than normal college students, whether because of greater heterogeneity in age and life position or as a function of their psychopathology. Among the implications of this conclusion is a methodological one, it will be more difficult to obtain statistically reliable differences among means when comparing patient groups. For the practice of psychotherapy, the implication is that each case must be assessed individually to determine whether self consciousness contributes to the clinical picture. While it does not do so for the patient population generally, in individual cases self consciousness and its resulting anxiety may require special attention and special handling during the course of treatment.

To explore the basis of this greater variability among pa tients, the effect on SC scores of a variety of covariates was examined.

As previously mentioned, neither global diagnosis nor prognosis contrib uted to level of SC. Whether or not the veteran has a service connected disability because of his illness, perhaps another indirect measure of 171

the degree of pathology, is unrelated, F(1,130)=1.532, ~> .20. Age is minimally correlated, ~(133)=-.08, ~> .30. Other covariates more pertinent to clinic treatment of the veteran will be discussed in the section of predicting premature termination, but it is clear that none of these variables which describe characteristics of the subject con tribute appreciably to the level or variability of self-consciousness.

In the absense of published materials setting the norms for psychopathological groups, the first goal of this research has been to evaluate hypotheses predicting performance of the SC scale among pa tients. Contrary to expectations, the group as a whole was very sim ilar to normal samples in total and subscale scores. Even the social anxiety scale, the measure of the degree to which scrutiny sets off disruptive self-evaluation, was not significantly different for pa tients than normals. In line with expectations, however, SC scores showed greater variability for the current sample, suggesting that self-consciousness may be of importance in individual cases. This is particularly relevant in light of the confirmation of the hypothesis that SC subscales would be more highly intercorrelated than previously reported. The patient's response to self-directed attention shows less independence between self-monitoring and anxiety than has been re ported for normals. Introspection or observation by another may lead reflexively to self-evaluation.

Factor structure of the Self-Consciousness scale. Because the SC scale is being applied to a new population, it is appropriate to reexamine the internal structure of the scale. To this end, an 172 intercorrelation matrix was generated and a factor analysis completed.

The average item correlates with the unadjusted total score between

.4 and .6. All are positively related, as expected. However, four of the 23 items have a correlation coefficient value less than .25. These items are numbers 3, 9, 12, and 22 (see Appendix A). The first three of these are the items scored in a reverse direction (e.g., 3: "Gen erally I'm not very aware of myself."). The attempt to create a scale balanced for the acquiescence response set has failed for this sample.

Unlike the normative sample, in which no item was endorsed in one direction by more than 85% of the group, three items were highly skewed.

Items 6 (13.5%), 13 (14.4%), and 21 (8.3%) were almost never rejected

(as being either "extremely unlike" or "somewhat unlike me") by pa tients. While the proportion of variance attributable to the response set is uncertain, it clearly constitutes a threat to the capacity to make sensitive distinctions and represents a major flaw in the scale.

Future research should assess the degree to which the SC scale is con founded with acquiescence.

A principal factor analysis was done, with Kaiser normaliza tion and varimax rotation, using the standard SPSS algorhythm (Nie,

Hull, Jenkins, Skinbrenner, & Bent, 1975). The communality was re stricted on some items, most notably on numbers 2 and 17, from the PU

SC subscale. Nevertheless, seven factors were selected which had eigenvalues greater than one and which cumulatively accounted for

58.1% of the variance. Of these, the two most important accounted for

31.7% of the variance initially and retained eigenvalues in excess of 173 one after iteration. While the entire seven factor matrix is pre sented in Appendix C, the following comments will concentrate on the two most important, which are duplicated in Table 17.

The factor structure obtained from patients fails to replicate that reported by Fenigstein et al. (1975). The expected three factors, corresponding to the subscales, are not evident. The SA-SC subscale is essentially reincarnated, in more robust form, as Factor I. The five most highly loaded items are from SA-SC. The only subscale item missing is 12, one of the items scored in the negative direction. Of lesser importance, but still significantly correlated at the .001 level is one item from PR-SC and two from PU-SC. With operative words such as "worry" and "concerned", they reinforce the interpretation of Factor

I as tapping a dimension related to anxiety and a disruptive fear of judgment. Although Factor II pulls two items from SA-SC and four from

PU-SC, PR-SC is the most frequently (five items) and most importantly represented subscale. The PR-SC items included emphasize congitive rather than affective self-awareness. Factor II may be described as a generalized introspective style in combination with a concern about personal motives and impact on others. The elements of anxiety are subdued, as are most references to emotionality. The factor emphasizes intellectualization and cognitive controls.

The PR-SC and PU-SC subscales, respectively, accounting for the greatest variance in previous reports, collapse into a secondary factor for this sample. They appear as essentially independent dimensions only among the minor factors. In Factor III, five of seven items Table 17

Communalities and Factor Matrix for the Self-Cansciousness Scale

ltem loadings

Suuscale Variable Communality Factor Factor 2 -·------PR-SC f 'm always tryir,g to figure myself out .627 '125 • 709* PU-SC I'm concerned about my styl'! of doing things .192 • 171 .353* PR--:JC Generally I'm not very aware of myself .406 -.113 -.027 SA-SS lt takes me ti•ne to overcome my shyness in new situations .852 .643* • 234 * f'R-SC~ 1 reflect (or think carefully) about myself -'1 lot .438 .186 . )45* PU-·SC I'm concerned about the way I present myself (cr come across to others) . 284 .112 .ld(l* I'R-·3C I'm often the subject: {or center) of my own fantasies • 334 .268* • 275* SJ\-SC I have trouble working when someone ls watching me .464 .606* • 16'• I'R-SC I never scrutinize myself (ol: examine or look at my8elf closely) .290 .060 .OWJ SI\-SC r ~;et cmharrasse.l very ensi!.y ·'•88 .526* .n1* l'li-·SC 1'111 self-conscl.ous about the way T look .528 • I 53 . 121 Si\-SC l don't find lt hnrd to talk to strange..--s .259 .100 .027 PP.<'C 1 'm generally attentive to (or aware of) my inner feelJ.ngs .321 .104 • 11, l Pl!-SC I. usually worry about making a good impressicr. .538 .303* . 325* l'R-<>C I'm constantly examJninr, my motives (or looking at my reasons for acting as I do) .451 . 151 • 591* SJ\-SC I feel anxious when I speak in front of a grou~ .218 .391* . 182 ru-se One of the last things I do before I leave my house is look in the mirror • 11•1 .139 .lSR I'R-SC 1 sometimes have the feeling that I'm off somewhere watching myself .413 . 1.85 .25)* l'U-SC I '•• con.:erned about what other p?.ot'le t!>ink of me .429 .3il* . J38* f"{-sr I'm alert to changes in MY mood .400 .025 • O!,z PU-SC I'm usually awnre of my appearan<:e .343 -.099 -.026 PR-SC r 'm aware of the way my mind works ~hen I work through a problem • 35'• -.095 • Ol, t Si\-SC Large groups mnke me nerv0us .63) .7Jv* . 116

l~_t_c. The Self-Consciousness subscales are Abbrevinted PR-SC for private, PU-SC for public, and SA-SC for social anxiety. *~" .001 175

loading, and the most highly correlated items among them~are from PU

SC. Likewise, Factor IV is dominated by a single subscale, in this case Private Self-Consciousness. But these are of lesser importance.

The important finding of this analysis is that SA-SC emerges essentially intact as the dominant feature of self-consciousness among a psycho pathological population. The importance of this subscale is to be highlighted throughout this research. Moreover, its emergence as the central factor underscores the difference between patients and stu dents: anxiety is the central feature of self-consciousness for those seeking psychotherapy.

Self-consciousness and ego strength. For the 94 inpatients completing the Ego Strength Scale (ES), the mean was 39.71 and the standard deviation 7.91. Table 18 shows the correlation of ES with the subscale and total scores from the Self-Consciousness scale.

The generalizability of previous findings about the SC scale would be enhanced if SC was minimally related to degree of psychopathology.

If the scale performed the same with schizophrenics as with normal students, one could be more confident in predicting avoidance of added stimuli to self-scrutiny in the clinical population. But, as already encountered with the factor analysis, it is clear that the SC scale is not treated in the same manner by patients as by normals. The signifi cant correlations with ES confirm this finding. Those who are most severely disturbed (low ES) are some 17 points higher in self-conscious ness than those with the mildest pathology. Alternatively, those with the highest SC scores admit to greater anxiety and symptomatology. 176

Table 18

Correlations for the Ego Strength Scale

with the Self-Consciousness Scale

Self-consciousness scales

Private Public Social anxiety Total

Correlation -.109 -.169 -.569 -.390

Probability .149 .052 .001 .001

Note. N=94. 177

Although each variable retains considerable independence from the other, with a common variance of 15.2%, the ES is a major contributor to total self-consciousness. This is contrary to the expectation ex pressed in Hypothesis 4.

Because the two scales are more related than expected overall, the subscales also are correlated with ES in a manner other than hypothesized. Private self-consciousness was expected to have a direct relationship with ES, on the assumption that it embodied the cognitive skills of introspection that would be more intact in the less disturbed patient. Instead, Table 18 shows a nonsignificant, inverse relation ship. Public self-consciousness, for which no correlation was pre dicted, proves to be negatively correlated to a marginally significant degree. It is apparent that the generalized introspective style

(Factor II, above) is, if anything, more typical of the most patho logical rather than the most intact elements of this sample. The social anxiety subscale relates to ES in the direction hypothesized, but to a much greater degree than anticipated. Almost a third of the variance of SA-SC (32.4%) can be traced to the psychopathology dimen sion captured in the ES scale. While PR and PU-SC may share minimal linear relationship with ES, it is clear that social anxiety--which proves to be the central dimension of self-consciousness for patients- takes on different meaning and increasing importance as the level of

ES drops. Table 19 summarizes the intercorrelation of SC subscales at four different levels of ES. It shows that the association of anxiety and public self-consciousness increases as ego resilience decreases. Table 19

Intercorrelations of the Self-Consciousness Subscales for Patients

at Different Levels on the Ego Strength Scale

Intercorrelations of self-consciousness subscales

Level of the Ego Strength Scale N Private with Private with Public with public social anxiety social anxiety

More than one SD below the mean 13 0 726** .359 .695**

Less than one SD below the mean 28 .170 -.049 .363*

Less than one SD above the mean 36 .479** .078 .284*

More than one SD above the mean 17 .651** .319 .188

*.E. <. 05 **.E.< .01 179

Other relationships are less clear. But, like those seen for screen

ings, the most disturbed of the participants show the least indepen

dence among the component subscales. Anxiety pervades self-awareness.

To summarize the findings up to this point:

1) Consistent with Hypothesis 1, the Self-Consciousness sub

scales are less distinct and independent for psychopathological samples

than for normal students. This is particularly true for those pa

tients--those attending screening appointments in this study--who are

exposed to situations which serve as special stimuli to self-awareness.

2) Contrary to Hypothesis 2, the mean levels of self-con

sciousness are quite consistent between therapy patients and normals.

Even the anxiety scale scores were not elevated in this sample. How

ever, the SA-SC subscale does take on special importance in a clinical

population since it lies at the core of the dominant factor of the

SC scale. The PR-SC and PU-SC subscales are neither so important nor

so independent as reported for the normative sample.

3) Patients are moderately more variable in their SC scores

than are normals, as stated in Hypothesis 3. For this reason, self

consciousness levels of an individual patient may comprise an impor

tant treatment issue, even though patients as a group show self

consciousness levels no greater or lower than normals.

4) Contrary to Hypothesis 4, ES and SC are importantly re

lated. Self-consciousness takes on a different meaning for subgroups within the patient population. Not only is SA-SC an increasingly

important dimension, but self-evaluation levels increase--as does 180

general self-consciousness--among the more seriously disturbed.

Hypothesis 5 is supported in part: SA-SC and ES are significantly

and inversely related. Contrary to expectation, and consistent with

its diminished role among patients, PR-SC is neither positively nor significantly correlated with the Ego Strength Scale.

Self-Consciousness and Premature Termination from Psychotherapy

A central hypothesis of this dissertation is that self-con sciousness levels can be used to predict premature termination from psychotherapy. This expectation is derived by analogy from self awareness theory, where research shows subjects experiencing cognitive dissonance avoid stimuli to objective self-awareness. It is also based on an analysis of causes of premature termination, which concluded that an absense of self-monitoring skills (i.e., insufficient self consciousness) leads to dropping out because of a lack of motivation, while excessive anxiety and self-consciousness results from an in sufficiently structured therapeutic relationship and also results in unilateral and early termination. Thus, an inverted U relationship was expected between self-consciousness (SC) and the percentage of patients dropping out of psychotherapy (Broen & Storms, 1961; Wine,

1971). The following discussion will present data by which to evaluate the central hypothesis.

The following section will be structured in four separate, but related sections. First will be a discussion of the direct relationship of SC scores to the number of sessions attended. Second will be an 181 examination of potential indirect effects in which other variables relating to length of stay are examined for their association with self-consciousness. Third, a discriminant analysis will assess the importance of SC scores relative to other variables in the prediction of premature termination. Fourth, data on other variables suggested in the literature as being related to perseverance in treatment will be reviewed.

Self-consciousness and continuation in treatment. Clinic rec ords of subjects completing the SC scale were reviewed and the number of sessions recorded. All veterans who attended six or more sessions were defined as continuers. Those who never began treatment comprised another comparison group, referred to in this report as treatment decliners. Table 20 presents the mean levels of SC for patients falling into the respective groups by length of stay. Since the literature is inconsistent in discriminating decliners from terminators, the groups are presented separately and combined. Because the number of termina tors is rather small, the data is pooled as well as presented on a ses sion by session basis. Yet, in none of these groupings is there any deviation of note from the grand mean. The relevant ANOVAs (see Table

21) are all nonsignificant. Even when SC scores are examined in terms of the sample specific factor structure, there is no evidence of rela tionship with perseverance in treatment; Factor 1: !(2,137)=.146, £>

.80, Factor 2: !(2,137)=.045, £> .90. The absense of differences in means suggests the absense of gross differences. But it only partially responds to the hypothesis. The prediction was that terminators would Table 20

Self-Consciousness Levels of Decliners, Terminators, and Continuers

Mean self-consciousness scores

Group Sessions N Private Public Social anxiety Total

Decliners 0 53 24.54 19.27 12.36 56.30

Terminators 1 9 22.55 20.33 12.78 55.66

2 3 24.00 22.00 13.00 59.00

3 6 30.00 21.83 13.83 65.66

4 2 20.50 17.50 12.50 50.50

5 1 22.00 9.00 12.00 43.00

Terminators 1,2,3,4,5 21 24.66 20.19 13.05 57.90

Decliners and 0-5 74 24.58 19.53 12.55 56.75 terminators

1-' Continuers 6 or more 59 24.67 19.51 12.39 57.27 00 N Table 21

ANOVAs for Self-Consciousness Levels of Decliners, Terminators, and Continuers

! (and associated z) for Self-Consciousness Scale scores

Contrasts df Private Public Social anxiety Total

0,1,2,3,4,5,6+ 6,126 1.046(.399)a 1.239(.291) .076(.998) .840(.542)

1,2,3,4,5,6+ 5,74 1.314(.267) 1. J88 (. 239) • 078 (. 995) .952(.453)

0-5 vs. 6+ 1,131 .008 (. 931) .000(.983) .030(.863) .055(.814)

0 vs. 1-5 vs. 6+ 2,130 .006(.994) .263(.769) .134 (. 87 5) .150(.861)

1-5 vs. 6+ 1,81 .008 (. 929) .087(.769) .002(.960) .017(.898)

aThe format of all entries is I(~). 184 include veterans with extremely low as well as extremely high SC scores. These subgroups could well cancel one another out in a statistic which pools their means.

A Chi-square analysis provides the appropriate test of the distribution of terminators across levels of SC. Each SC score was recoded into intervals of one standard deviation. The expectation, then, is that terminators would be overrepresented in the extreme intervals. In Table 22, a summary of the distribution on total SC, it is apparent that those terminating after three sessions roughly conform to the hypothesis. However, none of those staying other lengths of time do. There is no statistically significant difference between decliners, terminators, and continuers in distribution across SC, 2 X (18)=16.65, £>.SO. Nor are differences significant when length of 2 stay is plotted against distribution on PR-SC, x (18)=27.60, p > .05, 2 2 PU-SC, X (18)=22.64, £>.20, or SA-SC, X (18)=15.09, £>.60. Termina- tors are not different than continuers or decliners in their distribution on SC scores. When decliners are excluded, the relationship between SC and perseverance in treatment still fails to approach significance. Moreover, none of the groups varies much from a normal distribution. Terminators have no more extreme scorers and continuers are not concentrated near the mean. Figure 6 presents the percentage of people remaining in treatment at each level of total SC. There is no evidence of the hypothesized inverted U relationship. Quite clearly there is no direct association between the degree of self-consciousness and perseverance in treatment. Hypotheses 1 through 4 are not supported. Table 22

Total Self-C

----·------·-----~ Number of sessions attended

Dec liners Terminators Continuers ·--·--- Total score on the rcrcent in ~elf-Consciousness scale 0 2 3 r, 5 6 o1.· more each interval ------Hore than one SD 15.8% 10.0% 0.0% 28.6% 0.0% 50.0% 11.9% 14.3 below the mean

Lc>io!S thau one SD 33.37. 40.0% )3.37. 0.0% 100.0% 0.0% 39.0% 35.0 below the mean

Less than one SD 36.8% qQ.O% 66.7% 28.6% 0.0% 50.07. 33.9% 35.7 above the mean

Hore than rme SD 14. o~; 10.0:'( 0.0% 42.n. 0.0% 0.0% 15.3% i 5.0 abovf~~ the mean

Total N 57 10 3 7 z 2 59

Note. Percents arc tabulated vertically, that is, within session. 0 ..0 1-l ! I I i w0 ' I

! 71.1\ - /Woraqc -~~ g ---t---1---1---+- g L_ Percent ... Continuing .;J 0:::0 0 11) v to "c:o ....'1)11' .j.)

0 ·' o·---J Intcrv>~l 47 52 57 62 67 77 82 87 92 9"/ 202 107 tvteans Cases per Interval 1 l 3 4 8 1.4 10 5 18 7 7 4 l

Percent Contir,uc.r!3 Among Patients in St1ccessive IntcrvnJ.s on Total-SC. 187

Indirect effects of Self-Consciousness scores on premature ter-

mination. Even in the absense of a direct effect, SC scores may in-

fluence continuation in therapy by virtue of their common association

with other variables. Clinic records and the Intake Rating Scale pro-

vided information on a number of potential covariants of length of

stay. However, none of these variables successfully discriminated con-

tinuers from terminators. Client variables not significantly related 2 included: age, !_(2,131)=1.191, .e_> .30; presence of disability, x (2) 2 =.22, E.> .80; and rated prognosis, x (8)=12.99, E.> .10. Source of

referral was used as an index of the presenting problem since it re-

fleets the patients attempt to identify the nature of his problem as 2 physical or psychological. It was unrelated, X (16)=14.37, .E_> .50.

Prior perseverance in treatment at the clinic was employed as a rough

measure of the extent to which the patient was therapy-wise, that is,

skilled in the execution of the patient's role. It failed to dis- 2 criminate, X (6)=10.13, .E.> .10. One variable, motivation for treat- 2 ment, did discriminate those who failed to begin therapy, x (6)=15.05,

E.~ • 05. Those rated as poorly motivated dropped out without a single

session (72.2% versus 26.4% for the sample as a whole) while those

identified as excellently motivated usually continued for at least six

sessions (66.7% compared to 52.9% for the entire sample). But, even

motivation failed to discriminate between terminators and continuers, 2 X (3) = 1. 4 7 , E. ). • 50 •

Therapists do play a role in length of treatment. A comparison

of the seven therapists treating more than five research subjects each 188

2 (59 patients in all) showed a significant therapist effect, x (6)=

13.48, ~ ~.05. But this could not be attributed to student or staff 2 2 status, x (2)=.10, ~}.90; discipline, x (3)=2.10, ~).50; years of 2 clinical experience, x (1)=.001, p > .90; or even to the patients reac- 2 tion to their recommendations for treatment, X (4)=.76, ~}.90. Con-

sistency in treatment was unrelated, whether measured as the same ther 2 apist throughout treatment, x (1)=1.60, ~>.20, or as the same thera 2 pist who conducted the intake interview, x (1)=1.28, ~> .20. It made

no difference whether or not the actual treatment provided matched that 2 initially recommended by the therapist, X (2)=2. 62, .E_) .10, or even 2 what modality of treatment was involved, x (6)=3.35, .E_).70. Even the

number of visits to the clinic prior to the inception of treatment

failed to discriminate continuers, terminators, and those who failed 2 to ever begin treatment, x (4)=1.52, E._> .80. It will be recalled that

inpatients and outpatients varied in time between referral and either

intake or beginning of treatment. Yet, these variables also were not

demonstrably related to length of stay: for intake, f(l,80)=.000,

E._) .90, and for time before beginning therapy, f(1,80)=.381, ~>.50.

The practical consideration of distance of the clinic from the patient's

residence proved nonsignificant, f(1,80)=.328, E._} .50.

None of the readily available predictors of patient, therapist, or contextual factors in premature termination succeeded in discriminating between groups in a meaningful way. The differences in percentage retained among different therapists indicates that there is some variance to be accounted for. It remains a challenge for more finely 189

focused research. For the purposes of this project, it is sufficient

to note that it is futile to look for an indirect path by which SC

scores effect length of treatment when a direct association cannot be

found between the target and the many other variables hypothetically associated with it. For the record, Table 23 reports the association of self-consciousness to the purported covariates of perseverance in

therapy.

Discriminant analyses. A discriminant analysis optimizes what

ever power each of a set of variables has to distinguish criterion

groups. Despite the absense of significant univariate relationships,

the following set of variables was used to generate discriminant func

tions: days between referral and intake appointment (intake); rated prognosis; rated motivation; the patient's reactions to the recommended treatment, on a scale from 1 (negative) to 3 (positive); patient age; degree of disability; distance of residence from the clinic; and PR

SC, PU-SC, SA-SC, SA-SC or, in a separate analysis, total SC.

When the analysis included those declining treatment, this group was the one most effectively distinguished. Using capacity to minimize Wilk's Lambda (and hence to maximize the MANOVA between groups) as inclusion criterion, prognosis, distance, intake, and SA-SC were selected. Of the two discriminant functions the larger accounts for 95.41% of the explained variance, and separates decliners from terminators and continuers to a statistically significant degree, 2 Lambda=.691, x (8)=25.358, £=.014. The second function seeks to iden tify terminators but is less powerful and nonsignificant, Lambda=.980, Table 23

Association of Self-Consciousness Scores with Covariatcs of Premature Termination

Self-Consciousness scales

Variable Private Public Social anxiety Total

Patient variables ------2 Global dingnosis x2(l.2)=9.44,£.) .60 x (12)=17 .01 ,_p_ > .10 x2 (12)=10. 77 •P.. >.so x2 (l2)=l3.40,£>. JO Personality type x2(12)=7 .zo.£> .so x2 (l2)=18.02,p_ > .10 x2(12)=22. 9l,p_ < .os x2(12)=18.67,E_< .to Prognosis f_(4,l03)=.842,r_> .50 !_(l•' 103) =. 758,£_.,.. 50 !:_(4,103)=. 721,E_) .so £(1;,103}=.103,2_>.90 2 2 Percent disability x2 (3)"'. 79,p >.au x2(3),.9.36,_p_< .os x (3)=4.73,p_'>.lO x (3)=2.2l,_p.> .so 2 2 2 Referral source x (9)=6. 32 •£ >. 70 x2{9)=8.tt8,_p_> .40 x (9)=8.14,_p__?.50 x (9)=3. 70,_p_ > .90 Prior therapy !_(3, 130)=.488,£_ ') .60 _!<:(3,130)=.963,_£ ">.40 x_(3,130)=1.560,E_ > .20 £{3,130)=1 .021,£_> .10 ------Therapist and interaction variables

Experience x2(3)=3.24,_p_/ .30 x2 (3)=3.99,£).21) x2 (3)=4.42,£"> .20 x2(3)=7. 73,£ < .10 2 Discipline X (9)=t2.06,_p_'>.20 x2 (9)=4.03,_p_ ).90 x2{9)=7.13,£).60 X2 ( 9) =6. 7 1, .2.) . 60 2 Student or staff status x2(3)=2.94,_p_> .40 x o>=3.20,_p_ >.3o x2 (3) =5. 68,p_ >. 10 x2(3)=6.82,r_ <.10 Hotivation f_(3,104)=1.32,_£) .20 _!'_(3, 104)=1.69,£_) .10 !_(3, 104)=.151 ,_2. >. 90 _!'_(3, 10'•)=1. 34 ,P._). :!0 Patte,tt reaction to recommended !_(2,114)=.438,£_> .60 _!'_(2,114)=.388,p_ >.60 f{3,114)=.838,E_) .40 !_(3, 114)=.027 ·E.). 90 treatment

Clinic variables ·---·------2 Actual t hernpy modality deli.vered x (12)=l3. 27 ·E.). 30 . x2oz)=ll.20,_p__>.so x2(12)=11.58,E_ "> .40 x2(lz)~lf•• 6t,_r:'> .20 2 2 SC~ml' lhernplst throur,hout tre.,tment x2cn~r.97 .£.>.so x (3)=4.37,p_).20 XL ( 3) = 2 • 80 , £_) ·'• 0 x (J)='•.6!l,p) .10 Sessi•ms before the start x 2 (6)~s. 7l.p) .40 x2(6)=4.18,p_> .60 x2(6)=9.63,_r: > .10 x2(6)~5.6l,p ),t,o <>[ tre

2 X (3)=1. 36 7, .E.> .10. In combination, the two equations achieve 70% accuracy in identifying decliners (who have the higher scores based on the first set discriminant function coefficients replicated in

Table 24). But terminators match their predicted group only 36.8% of the time, and continuers are correctly placed 36.2% of the time.

The overall accuracy is 44.19%, less than 59.6% that would have resulted from predicting that all cases are continuers. If total SC is used instead of subscale scores, it is not included in the discriminant function. The accuracy of prediction remains best for decliners and an inadequate 48.84% overall.

When the discrimination was exclusively between terminators and continuers, two additional variables were added: the experience of the therapist and the number of days between referral and the beginning of treatment. The resulting analysis included the variables presented in Table 25. Terminators had a group mean of -.69007, while continuers averaged .24772. This difference is not significant, Lambda 2 =.849, x (5)=7.93, .E_> .10, and the resulting function is 61.84% accurate. With 55 continuers of 78 cases, a uniform prediction of continuation would have been more effective, achieving 70.5% accuracy.

Table 25 also shows the variables selected when total SC replaces subscales in the analysis. Total SC is included in the discriminant function. But once again the function is nonsignificant statistically, 2 Lambda=.83612, X (5)=8.860, .E_>.10, and unreliable practically, classifying cases with only 66.3% accuracy.

As indicated in Hypothesis 7, self-consciousness is included 192

Table 24

Standardized Discriminant Function Coefficients for

Continuers, Terminators, and Decliners

Variable Coefficients with function 1 Coefficients with function 2

Intake .580 -.161

Distance .708 -.196

Prognosis -.658 -.737

Social anxiety .384 -.653 Table 25

Standardized Discriminant Function Coefficients for Continuers and Terminators

With Self-Consciousness subscales With Total Self-Consciousness

Variable Coefficient Variable Coefficient

Intake -.554 Intake -.846

Begin therapy -.518 Pretherapy sessions -.599

Age .682 Age .794

Distance -.605 Distance -.577

Social anxiety -.617 Total self- -.415 consciousness 194 in the discriminant function for three of the four analyses. It does contribute in the prediction of continuation in treatment. But the contribution is in the context of a failing cause. The set of vari ables on which information has been collected are inadequate to re liably classify cases. This is surprising, in that at least some of the variables have been shown to be related to premature termination, albeit in the inconsistent fashion that typifies research in the area.

Patient age and prognosis, and therapist experience have been signifi cantly associated in previous work. A number of variables included here related to the context of treatment, clinic policy and procedures.

As such they tap a less researched area. Their importance is less clearly established, although the recurrent presence in the discrimi nant functions of variables such as days before intake, days before beginning therapy, number of visits before treatment, and distance between residence and clinic makes it clear that that context is as worthy of future research as are the more traditional patient, thera pist, and interaction variables.

In the absense of a pattern of significant findings, it is difficult to interpret the relative importance of self-consciousness or to explain the absense of the hypothesized link between self-evaluative anxiety and flight from therapy. A clear possibility is that the hy pothesis is wrong, that anxiety is not a determining variable in length of stay, or, more specifically, that high levels of self-evaluative anxiety may increase reliance on the therapist for support and coun teract a tendency to avoid therapy. People enter treatment because 195

they are in pain in their daily life. They may be quite willing to

tolerate whatever additional discomfort arises during the course of

therapy because of sustained self-scrutiny. But the obtained results

do not lead to this conclusion, they just fail to rule it out.

Two alternative explanations can be briefly suggested. If

self-consciousness is a determining variable in perseverance in treat ment, it is the operative degree of affect aroused that is critical, not just the level of self-focused attention. In measuring degree of

self-consciousness, this research is tapping a precursor of the end state of interest. Level of self-consciousness may be a necessary but not a sufficient condition for disruptive anxiety. Indeed, the factor structure which shows SA-SC to be separate from the other subscales suggests this. The impact of level of self-consciousness is mediated by a related but untapped variable: skill in handling self-focused attention. Highly self-conscious people may have related abilities to laugh at themselves, to maintain a meditative distance between the ob served and observing self, or to react with permissive acceptance to their own qualities. Any of these abilities, among many others, may permit high levels of self-observation without undesirable sequellae.

Or, as Wicklund (1975) suggests, the person may like significant features of the self observed. The remarkable human capacity to re interpret events and their meaning is central to our extensive efforts to preserve self-esteem. Self-evaluation need not lead to self devaluation. While patients may be generally less happy than others, they nonetheless are motivated to justify and preserve the value of 196 their character (Rogers, 1961; Sullivan, 1954). Their skills in managing self-consciousness may be the critical element in determining their capacity to adapt to the requirements of verbal therapy, and it is an index of these skills that is missing from the present research.

A second, and somewhat related, explanation of the nonsignificant results also is available. It is not only the patient's skills at managing self-consciousness that matters. Therapists also vary in the degree to which they consciously or unconsciously control the tone and intensity of the patient's self-exploration. A highly self-con scious patient, lacking skills to handle that self-scrutiny, might still do well if paired with a therapist whose permissiveness, warmth, patience, or ironic perspective on human failings helped to insulate the patient from excessive self-evaluation. As stated in the review, the thrust of the research on premature termination has shifted to interaction variables. To examine self-consciousness as a quality of the patient alone is somewhat anachronistic. It is justified in this first attempt to examine the impact of the variable. But a more sophis ticated investigation, with a higher probability of significant re sults, would consider the interaction of patient, therapist, and con text in measuring the operative level of self-consciousness.

Other hypotheses about premature termination. In line with the skills approach to self-consciousness described above, it was hypothesized that previous exposure to psychotherapy would provide the patient with an education about handling self-exploration in psycho- 197

therapy, and so reduce the rate of premature termination. Of the 134

cases available, 92 had no previous referrals, 22 had failed to begin

treatment on all previous contacts with the Mental Hygiene Clinic, 18

had previously completed a course of six or more sessions, and only

two had dropped out during the early sessions. So, the sample is too

small for comparing former terminators and continuers specifically.

The only apparent variations from expected frequencies are the con-

tinuation of previous patterns among those failing prior referrals

(68.2% again failed, as opposed to a sample mean of 40.3%) and those continuing (55.6% remained after 6 sessions, compared to 41.8% samplewide). But these differences did not translate into significant Chi-

square valued. When the 24 cases terminating between sessions one and five were contrasted with those continuing, no difference in prior 2 experience was evident, x (3)=1.28, .E_).70. The best test of Hypothe- sis 6 sets two levels of experience by pooling former terminators with continuers, and decliners with those never previously referred. But again, there is no evident relationship with the length of the current 2 course of therapy, X (2)=1.71, .E_).30. The one test that might intimate the importance of self-consciousness management skills, or at least of previous therapy experience at the clinic, fails to find any differences in the current course as a function of previous treatment experience.

Hypothesis 8 specifies that therapist skills and experience will be related to length of treatment. As mentioned previously, there is an effect attributable to therapists, since this does affect 198 the number of sessions attended. But the available indices of skill- fulness are not significantly associated with perseverance in treat 2 ment. Neither therapist discipline, x (3)=2.09, ~>.50, student versus 2 2 staff status, x (2)=.104, £).90, nor experience, x (16)=16.03, £ >

.40, appeared to contribute. No support was provided for the potential explanation of self-consciousness management in the interaction of patient and therapist.

An additional expectation (Hypothesis 5) was that the effects of public self-consciousness, specifically, would be more pronounced among those receiving group therapy. It is in this modality that at tention to the social aspects of this self is maximized. Of the 21 patients referred for group treatment, 4 failed to begin, 5 dropped out, and 12 remained through the initial 6 sessions. The two cases quitting after the first session were 4.52 higher than the sample mean of 18.48 on the PU-SC subscale. But the other terminators were below average. The ANOVA was nonsignificant overall, F(4,16)=.524, ~> .70.

There is no evidence of a special linkage between PU-SC and perseverance in treatment of those assigned to group therapy.

To recapitulate, there was little support for any of the hypotheses linking premature termination to self-consciousness. Terminators were not overrepresented among those highest or lowest in SC scores.

The expected inverted U relationship between percent remaining and location in the SC distribution failed to appear. Nor did the inclu sion of decliners as an additional comparison group clarify the rela tionship. No differences in means were observed. While decliners 199 proved easiest of the three groups to discriminate, and SC scores contributed to that discriminant function, little accuracy was achieved in classifying patients as decliners, terminators, or discriminators.

None of the variables examined as potential covariates of premature ter mination showed significant differences between groups. Similarly, there was no support for the special relationships hypothesized between patient's previous clinic contact and perseverance, therapist disci pline or experience and sessions attended, or PU-SC levels and the behavior of group therapy participants.

Personality Type and Self-Consciousness Levels

The analysis of five personality styles (see pp. 125-133) led to predictions of different self-consciousness levels and subscale pat terns for each type. This chapter will review findings which relate scores from the Self-Consciousness scale (SC) to ratings of typology obtained from the intake interviewer. Methodological issues arising from both scales must, however, be mentioned first.

Methodological issues. As discussed at the beginning of this section, psychiatric patients failed to respond to the logical nega tive included in three items from the SC scale intended to counter balance the acquiescent response set. As such the items had reduced means and low correlations with the total scale score. These items are not evenly distributed across the subscales; two are from PR-SC and one from SA-SC. These scales have correspondingly diminished means for the average item. This presents a problem for evaluating the hy- 200

pothesis that certain types will have differential elevation of sub

scale averages. Additional statistical manipulation was required to

correct this problem. Each subscale average was adjusted (for PR-SC,

PU-SC, and SA-SC, respectively by .0122, -.3324, and .3101) to equate

it with the grand mean for all SC items over the 116 cases involved.

Despite this adjustment based on pooled results, subscale means for each personality type remain free to vary. Within type comparisons of subscale averages become meaningful since no scale retains a built in elevation samplewide.

The second methodological issue pertains to the reliability of clinicians' use of the typology scale. With K=.374, for the reliabil ity study, there is a likelihood that many differences will be obscured by error variance. This will be a particular problem for obsessive and passive types, who proved to be the most difficult to classify.

In reality, discrepancies will be more robust than indicated by dif ferences in means or by significance tests.

Self-concsiousness levels for five global personality types.

At the conclusion of the review, it was argued that types are inde pendent of degree of pathology. They are not, however, independent of global diagnosis, which must be considered a flawed index of inten sity of disturbance. Table 26 presents the data relating type to diagnosis. A Chi-Square analysis shows the two variables to be signif 2 icantly related, x (16)=99.06, E< .001. Accordingly, SC levels will be evaluated separately for each variable. Table 27 summarizes the scale and subscale means used in these analyses. None of the com- Table 26

Global Diagnosis and Personality Type

Personality type Diagnosis N

Psychotic Neurotic Personality OBS Other disorder

Hysteric 3 6 4 1 9 23

Paranoid 35 3 3 0 5 46

Obsessive 10 30 3 3 17 63

Passive 16 20 6 0 8 50

Sociopath 3 3 10 0 1 17

N 67 62 26 4 40 199

N 0 t-' 202

Table 27

Self-Consciousness Scale Means (with ANOVAs)

as a Function of Diagnosis and Personality Type

Self-Consciousness subscales

Diagnosis N Private Public Social Anxiety Total

Global diagnosis

Psychotic 38 23.74 18.81 12.42 55.50

Neurotic 35 24.43 18.14 12.14 54.66

Personality 18 23.56 21.94 14.66 60.17 disorder

OBS and other 27 22.93 18.59 11.77 53.30

E_(3,114) .322,.E_) .80 2.325,£. < .10 1.27l,.E_> .20 1.157,.E_>.30

Personality type

Hysteric 11 23.28 17.91 10.73 51.91

Paranoid 25 24.32 17.84 11.96 54.92

Obsessive 44 24.39 19.38 11.64 55.36

Passive 23 22.96 18.87 14.79 56.61

Sociopath 14 24.65 22.21 14.22 61.72

E_(4' 112) .303,.E_>.80 1.892,.E_>.10 2.275,.E_(.10 1.104,.E_>.30 203 parisons are statistically significant. When public self-conscious ness (PU-SC) is examined as a function of diagnosis, the higher scores of those with personality disorders result in a marginally signifi cant difference. While this is consistent with the extraverted pro file from Miller and Magaro's (1977) findings and is in the direc tion hypothesized, in the context of multiple tests, the probability of chance variation of this degree is too high to regard this as validating expectations. Similarly, the findings on SC as a function of personality type fall short of statistically reliable differences.

The elevated Social Anxiety (SA-SC) scales for the passive, but not sociopathic, patients was anticipated from the literature review.

But the differences among the five groups is insufficient to be clinically or statistically meaningful. Because somewhat different factors were found for this psychopathological sample than for the norm group, ANOVAS were also computed for the two major factors as a function of personality type. Although the sociopathic and passive types were above the mean on the anxiety factor, it proved non significant, !_(4,112)=1.725, .£.> .10. On the introspection factor, pas sive types were the on:Ly divergent group, obtaining lower scores.

But, again, the total variability was nonsignificant, !_(4,112)=2.007,

.£. > .05.

Instructions did permit clinicians to rate a secondary type if this was felt to be necessary. While the matrix in Table 27 used only primary types, the function may have been confounded by the inclusion of mixed or unclear diagnoses. A series of ANOVAS was also 204 computed for those patients described in terms of only one type.

(See Table 28). Neither the relative ranking nor the magnitude of the differences were much effected.

To check the possibility that significant differences in the general distribution were disguised by the presence of extreme scores, a Chi-Square was computed for each SC scale after it had been re- computed into frequency intervals of one standard deviation each. For private self-consciousness (PR-SC), the data were unremarkable and 2 analysis was nonsignificant, x (12)=7.20, ~> .80. Public self-con 2 sciousness was also nonsignificant, x (12)=18.02, ~ ).10. Hysterics and sociopaths, the smallest sample groups, were not represented in all sectors of the distribution. Sociopaths were consistently high scorers, with none more than a standard deviation below the mean and

35.7% in the highest interval. Hysterics, who might also be expected to be elevated in PU-SC, were overrepresented in the interval immediately above the mean but also had a disproportional number of extremely low scores. Paranoids, expected to be consistently low, did have a preponderance of low scorers but also had the widest scatter.

Obsessives and passive types were, as hypothesized, intermediate.

Frequency Distributions for the remaining Self-Consciousness scales are presented in Tables 29 and 30. As expected, hysterics and obsessives were low or moderate in anxiety. Passive types were usually above mean but had a restricted variability. In contrast, paranoids were below average in total but were overrepresented at the extremes.

Sociopaths were likewise highly variable, but typically above the mean. Table 28

Self-Consciousness Scale Means (with ANOVAs) as a Function of Unmixed Personality Type

Self-Consciousness subscales

Personality type N Private Public Social anxiety Total

Hysteric 8 23.12 18.37 10.25 51.75

Paranoid 22 25.41 18.23 12.04 56.59

Obsessive 40 24.40 19.70 11.57 55.62

Passive 20 22.75 18.65 14.80 56.20

Sociopath 12 25.75 22.50 13.67 62.67

!_(4,97) .767, .E_ >.50 1. 66 3' .E_ > . 10 1.832, .E_>.10 1.051, .E_ > .30

N 0 Ul Table 29

Distribution of Social Anxiety Scores by Personality Type

Intervals

Personality type N More than one SD Less than one SD Less than one SD More than one SD below the mean below the mean above the mean above the mean

Hysteric 11 18.2% 54.5% 18.2% 9.1%

Paranoid 25 24.0% 24.0% 36.0% 16.0%

Obsessive 44 11.4% 45.5% 34.1% 9.1%

Passive 23 0.0% 21.7% 69.6% 8.7%

Sociopath 14 14.3% 7.1% 57.1% 21.4%

N 0 "' Table 30

Distribution of Total Self-Consciousness Scale Scores by Personality Type

Intervals

More than one SD Less than one SD Less than one SD More than one SD Personality type N below the mean below the mean above the mean above the mean

Hysteric 11 9.1% 63.6% 27.3% 0.0%

Paranoid 25 24.0% 28.0% 20.0% 28.0%

Obsessive 44 18.2% 36.4% 38.6% 6.8%

Passive 23 8.7% 43.5% 39.1% 8.7%

Sociopath 14 7.1% 21.4% 42.9% 28.6%

N 0 -...J 208

The variability of the last two groups probably contributed to the failure of the ANOVA described above. In significance testing, based on distribution, personality style did relate to anxiety levels, 2 X (12)=22.91, ~(.05. Table 30 shows a somewhat similar distribution for total Self-Consciousness scale scores, but without the above mean congregation of passive style scorers. With only the sociopathic types disproportionately high, the degree of association is only 2 marginal, x (12)=18.67, £_(.10.

The data provides some tenuous support for the hypothesized link between personality style and self-consciousness. It is most consistently, and least surprisingly, encountered in support of an association between passive and character disorder types and anxiety in response to self-focused attention. It is interesting to note this commonality between two groups whose symptomology is so divergent.

These findings support an interpretation of the sociopath's perapatetic search for social interaction, stimulation, and excitement as an attempt to use constant activity as a cover by means of which to avoid self-evaluation. Anxiety is much more evident in passive types, as their methods of escape into states of subjective self-awareness are more limited. But, for both groups, avoidance of a tendency toward critical self-evaluation seems central.

The data in support of such an interpretation is, however, neither robust nor consistently apparent across significance tests.

Two factors seem to contribute to this inconsistency. The reliability of the type rating scale, while adequate, is sufficiently low that 209 misclassifications seem certain to increase error variance. The consistency with which sociopaths were distributed in SC scores

(Tables 29 and 30) is partially attributable to the relative reli ability of this subscale. The other subscales are not so well served.

In addition, the variability that characterizes patients reduces the power of those statistical tests based on a regression model. As described previously, patients are more variable than college students.

The power of ANOVA and t-tests is correspondingly diminished.

Comparisons between specific personality types. Specific predictions were made regarding the rank order of the types. On the total SC score, obsessives and passive types were predicted to be similar, as indeed they are, and higher than hysterics and sociopaths.

Likewise, obsessives were hypothesized to exceed hysterics and socio paths on PR-SC, but to have a lower mean on PU-SC. These planned comparisons are rendered inappropriate because of the dissimilarity of the scale scores of hysterics, lowest overall, and sociopaths, who proved the highest. Apparently, the primarily nonpsychotic hys terics in this sample were well enough defended that self-conscious ness of any type was minimized. Public self-consciousness as well as social anxiety were low. This is in keeping with the description of hysterics in terms of their characteristic naivete and la belle indif ference. Of the comparisons specified in Hypotheses 1 and 2, only that between combined obsessive and passive groups with paranoids re mains feasible. In this case, the difference is as expected, with paranoids scoring lower, but it falls short of statistical significance, 210

!_(90)=-.29, p> .70.

Table 31 summarizes expected ranks (duplicated from p. 135) and contrasts them with those obtained. A series of pairwise comparisons, based on these hypothesized rankings were also completed.

Significant differences between types were found for the Social

Anxiety Subscale (SA-SC). Passive types exceeded both hysterics, !_(32)

=3.08, .E_ < .01, and obsessives, !_(65)=2.86, .£ < .01, and were marginally greater than paranoid patients, !_(46)=1.88, .E_=.066. Rankings of other subtypes did not entirely conform with expectations. Obsessives, expected to be second highest in SA-SC, were not statistically different than hysterics or paranoids. Instead, sociopaths had the elevated scales. Although t-tests failed to show significant differences, the

Mann-Whitney U-test revealed that sociopaths significantly exceeded obsessives, U =2.017, E._ ( .05, and marginally exceeded hysterics, !!_= -z

43.5, n =11, .!!_ =14, .10 in Social Anxiety. This represents the 1 2 ..E.< only instance in the set of pairwise comparisons for which nonparametric tests showed greater sensitivity than !_-tests. Despite the partially unexpected ranking of the sociopaths, the essential feature of the hypothesis for the SA-SC subscale was confirmed in that passive types proved the most anxious.

With the major exception of the rank of hysterics, hypothesized relationships were also observed for the Public Self-Consciousness (PU-

SC) subscale. Sociopaths obtained marginally higher scores than obsessives, ~(56)=1.80, r_=.078, and exceeded passive types, !_(35)=2.28, r_=

.029, and paranoids !_(37)=2.46, _£=.019, to a statistically reliable Table 31

Expected and Observed Rankings of Personality Types on Self-Consciousness Scores

Self-Consciousness scales

Expected Observed

Social So.oial Personality type Private Public anxiety Total Private Public anxiety Total

Hysteric 3 1 3 3 4 4 5 5

Paranoid 5 5 3 5 3 5 3 4

Obsessive 1 3 2 2 2 2 4 3

Passive 1 3 1 1 5 3 1 2

Sociopath 3 1 3 3 1 1 2 1 212 degree. Contrary to expectations, hysterics were ranked second lowest on PU-SC and varied significantly from sociopaths, ~(23)=2.67, z=.014.

Parametric and nonparametric tests revealed no statistically reliable differences between personality types on private self-consciousness

(PR-SC). The other trends toward significance occurred on the total score. Sociopaths were marginally higher than obsessives, ~(56)=

1.76, z=.083, and significantly higher than hysterics, ~(23)=2.31, z=.030, on total scores. While sociopaths were the "unknown quantity", as revealed in the qualifying comments in the discussion of their anxiety level, their highest rank on both total SC and PR-SC is incon sistent with what was predicted. Neither was the consistently low score of hysterics hypothesized. While rationales for the observed results have been provided, results from both groups may be anomalous, a happenstance due to the relatively small sample size. It is this combination of small groups and high standard deviations that accounts for the dearth of significant differences. The means do differ. In the case of SA-SC particularly, they vary much in the manner that was predicted, confirming the observations from the previous section. Like wise, on PU-SC, the sociopaths achieved the highest rank in the manner predicted. But the mutual validation of the subscale and the distinc tion between types is marred by the unexpected low scores from the hysterics. Both PU-SC and PR-SC, less distinct and central for patho logical groups than for normals, show scant evidence of being related to personality style. For PR-SC especially, the results suggest little relevance to diagnosis and offer no support for the hypothetical link 213 of introspection and personality types.

Subscale item averages and personality types. In the attempt to demonstrate a link between personality style and management of self focused attention, comparisons are possible within, as well as across personality types. Hypothesis 3 specified an expected pattern of scores over the three subscales for each type. Table 32 replicates those predictions, showing the expected rank order of the average sub scale item. Tests against these hypotheses would prove misleading if unrefined subscale item averages were used. For all types the average

PU-SC item was highest. Except for passive types the average of the second greatest magnitude was PR-SC. Differences in the relative elevation of subscale averages across types are obscured by the uniform elevation of the PU-SC items. As previously discussed, a correction was made to eliminate subscale elevation sample-wide prior to further analysis.

The observed rank for the revised item averages are presented in Table 32. Table 33 includes the means. Only for the passive types is there a significant difference. The average from the SA subscale is significantly higher than that from PR-SC, ~(23)=3.25, £<.01, or

PU-SC, ~(28)=3.02, ~~.01. This is consistent with the expected high level of anxiety. Contrary to expectation, the passive types reported little attention to internal experience. A marginally significant difference was observed among sociopaths, for whom PU-SC exceeded PR

SC items, ~(14)=2.06, £=.06, as expected. For the other personality types, differences are insufficient to interpret reliably. Hysterics Table 32

Expected and Observed Subscale Ranks for Each Personality Type

Self-Consciousness subscales

Expected Observed

Personality type Private Public Social anxiety Private Public Social anxiety

Hysteric 2 1 2 1 2 3

Paranoid 1 3 2

Obsessive 1 2 1 2 3

Passive 1 2 1 3 2 1*

Sociopath 2 1 3 1 2

* Significantly greater than lower ranking subscale averages,£ <.01~ 215

Table 33

Self-Consciousness Subscale Item Means for Personality Types

Self-Consciousness subscale

Personality type Private Public Social anxiety

Hysteric 2.34 2. 24 2.10

Paranoid 2.44 2.22 2.30

Obsessive 2.45 2.45 2.25

Passive 2.31 2.37 2. 77

Sociopath 2.48 2.85 2.68 216 failed to peak on public self-awareness as expected. For obsessives the difference between PR-SC and PU-SC items, expected to be signifi cant, failed to appear. In fact, when mixed types are excluded, the remaining 40 obsessives are minimally higher on Public Self-Conscious ness (mean item value of 2.47) than on Private Self-Consciousness

(2.44). In all other respects, the exclusion of mixed types affected neither ranks or significance levels. In view of the results from the factor analysis, the distinction between PR-SC and PU-SC is aca demic. Table 34 compares the means of the Introspection and Anxiety

Factors. As the former corresponds more closely with PR-SC, although with PU-SC elements, and the later is highly related to SA-SC, a comparison is possible using the hypothesized rankings as recapitulated in the above review section. Excluding the paranoids, for whom no relative rankings were suggested, and the sociopaths, for whom the anxiety level was problematic, the relative level of the two factor scores is in the direction expected.

Comparisons among subscales offer some support for the hypoth esized link between self-consciousness and personality style. Pre dicted differences were found in which the anxiety scale was elevated among passive types. A marginal distinction between public and private self-consciousness was found for sociopaths. But, as with the preced ing sections of analysis, the results are neither so consistent with hypothesis nor so robust as might be hoped. The distinctions between personality types may have been too gross for sensitive testing.

Further research might profitably use more thoroughly validated in- 217

Table 34

Self-Consciousness Scale Factor Score Means for Personality Types

Personality type Anxiety factor Introspection factor

Hysteric -.41 .osa

Paranoid -.15 .o2a

Obsessive -.02 .04a

Passive .1sa -.44

Sociopath .41a .08 aHigher ranking factor score. 218

struments such as the Millon Multiaxial Clinical Inventory (Millon,

1977) or more intensive diagnostic assessments prior to rating per

sonality style. Such steps would likely increase the frequency of

statistically significant results. But, with the possible exception

of the social anxiety in the absense of introspection that typifies

passive types, the current research suggests that any distinctions

that are eventually made will have more theoretical than applied im

plications. Self-consciousness is not so closely linked to any

specific diagnostic group or personality type as to be of particular

use in assessment or to automatically present an issue to be addressed

in psychotherapy.

Discussion

Implications for self-awareness theory. The results of this

research provide additional information with which to evaluate the

Self-Consciousness scale (SC) and self-awareness theory. Two poten

tial problems with the SC scale have been identified: the lack of

control of the acquiescence response set, and the impact of transient

emotional states on a presumably dispositional measure. Since neither

issue was directly tested, further research is needed. The SC scale

has three items for which the maximum self-consciousness scores re

quires a negative answer. These items were among the four with the

lowest correlation with the total score, and none of these was im portantly loaded on the two major factors. It may be that patients

failed to understand the double negative required or a pervasive 219 response set may be in operation. While there was no measure of the influence of acquiescence, the low degree of association between the three contrary items included to control it and the rest of the scale suggests that it is considerable. Three other items were rejected by fewer than 15% of the sample. This extreme skewness also undermines the sensitivity of the scale and may be taken as further evidence of an artifactual response set. The SC scale would perform better if revised to be better balanced for acquiescence.

The 12 patients completing the scale immediately prior to their screening appointments were significantly higher in SC than other comparison groups. In the absense of any apparent differences from other outpatients seen at the clinic, the most plausible explanation is that the common feature in their situation, the imminent screening appointment, acted as a powerful manipulation increasing self-focused attention. The validation research of the scale has demonstrated that the scale taps a dispositional tendency comparable in effect to stimuli increasing objective self-awareness (Carver & Scheier, 1978; Fenigstein,

1979; Scheier & Carver, 1977). But perhaps, as is suggested in this study and in the results reported by Buss and Scheier (1976), the dispositional and situational elements of self-attention are indepen dent and additive. Further research is indicated on this topic. To the extent that the scale proves sensitive to state variables, the generalizability of its results decrease and the sensitivity of com parisons--at least in studies such as this, in which the influence of state variables accumulates as error variance--is proportionally de- 220

creased.

Some findings with normal subjects fail to generalize to a

clinical population. The increased correlation between SA-SC and the

remaining subscales is not the point. In fact, the anxiety factor, al-

though more central for patients, is quite similar to that obtained

from students. The important difference is the disappearance of dis-

tinct public and private aspects of self-awareness. For a clinical

population, these two subscales merge featurelessly into a secondary

factor. This research suggests that little confidence be placed on

the meaning of subscale scores. A two factor model of self-conscious-

ness--which is entirely adequate for Duval and Wicklund's (1972)

formulations--may be more appropriate. In the absense of further con-

firmation, it is suggested that little importance be attached to the

PR-SC and PU-SC scores, at least for clinical populations. This finding

counters the trend in the research literature which treats the public- private dichotomy as the fundamental contribution of self-consciousness

research. Buss (1980) reports a range of experiments on normal sub-

jects in which the two dimensions have behavioral consequences which are either independent or opposing. He presents this research as:

evidence demonstrating how different the two self-consciousness traits are, despite their moderate positive correlation. People who. reflect about their psychological "insides" are very different from people who are concerned about themselves as social objects. (p. 61)

Because the distinct behavioral implications of PR-SC and PU-SC have been replicated with normals, the current research does not challenge those findings. In fact, the absense of the distinction should be 221 replicated in future research. But the present results indicate that, for a clinical population at least, introspective behavior is too un differentiated for the dichotomy to be applicable to the population as a whole. A question for future researchers is whether it is the psy chiatric patient or the college student that is a more representative group on which to build a model of self-consciousness.

This does not imply that the results of the current research are entirely discontinuous with previous studies. The social anxiety sub scale, specifically, generalizes quite well. The consistency in mean scores across samples, very different in life situations, suggests the broad applicability of the central dimension. The internal consistency of SA-SC items was retained. This research provided some needed vali dation for the subscale. It makes sense that anxiety accounts for a greater proportion of the explained variance among patients than normals. The significant and important inverse relationship with the

Ego Strength scale was hypothesized and is consistent with clinical wisdom. The SA-SC appears to be a valid index of anxiety about evalua tion in a social context.

This research also supports a central tenant of self-awareness theory in clarifying the relationship of self-consciousness and anxiety.

Although never discussed in those terms specifically, the self-denigra tion, the heightened acitvity and conformity, and the attempted escape into subjective self-awareness all suggest that objective self-aware ness is a major source of anxiety. Yet, paradoxically, people regular ly and willingly maintain this state of consciousness. Self-denigra- 222 tion cannot be an automatic consequence. The factor structure of the

SC scale, for both normal and pathological samples, includes anxiety as one dimension among others. For patients, this aspect is more pronounced and self-consciousness is more likely to be expressed as anxiety. But neither patients nor normals need inevitably respond with evaluation. A non-affective response of self-observation, whether or not it is differentiated into public and private components, is also available. Wicklund (1975b) acknowledges that objective self-awareness may be acceptable or even enjoyable when behavior exceeds stands. Re search with the SC scale suggests that a non-evaluative self-awareness is also normative. The implication is clear that some intervention may shift the internal structure of self-consciousness to lessen the role of anxiety. With patients, for whom anxiety takes on a more cen tral role, this is a particularly relevant project.

Implications for future research and clinical applications.

Two clinical issues were examined in this study, incidence of premature termination and classification of personality type. Despite the ex pectation of significant association, self-consciousness had minimal practical impact on either variable. The attempt to understand this leads to several methodologicql issues and suggestions for future re search. Greater differences between types are probable if the type of rating scale was made more reliable or replaced with a more thoroughly validated scale. A replication with larger samples, especially of hysterics and sociopaths, would be of interest. In this study, signif icant differences were found on the SA-SC in comparison with passive 223 patients and sociopaths, and on PU-SC and total SC with sociopaths.

More statistically reliable differences may be anticipated in replica tion studies. The absolute lack of significant findings in the pre diction of termination suggests that self-consciousness levels are not related. If the central assumption, that dropping out represents flight from excessive anxiety, has any value, then it refers to state anxiety aroused during the course of treatment. Rather than using a primarliy dispositional measure prior to therapy it would be profitable to examine variables more directly linked to the nature of interaction in treatment. These would include indices of patient skill in handling anxiety and therapist skills in modulating self-consciousness. How self-awareness is handled appears more critical than level of self- consciousness.

The findings in the present study suggest no particular modifi cations in the practice of psychotherapy. If anything, they offer the reassurance that the degree of self-consciousness on the part of the patient is unlikely to be associated with diagnosis or flight from therapy. More disturbed patients may be expected to display heightened anxiety as part of a generally excessive self-consciousness. Passive types should display social anxiety in the absense of evidence of much introspection. But patients as a group differ little from normals in self-consciousness levels. What differences do appear are not demon strably important in the process of therapy.

The findings relevant to clinical practice are, then, essen tially negative. This is generally reassuring in that hypothetical 224 aversive consequences are not rE2gularly or inevitably associated with the very basic therapeutic proct2ss of insight. While self-conscious ness is functionally associated with concern about social evaluation, there is no evidence that this Elnxiety assumes a major role in the clinical setting. SUMMARY

The Self-Consciousness scale (SC) was administered to 230 veterans referred for outpatient psychotherapy. In the absense of previous published reports on the use of this scale with a clinical population, this study was in part a generalization study. It checked the external validity of results obtained from the college students sampled to date. Means from patients varied little from those pre viously reported. There is no evidence that gross level of self consciousness, for the total score or subscales, differs for patients or varies across age, education, social class background, or any of the other variables on which veterans are discrepant from students.

However, definite limits to the generalizability of previous findings were encountered. Patients are significantly more variable in their scores than students. The subscale scores are also more highly cor related. The response to self-focused attention, with three identifi ably different components for students, was less differentiated for veterans. A factor analysis reveals only two major factors, with so cial anxiety assuming a more central role. This reaffirms a central assumption of self-awarenes-s theory: that self-consciousness is easily transformed into aversive self-evaluation. But it also indicates that caution should be exercised in applying the SC scale to non-student

225 226 populations. The total score, primarily as a result of the Social

Anxiety subscale, is significantly and inversely related to the Ego

Strength scale. The more pathological the sample, the higher the SC scores that may be expected. Those who admit to symptoms also report a dispositional tendency to examine themselves and to be made anxious by this evaluation.

Self-consciousness was hypothesized to relate to perseverance in treatment. Highly self-evaluative patients were expected to flee the extremely focused attention engendered by therapy. No support was found for this hypothesis. Neither self-consciousness nor the other patient, therapist, or contextual features examined successfully dif ferentiated those terminating unilaterally within five sessions from those continuing in verbal therapy. While the Social Anxiety (SA-SC) subscale or the total SC score alternately contributed to discriminant functions, these functions were nonsignificant and unsuccessful in classifying continuers and terminators at better than chance levels.

It appears that the relationship of patient and therapist may be a more powerful determinant of perseverance in treatment than is the patient's dispositional tendency toward self-scrutiny.

Different levels of self-consciousness were expected to char acterize different personality types from among the patient sample.

The tests of this hypothesis generated mixed results. An ANOVA re vealed no significant difference in total self-consciousness among hysterics, paranoids, obsessives, passive types, and sociopaths. But pairwise comparisons did indicate that sociopaths had relatively more 227 elevated scores, significantly so when contrasted with hysterics.

Contrary to expectation, obsessives achieved the second highest, rather than the highest, mean score on private self-consciousness (PR-SC) and hysterics were next to lowest on public self-consciousness (PU-SC) when expected to be among the top two. As hypothesized, sociopaths were the highest of the personality styles on PU-SC and passive types had the highest mean for SA-SC, to a statistically reliable degree for two of the four comparisons. With the exception of hysterics, each type also had the hypothesized peak subscale score. In the face of the general variability of patient scores and the less than optimal reliability of the personality rating task, these results lend modest support to the hypothesized link between personality style and self-consciousness lev els. The degree of association obtained is, however, insufficient to be of practical use in diagnosis or anticipating issues in therapy.

The self-consciousness generated in psychotherapy is related to self-evaluation. But the association is neither so consistent nor so strong as to constitute a threat to treatment in the typical case. REFERENCE NOTES

1. Crandell, J.S. Objective self-awareness, anxiety, and self-esteem. Unpublished manuscript, 1978. (Available from John S. Crandell, Northwestern Community Mental Health Center, 259 East Erie, Chicago, Illinois 60611).

2. Turner, R.G. Personal communication, January 29, 1980.

228 REFERENCES

Adams, H.B., & Cooper, G.D. Three measures of ego strength and prognosis for psychotherapy. Journal of Clinical Psychology, 1962, ~. 490-494.

Affleck, D.C., & Mednick, S.A. Use of the Rorschach Test in the prediction of the abrupt termination in individual psychotherapy. Journal of Consulting Psychology, 1959, ~. 125-128.

American Psychiatric Association. DSM III: Diagnostic and statis tical manual of mental disorders. Washington, D.C.: American Psychiatric Association, 1968.

American Psychiatric Association. Diagnostic and statistical manual of mental disorders (3rd ed.). Washington, D.C.: American Psychiatric Association, 1980.

Applezweig, M.H., Dibner, A.S., & Osbourne, R.I. PEAQ: a measure of psychopathic behavior. Journal of Psychology, 1958, ~. 26-30.

Auld, F., & Eron, L.D. Use of the Rorschach scores to predict whether patients will continue psychotherapy. Journal of Consulting Psychology, 1953, !2, 104-109.

Auld, F., & Myers, J.K. Contributions to a theory for selecting psychotherapy patients. Journal of Clinical Psychology, 1954, lQ_, 56-60.

Baekeland, F., & Lundwall, L. Dropping out of treatment: a critical review. Psychological Bulletin, 1975, 82, 738-783.

Bailey, M.A., Warshaw, L., & Eichler, R.M. Study of factors related to length of stay in psychotherapy. Journal of Clinical Psy chology, 1959, 12, 442-444.

Barron, F. An ego-strength scale which predicts response to psycho therapy. Journal of Consulting Psychology, 1953, !2, 327-333.

Barron, F., & Leary, T.F. Changes in psychotherapy patients with and without psychotherapy. Journal of Consulting Psychology, 1955, _11_, 239-245.

229 230

Bassos, C. A. Two modes of consciousness and their pseudo-modes. Psychotherapy: Theory, Research, and Practice, 1976, ]l, 335- 342.

Bern, S. L. Measurement of psychological anrogyny. Journal of Con sulting and Clinical Psychology, 1974, 42, 155-162.

Block, J. Challenge of response sets. New York: Appleton, Century, and Crofts, 1965.

Block, J., & Thomas, H. Is satisfaction with self a measure of adjustment? Journal of Abnormal and Social Psychology, 1955, 21, 254-259.

Brandt, L. W. Rej~ction of psychotherapy. Archives of General Psy chiatry, 1964, lQ, 310-313.

Brandt, L. W. Studies of 'dropout' patients in psychotherapy: a review of findings. Psychotherapy: Theory, Research, and Practice, 1965, ~. 6-12.

Brehm, S. S. Developmental aspects of objective self-awareness. Dissertation Abstracts International, 1974, 34, 4654B-4655B.

Brill, M. Q., & Storrow, H. A. Social class and psychiatric treatment. Archives of General Psychiatry, 1960, 1, 340-344.

Broen, W. E., & Storms, L. H. Reaction potential ceiling and response decrements in complex situations. Psychological Review, 1961, ~' 405-415.

Brown, J. S., & Kosterlitz, N. Selection and treatment of psychiatric outpatients. Archives of General Psychiatry, 1964, 11, 425-438.

Bugenthal, J. F. T. Search for authenticity. New York: Holt, Win ston and Rhinehart, 1965.

Buss, A. H. Self-consciousness and social anxiety. San Fransisco: Freeman, 1980.

Buss, A. H., & Plomin, R. Temperament theory of personality develop ment. New York: Wiley-Interscience, 1975.

Buss, D. M., & Scheier, M. F. Self consciousness, self awareness, and self attribution. Journal of Research in Personality, 1976, lQ, 463-468.

Butler, J. H., & Haigh, G. V. Change in the relation between self con cepts and ideal concepts consequent upon client-centered counsel ing. In C. R. Rogers and R. F. Dymond (Eds.), Psychotherapy and personality change. Chicago: University of Chicago Press, 1954, 55-75. 231

Byrne, D. Introduction to personality. Englewood Cliffs, New Jersey: Prentice-Hall, 1974.

Campbell, D., & Fiske, D. Convergent and discriminant validation by the multitrait-multimethod matrix. Psychological Bulletin, 1959, 56, 81-105.

Caracena, P. F. Elicitation of dependency expressions in the initial stage of psychotherapy. Journal of Counseling Psychology, 1965, .!1_, 268-274.

Cartwright, D. S. Success in psychotherapy as a function of certain actuarial variables. Journal of Consulting Psychology, 1955, J2, 357-363.

Carver, C. S. Aggression as a function of objective self-awareness and attitudes toward punishment. Journal of Experimental Social Psychology, 1975, !l, 510-519.

Carver, C. S., & Glass, D. C. Self-consciousness scale: a discriminant validity study. Journal of Personality Assessment, 1976, 40, 169-172.

Carver, C. S., & Scheier, M. F. Self focusing effects of dispositional self consciousness, mirror presence, and audience presence. _J_o_u_r_n_a_l_o_f_P_e_r_s_o_n_a_l_i_t.,_y_a_n_d_S_o_c_l_·a_l_P_s~y.__ch_o_l_o_..g..._y, 19 7 8, 36, 324-332 .

Chameides, W. A., & Yamamoto, J. Referral failures: a one year follow up. American Journal of Psychiatry, 1973, 130, 1157-1158.

Chase, P. H. Self concepts in adjusted and maladjusted hospital patients. Journal of Consulting Psychology, 1957, ~' 495-497.

Clayton, M. R., & Graham, J. R. Predictive validity of Barron's ES scale: the role of symptom acknowledgement. Journal of Consul ting and Clinical Psychology, 1979, 47, 424-425.

Clopton, J. R. Development of special MMPI scales. In C. S. Newmark (Ed.), MMPI: clinical and research trends, New York: Praeger, 1979, 354-372.

Cohen, J. Coefficient of agreement for nominal scales. Educational and Psychological Measurement, 1960, 20, 37-46.

Collins, P. T. Objective self awareness, performance, and self esteem of overweight and normal weight high school students. Disser tation Abstracts International, 1976, 36, 4223B. 232

Conrad, D. C. Duration of the therapeutic relationship and therapists' successive judgements of patients' mental health. Journal of Clinical Psychology, 1954, !Q, 229-233.

Corotto, L. V., & Cornutt, R. H. Ego strength: a function of the measuring instrument. Journal of Projective Techniques, 1962, 26, 228-230.

Cottrell, N. B. Performance in the presence of other human beings: mere presence, audience, and affiliation effects. In E. C. Symond, R. A. Hoppe, & G. A. Milton (Eds.), Social facilitation and imita tive behavior. Boston: Allyn & Bacon, 1968, 91-110.

Crites, J. 0. Ego-strength in relation to vocational interest develop ment. Journal of Counseling Psychology, 1960, 1. 137-143.

Crumpton, E., Cantor, J. M., & Batiste, C. Factor analytic study of Barron's Ego Strength scale. _J_o_u_r_n_a_l___ o_f __ C_l_i_n_l_·c_a_l ___ P_s~y~c_h_o~l_o~gy~, 1960, j&, 283-291.

Csikszentmihalyi, M. Flow experience. In D. Goleman & R. J. Davidson (Eds.), Consciousness: the brain, states of awareness, and alter nate realities. New York: Irvington, 1979.

Cumming, R. D. Philosophy of Jean-Paul Sartre. New York: Modern Library, 1965.

Cummings, W. H. Audience observation as a noxious stimulus: escape and punishment contingencies. Dissertation Abstracts Internation al, 1976, 36, 4224B.

Dahlstrom, W. G., Welsh, G. S., & Dahlstrom, L. E. MMPI handbook (Vol. 2). Minneapolis: University of Minnesota Press, 1975.

Dana, R. H. MMPI performance and electroshock treatment. Journal of Clinical Psychology, 1957, !l, 350-355.

David, K. H. Ego-strength, sex differences, and description of self, ideal, and parents. Journal of General Psychology, 1968, ~. 79-81.

Davis, D., & Brock, T. C. Use of the first person pronoun as a function of increased objective self-awareness and performance feedback. Journal of Experimental Social Psychology, 1975, ll• 381-388.

DeLoach, S. S. Level of ego development, degree of psychopathology and continuation or termination of outpatient psychotherapy involve ment. Dissertation Abstracts International, 1977, li• 5348B. 233

Dengrove, E., & Kutash, S. B. Why patients discontinue treatment in a mental hygiene clinic. American Journal of Psychotherapy, 1950, i· 457-472.

DeWald, P. Psychotherapy: a dynamic approach. New York: Basic Books, 1971.

Distler, L. S., May, P. R. A., & Tuma, A. H. Anxiety and ego strength as predictors of response to treatment in schizophrenic patients. Journal of Consulting Psychology, 1964, 28, 170-177.

Dodd, J. Retrospective analysis of variables related to the duration of treatment in a university psychiatry clinic. Journal of Ner- vous and Mental Disease, 1970, 151, 75-85.

Doyle, E. Rorschach sex responses in relation to defense style of normal subjects. Unpublished doctoral dissertation, Loyola University of Chicago, 1968.

Duval, S. Conformity as a function of perceived level of personal uniqueness and being reminded of the object status of self. Unpublished docotral dissertation, University of Texas at Austin, 1972.

Duval, S., & Hensley, V. Extensions of objective self-awareness theory: the focus of attention- causal attribution hypothesis. In J. H. Harvey, W. J. Ickes, & R. F. Kidd (Eds.), New directions in attri bution research (Vol. 1). New York: Wiley, 1976, 165-198.

Duval, S., & Wicklund, R. A. Theory of objective self-awareness. New York: Academic Press, 1972.

Duval, S., & Wicklund, R. A. Effects of objective self awareness on attribution of causality. Journal of Experimental Social Psycho ~. 1973' 2_, 17-31.

Duval, S., Wicklund, R. A., & Fine, R. L. Avoidance and objective self awareness under conditions of high and low intra-self discrepancy. In S. Duval & R. A. Wicklund, Theory of objective self awareness. New York: Academic Press, 1972, 16-21.

Dymond, R. F. Adjustment changes over therapy from self sorts. In C. R. Rogers & R. F. Dymond (Eds.), Psychotherapy and personality change. Chicago: University of Chicago Press, 1954.

Endicott, N. A., & Endicott, J. Prediction of improvement in treated and untreated patients using the Rorschach Prognostic Rating Scale. Journal of Consulting Psychology, 1964, 28, 342-348. 234

Ends, E. J., & Page, C. W. Study of functional relationships among measures of anxiety, ego strength and adjustment. Journal of Clinical Psychology, 1957, ll• 148-150.

Exner, J. E. Self focus sentence completion: a study of egocentri city. Journal of Personality Assessment, 1970, 12, 437-455.

Eysenck, H. J. Classification and the problem of diagnosis. In H. J. Eysenck (Ed.), Handbook of abnormal psychology. London: Pitman Medical Publishing Company, 1960, 1-31.

Feldman, R., Lorr, M., & Russell, S. B. Mental hygiene clinic case survey. Journal of Clinical Psychology, 1958, ~. 245-250.

Fenigstein, A. Self consciousness, self attention, and social inter action. Journal of Personality and Social Psychology, 1979, 12, 75-86.

Fenigstein, A., Scheier, M. F., & Buss, A. H. Public and private self consciousness: assessment and theory. Journal of Consulting and Clinical Psychology, 1975, ~. 522-527.

Fiester, A. R., Mahrer, A. R., Giambra, L. M., & Ormiston, D. W. Shaping a clinic population: the dropout problem reconsidered. Community Mental Health Journal, 1974, !Q, 173-179.

Fiester, A. R., & Rudestam, K. E. Multivariate analysis of the early dropout process. Journal of Consulting and Clinical Psychology, 1975, 43, 528-535.

Figetakis, N. Process-reactive schizophrenia: ego-strength and select ed psychosexual dimensions. Dissertation Abstracts International, 1964, 12· 625. Fisher, S., & Greenburg, R. P. Scientific credibility of Freud's theories and therapy. New York: Basic Books, 1977.

Fiske, D. W., Cartwright, D. S., & Kirtner, W. L. Are psychothera peutic changes predictable? Journal of Abnormal and Social Psy chology, 1964, 69, 418-426.

Fowler, R. D., Tell, S. K., & Coyle, F. A. Measurement of alcoholic response to treatment by Barron's Ego-Strength scale. Journal of Psychology, 1967, ~. 65-68.

Frank, J.D. Influence of patients' and therapists' expectations on the outcome of psychotherapy. British Journal of Medical Psychol £gy, 1968, ~. 349-356. 235

Frank, J.D., Gliedman, L. H., Imber, S. D., Nash, E. H., & Stone, A. R. Why patients leave psychotherapy. Archives of Neurology and Psy chiatry, 1957, ll• 283-299.

Freedman, A. M., Kaplan, H. I., & Sadock, B. J. Modern synopsis of the comprehensive textbook of psychiatry. Baltimore, Maryland: Williams & Wilkins, 1972.

Freedman, N., Englehardt, D. M., Hankoff, L. D., Glick, B. S., Kaye, H., Buchwald, J., & Stark, P. Drop-out from outpatient psychiatric treatment. Archives of Neurology and Psychiatry, 1958, 80, 657- 666.

Freeman, H. E. Attitudes toward mental illness among relatives of former patients. American Sociological Review, 1961, 26, 59-66.

Fulkerson, S. C., & Barry, J. R. Methodology and research on the prog nostic use of psychological tests. Psychological Bulletin, 1961, 58, 177-204.

Furlong, W. B. Flow experience: fun in fun. Psychology Today, 1976, !Q, 35-80.

Gaines, L. S., & Fretz, B. R. Reliability of the Barron Ego Strength Scale in an individual form. Psychological Reports, 1969, 25, 513-514.

Gallagher, J. J. Test indicators for therapy prognosis. Journal of Consulting Psychology, 1954, ~. 409-413.

Garfield, S. L. Note on patients' reasons for terminating therapy. Psychological Reports, 1963, ll• 38.

Garfield, S. L. Research on client variables in psychotherapy. In S. L. Garfield & A. E.Bergin (Eds.), Handbook of psychotherapy and behavior change. New York: Wiley, 1978, 191-232.

Garfield, S. L., Affleck, D. C., & Muffly, R. Study of psychotherapy interaction and continuation in psychotherapy. Journal of Clin- ical Psychology, 1963, 12, 473-478.

Garfield, S. L., & Kurz, M. Evaluation of treatment and related pro cedures in 1216 cases referred to a mental hygiene clinic. Psy chiatric Quarterly, 1952, ~. 414-424.

Gedo, J. E., & Goldberg, A. Models of the mind: a psychoanalytic theory. Chicago: University of Chicago Press, 1973.

Geller, V., & Shaver, P. Cognitive consequences of self aw~reness. Journal of Experimental Social Psychology, 1976, ll• 99-108. 236

Gertler, R., Raynes, R. E., & Harris, N. Assessment of attendance and outcome at an outpatient alcoholism clinic. Quarterly Journal of Studies in Alcohol, 1973, ~. 3, part A.

Getter, H., & Sundland, D. M. The Barron Ego Strength scale and psy chotherapy outcome. Journal of Consulting Psychology, 1962, ~. 195.

Gibby, R. B., Stotsky, B. A., Hiler, E. W., & Miller, D. R. Validation of Rorschach criteria for predicting duration of therapy. Journal of Consulting Psychology, 1954, ~. 185-191.

Gibby, R. B., Stotsky, B. A., Miller, D. R., & Hiler, E. W. Prediction of duration of therapy from the Rorschach test. Journal of Con sulting Psychology, 1953, ll• 348-354.

Gliedman, L. H., Stone, A. R., Frank, J. D., Nash, E., & Imber, S.D. Incentives for treatment related to remaining or improving in psy chotherapy. American Journal of Psychotherapy, 1957, 11, 589-598.

Goldstein, A. P. Therapist and client expectation of personality change in psychotherapy. Journal of Counseling Psychology, 1960, 1_, 180-184.

Gottesman, I. I. More construct validation of the Ego Strength scale. Journal of Consulting Psychology, 1959, ~. 342-346.

Gough, H. G., & Heilbrun, A. B. Adjective checklist manual. Palo Alto, California: Consulting Psychologists Press, 1965.

Grace, D. P. Experimental study of the validity of Barron's Ego Strength scale. Dissertation Abstracts International, 1960, ll, 238-239.

Graham, J. R. MMPI: a practical guide. New York: Oxford University Press, 1977a.

Graham, J. R. Review of Minnesota Multiphasis Personality Inventory special scales. In P. McReynolds (Ed.), Advances in psychological assessment (Vol. 4). San Francisco: Jossey-Bass, 1977b, 11-55.

Gravitz, M. A. Use of a short form of Barron's Ego Strength scale with normal adults. Journal of Clinical Psychology, 1970, ~. 223.

Greenfield, N. S., Roessler, R., & Crosley, A. P. Ego strength and length of recovery from infectious mononucleosis. Journal of Nervous and Mental Disease, 1959, 128, 125-128. 237

Greenson, R. B. Technique and practice of psychoanalysis (Vol. 1). New York: International Universities Press, 1967.

Guilford, J. P., & Zimmerman, W. S. Guilford-Zimmerman Temperament Survey: manual of instructions and interpretations. Beverly Hills, California: Sheridan Supply Company, 1949.

Gulas, I. Client-therapist congruence in prognostic and role expec tations as related to client's improvement in short-term psycho therapy. Dissertation Abstracts International, 1974, li• 2430B.

Gundlach, R. H., & Geller, M. Problem of early termination: is it really the terminee? Journal of Consulting Psychology, 1958, ~. 410.

Hawkinson, J. R. Study of the construct validity of Barron's Ego Strength scale with a state mental hospital population. Disser- tation Abstracts International, 1962, ~. 4081.

Heilbrun, A. B. Male and female personality correlates of early termi nation in counseling. Journal of Counseling Psychology, 1961, ~. 31-36.

Heilbrun, A. B. Psychological factors related to counseling readiness and implication for counselor behavior. Journal of Counseling Psychology, 1962, ~. 353-358.

Heilbrun, A. B. History of self-disclosure in females and early de fection from psychotherapy. Journal of Counseling Psychology, 1973, 20, 250-257.

Heine, R. W., & Trosman, H. Initial expectations of the doctor-patient interaction as a factor in continuance in psychotherapy. Psychi atry, 1960, ~. 275-278.

Heitler, J. B. Preparatory techniques in initiating expressive psycho therapy with lower-class unsophisticated pa~ients. Psychological Bulletin, 1976, 83, 339-352.

Herron, W. C., Guido, S. M., & Kanter, R. E. Relationships among ego strength measures. Journal of Clinical Psychology, 1965, 11, 403-404.

Hiler, E. W. Analysis of patient-therapist compatibility. Journal of Counseling Psychology, 1958, ~. 341-347.

Hiler, E. W. Initial complaints as predictors of continuation in psy chotherapy. Journal of Clinical Psychology, 1959, 12, 244-245. 238

Himelstein, P. Further evidence on the Ego Strength scale as a measure of psychological health. Journal of Consulting Psychology, 1964, 28, 90-91.

Hoehn-Saric, R., Frank, J. D., Imber, D. S., Nash, E. H., Stone, A. R., & Battle, C. C. Systematic preparation of patients for psycho therapy: I. effects on therapy behavior and outcome. Journal of Psychiatric Research, 1964, ~. 267-281.

Hollon, S., & Mandell, M. Use of the MMPI in the evaluation of treat ment effects. In J. N. Butcher (Ed.), New developments in the use of the MMPI. Minneapolis: University of Minnesota Press, 1979, 241-302.

Holmes, D. S. Male-female differences in MMPI ego strength: an arti fact. Journal of Consulting Psychology, 1967, ]l, 408-410.

Horenstein, D Correlates of initial client disturbance: expectations for therapy, dropouts, resistance, and demographic description. Journal of Clinical Psychology, 1975, ]l, 709-715.

Horenstein, D., & Houston, B. K. Expectation-reality discrepancy and premature termination from psychotherapy. Journal of Clinical Psychology, 1976, ~. 373-378.

Horton, M., & Kriauciunas, R. Minnesota Multiphasic Personality Inven tory differences between terminators and continuers in youth counseling. Journal of Counseling Psychology, 1970, ll• 98-101.

Hunter, C. G., & Goodstein, L. D. Ego strength and types of defensive and coping behavior. Journal of Consulting Psychology, 1967, ]1, 432.

Ickes, W. J., Wicklund, R. A., & Ferris, C. B. Objective self-aware ness and self-esteem. Journal of Experimental Social Psychology, 1973, 1, 202-219.

Imber, S. D., Frank, J.D., Gliedman, L. H., Nash, E. H., & Stone, A. R. Suggestibility, social class and the acceptance of psychotherapy. Journal of Clinical Psychology, 1956, Jl, 341-344.

Imber, S. D., Nash, E. H., & Stone, A. R. Social class and duration of psychotherapy. Journal of Clinical Psychology, 1955, 11, 281-284.

Insko, C. A., Worchel, S., Songer, E., & Arnold, S. E. Effort, object ive self-awareness, choice, and dissonance. Journal of Person ality and Social Psychology, 1973, ~. 262-269. 239

Jones, R. D. Study of the relationship among ego strength, persistence, and degree of success in rehabilitation activity. Dissertation Abstracts International, 1969, 30, 1868A.

Kamin, I., & Caughlin, J. Subjective experiences of outpatient psycho therapy. American Journal of Psychotherapy, 1963, 12, 660-668.

Kassebaum, C. G., Couch, A. S., & Slater, P. E. Factorial dimensions of the MMPI. Journal of Consulting Psychology, 1959, ~. 226-236.

Katz, J., & Solomon, R. Z. Patient and his experiences in an outpa tient clinic. Archives of Neurology and Psychiatry, 1958, 80, 86-92.

Kennedy, E. On becoming a counselor. New York: Seabury Press, 1977.

Kidd, A. H. Ego Strength scale and the Goldberg scale in evaluating clinic outpatients. Journal of Clinical Psychology, 1968, 24, 438-439.

King, G. F., & Schiller, M. Note on ego strength, defensiveness, and acquiescence. Psychological Reports, 1958, ~' 434.

King, G. F., & Schiller, M. Ego strength and type of defensive be havior. Journal of Consulting Psychology, 1960, 24, 215-217.

Kleinmuntz, B. Extension of the construct validity of the Ego Strength scale. Journal of Consulting Psychology, 1960, ~. 463-464.

Kohut, H. Analysis of the self. New York: International Universities Press, 1971.

Kohut, H. Restoration of the self. New York: International Univer sities Press, 1977.

Korman, M. Ego strength and conflict discrimination. Journal of Con sulting Psychology, 1960, ~. 294-298.

Kotkov, B. Favorable clinical indications for group attendance. Inter national Journal of Group Psychotherapy, 1958, ~' 419-427.

Kotkov, B., & Meadow, A. Rorschach criteria for predicting continu ation in individual psychotherapy. Journal of Counseling Psych ology, 1953, !l, 16-20.

Kreitman, N. Reliability of psychiatric diagnosis. Journal of Mental Science, 1961, 107, 876-886. 240

Lesser, W. Relationship between counseling progress and empathic under standing. Journal of Counseling Psychology, 1961, ~. 330-336.

Levine, D., & Cohen, J. Symptoms and ego strength as predictors of the outcome of hospitalization in functional psychoses. Journal of Consulting Psychology, 1962, ~. 246-250.

Lewinsohn, P. M. Dimensions of MMPI change. Journal of Clinical Psy chology, 1965, ~. 37-43.

Liebling, B. A., Seiler, M., & Shaver, P. Self-awareness and cigarette smoking behavior. Journal of Experimental Social Psychology, 1974, lQ_, 325-332.

Liebling, B. A., & Shaver, P. Evaluation, self-awareness, and task per formance. Journal of Experimental Social Psychology, 1973, ~. 297-306.

Lorr, M., Katz, M. M., & Rubenstein, E. A. Prediction of length of stay in psychotherapy. Journal of Consulting Psychology, 1958, 22, 321-327.

Lorr, M., McNair, D. M., Michaux, W. W., & Raskin, A. Frequency of treatment and change in psychotherapy. Journal of Abnormal and Social Psychology, 1962, ~. 281-292.

Lubarsky, L. Clinician's judgements of mental health. Archives of General Psychiatry, 1969, 33, 750-756.

MacKinnon, R. A., & Michels, R. Psychiatric interview in clinical prac tice. Philadelphia: Saunders, 1971.

Malzer, R. L. Clients' pretherapy expectations and preferences, and their relationship to early termination of individual psychother- ~- Unpublished doctoral dissertation, Northwestern University, 1980.

Mandler, G., & Sarason, S. B. Study of anxiety and learning. Journal of Abnormal and Social Psychology, 1952, ir• 166-173.

Marlowe, D., & Crowne, 0. P. Approval motive. New York: Wiley, 1964.

Martin, P. J., Moore, J. E., Sterne, A. L., & McNairy, R. M. Therapists prophecy. Journal of Clinical Psychology, 1977, 33, 254-258.

McAdoo, W. G., & Roeske, N. A. Comparison of defectors and continuers in a child guidance clinic. Journal of Consulting and Clinical Psychology, 1973, 40, 328-334. 241

McDonald, P. J. Reactions to objective self-awareness. Dissertation Abstracts International, 1977, 12, 3679B-3680B.

McGuire, R. J. Classification and the problem of diagnosis. In H. J. Eysenck (Ed.), Handbook of abnormal psychology. San Diego, Cali fornia: Knopf, 1973, 3-33.

McNair, D. M., Callahan, D. M., & Lorr, M. Therapist 'type' and patient response to psychotherapy. Journal of Consulting Psychol- £81., 1962, 26, 425-429.

McNair, D. M., Lorr, M., & Callahan, D. H. Patient and therapist in fluences on quitting psychotherapy. Journal of Counseling Psychol- ~· 1963, ~. 10-17.

Head, G. H. Mind, self and society. Chicago: University of Chicago Press, 1934.

Meltzoff, J., & Kornreich, H. Research in psychotherapy. New York: Atherton Press, 1970.

Mendelsohn, G. A. Effects of client personality and client-counselor similarity on the duration of counseling: a replication and extension. Journal of Counseling Psychology, 1966, Jl, 228-234.

Mensh, D. N., & Golden, J. H. Factors in psychotherapeutic success. Journal of Hissouri State Hedical Association, 1951, 48, 180-184.

Hiller, I. W., & Magaro, P. A. Toward a multivariate theory of person ality styles, measurement and reliability. Journal of Clinical Psychology, 1977, ]i, 460-466.

Hillon, T. H. Hanual for the Hillon Multiaxial Clinical Inventory. Hinneapolis: Interpretive Scoring Systems, 1977.

Horse, S., Gergen, K. Social comparison, self-consistency, and the concept of self. Journal of Personality and Social Psychology, 1970, li· 264-273. Mozdzierz, G. J., Macchitelli, J. F., Conway, J. A., & Krauss, H. H. Personality characteristic differences between alcoholics who leave treatment against medical advice and those who don't. Journal of Clinical Psychology, 1973, 29, 78-82.

Myers, J. K., & Schaffer,. L. Social stratification and psychiatric practice: a study of an outpatient clinic. American Sociological Review, 1954, J1, 307-310. 242

Nash, E. H., Frank, J.D., Gliedman, L. H., Imber, S.D., & Stone, A. R. Some factors related to patients remaining in group psychotherapy. International Journal of Group Psychotherapy, 1957, z, 264-274. Nie, N.H., Hull, C. H., Jenkins, J. G., Steinbrenner, K., & Bent, D. H. Statistical package for the social sciences. New York: McGraw Hill, 1975.

Ornstein, R. E. Psychology of consciousness. New York: Penguin, 1972.

Otis, A. S. Manual of directions for Gamma Test. New York: Harcourt, Brace, and World, 1954.

Overall, B., & Aronson, H. Expectations of psychotherapy in patients of lower socioeconomic class. American Journal of Orthopsychiatry, 1963, 33, 421-430.

Parloff, M. B., Waskow, I. E., & Wolfe, B. E. Research on therapist variables in relation to process and outcome. In S. L. Garfield & A. E. Bergin (Eds.), Handbook of psychotherapy and behavior change. New York: Wiley, 1978.

Pasternak, L. Self-acceptance and objective self-awareness. Disser tation Abstracts International, 1978, ]l,- 3856B.

Pavia, A. Imagery and verbal processes. New York: Holt, Rinehart & Winston, 1971.

Pervin, L. A. Personality: theory, assessment, and research. New York: Wiley, 1970.

Pfuetze, P. E. Self, society, existence. New York: Harper, 1954.

Piaget, J. Construction of reality in the child. New York: Basic Books, 1954.

Piaget, J. Judjment and reasoning in the child. Totowa, New Jersey: Littlefield & Adams, 1966.

Pryor, J. B., Gibbons, F. X., Wicklund, R. A., Fazio, R. , & Hood, R. Self focused attention and self-report validity. Journal of Per- sonality, 1977, 45, 513-527.

Quay, H. Performance of hospitalized psychiatric patients on the Ego Strength scale of the MMPI. Journal of Clinical Psychology, 1955, l!_, 403-405.

Rapaport, E. G. Lower class patient attitudes and expectations related to· dropout from psychotherapy. Dissertation Abstracts Inter- national, 1976, 36, 6397B. 243

Reschke, S. E. Mental health factors in college students. Disser tation Abstracts International, 1967, ~. 1251B.

Robinson, H. A., Redlich, F. C., & Myers, J. K. Social structure and psychiatric treatment. American Journal of Orthopsychiatry, 1954, 24, 307-:-316.

Rogers, C. R. Process conception of psychotherapy. American Psychol ogist, 1958, 11. 142-149.

Rogers, C. R. On becoming a person. Boston: Houghton Mifflin, 1961.

Rogers, C. R., & Dymond, R. F. (Eds.). Psychotherapy and personality change. Chicago: University of Chicago Press, 1954.

Rogers, L. S., Knauss, J., & Hammond, K. R. Predicting continuation in therapy by means of the Rorschach Test. Journal of Counseling Psychology, 1951, 12. 368-371.

Roos, P. Performance of psychiatric patients on two measures of ego strength. Journal of Clinical Psychology, 1952, .!..§_, 48-50.

Rosen, A. Diagnostic differentiation as a construct validity indicator for the MMPI Ego Strength scale. Journal of General Psychology, 1963, ~. 293-297.

Rosenthal, D., & Frank, J. D. Fate of psychiatric clinic outpatients assigned to psychotherapy. Journal of Nervous and Mental Disease, 1958, 127, 330-343.

Rosenzweig, S. P., & Folman, R. Patient and therapist variables affect ing premature termination in group psychotherapy. Psychotherapy: Theory, Research and Practice, 1974, ll• 76-79.

Rubenstein, E. Q., & Lorr, M. Comparison of terminators and remainers in outpatient psychotherapy. Journal of Clinical Psychology, 1956, Q, 345-349.

Saltzman, C., Luetgert, M. J., Roth, C. H., Creaser, J., & Howard, L. Formation of a therapeutic relationship: experiences during the initial phases of psychotherapy as predictors of treatment duration an outcome. Journal of Consulting and Clinical Psychology, 1976, 44, 546-555.

Sartre, J. P. No exit and three other plays. New York: Knopf, 1946.

Sartre, J. P. Being and nothingness. New York: Philosophical Library, 1956. 244

Schaffer, L., & Myers, J. K. Psychotherapy and social stratification. Psychiatry, 1954, !l, 83-93.

Scheer, N., & Barton, G. M. Comparison of patients discharged against medical advice with a matched control group. American Journal of Psychiatry, 1974, 131, 1217-1220.

Scheier, M. F. Self-awareness, self-consciousness, and angry aggres sion. Journal of Personality, 1976, 44, 627-644.

Scheier, M. F., Buss, A. H., & Buss, D. M. Self-consciousness, self report of aggressiveness and aggression. Journal of Research in Personality, 1978, Jl, 133-140.

Scheier, M. F., & Carver, C. S. Self focused attention and the ex perience of emotion: attraction, repulsion, elation and depres sion. Journal of Personality and Social Psychology, 1977, 35, 625-636.

Scheier, M. F., Carver, C. S., Schulz, R., Glass, D. C., & Katz, I. Sympathy, self-consciousness, and reactions to the stigmatized. Journal of Applied Social Psychology, 1978, ~. 270-282.

Sheehan, D., & Surman, 0. Hypnosis and psychosomatic medicine clinic. InT. P. Hackett & N. H. Cassem (Eds.), Massachusetts General Hospital handbook of general hospital psychiatry. St. Louis: Mosby, 1978, 536-545.

Sherry, R. Patients' and therapists' differential expectations of psychotherapy as correlates of patient dropout. Dissertation Abstracts International, 1977, 38, 380B-381B.

Silverman, J. Validity of the Barron Ego Strength scale in an individ ual form. Journal of Consulting Psychology, 1963, ~. 532-533.

Sloane, R., Cristal, A., Pepernick, M., & Stapler, F. Role preparation and expectancy of improvement in psychotherapy. Journal of Nervous and Mental Disease, 1970, 150, 18-26.

Smith, M. B. Perspectives on selfhood. American Psychologist, 1979, 11_, 1053-1063.

Snyder, M. Self-monitoring and expressive behavior. Journal of Person ality and Social Psychology, 1974, 30, 526-537.

Spiegel, D. SPI discriminators among four psychological health-sick ness levels. Journal of Consulting and Clinical Psychology, 1969, 33, 750-756. 245

Spielberger, C. D., Gorsuch, R. L., & Lushene, R. E. Manual for the State-Trait Anxiety Inventory. Palo Alto, California: Consul ting Psychologists Press, 1970.

Spitzer, R. L., & Fleiss, J. L. Reanalysis of the reliability of psychiatric diagnosis. British Journal of Psychiatry, 1974, 125, 341-347.

Spitzer, R. L., Fleiss, J. L., Endicott, J., & Cohen, J. Mental Status schedule. Archives of General Psychiatry, 1967, ~. 479-493.

Stein, K. B., & Chu, C. L. Dimensionality of Barron's Ego Strength scale. Journal of Consulting Psychology, 1967, 11, 153-161.

Stern, S. L., Hoare, S. F., & Gross, S. J. Confounding of personality and social class characteristics in research on premature termin ation. Journal of Consulting and Clinical Psychology, 1975, 43, 341-344.

Straker, M., Davanloo, H., & Moll, A. Psychiatric clinic dropouts. Laval Medicine, 1967, 38, 71-77.

Strickland, B. R., & Crowne, D. P. Need for approval and the premature termination of psychotherapy. Journal of Consulting Psychology, 1963, 12· 95-101.

Sullivan, H. S. Psychiatric interview. New York: Norton, 1954.

Sullivan, P. L., Miller, C., & Smelser, W. Factors in length of stay and progess in psychotherapy. Journal of Consulting Psychology, 1958, ~. 1-9.

Taft, R. Validity of the Barron Ego-Strength scale and the Welsh Anxiety Index. Journal of Consulting Psychology, 1957, ll• 247- 249.

Tamkin, A. S. An evaluation of the construst validity of Barron's Ego Strength scale. Journal of Clinical Psychology, 1957, }1, 156-158.

Tamkin, A. S., & Klett, C. J. Barron's Ego-Strength scale: a replica tion of an evaluation of its construct validity. Journal of Con sulting Psychology, 1957, ll• 412.

Taulbee, E. S., Relationship between certain personality variables and continuation in psychotherapy. Journal of Consulting Psychology, 1958, ~. 83-89.

Taylor, J. W. Relationship of success and length in psychotherapy. Journal of Consulting Psychology, 1956, 20, 332. 246

Teter, J. T., & Dana, R. H. Construct validation of the Barron Ego Strength scale. Psychological Reports, 1964, 12, 525-526.

Tolman, R. S., & Meyer, M. M. Who returns to the clinic for more ther apy? Mental Hygiene, 1957, ~, 497-506.

Trice, H. M., & Roman, P.M. Sociopsychological predictors of affilia tion with Alcoholics Anonymous. Social Psychiatry, 1970, 1, 51-60.

Turner, R. G. Self-consciousness and anticipatory belief change. Per sonality and Social Psychology Bulletin, 1977, 1, 438-441.

Turner, R. G. Consistency, self-consciousness, and the predictive validity of typical and maximal personality measures. Journal of Research in Personality, 1978, Jl, 117-132.

Turner, R. G., & Keyson, M. Relationships between client self-percep tions of self-consciousness levels and therapist awareness of these perceptions. Journal of Consulting and Clinical Psychology, 1978, ~. 1586-1587.

Turner, R. G., & Peterson, M. Public and private self-consciousness and emotional expressivity. Journal of Consulting and Clinical Psychology, 1977, 45, 490-491.

Turner, R. G., Scheier, M. F., Carver, C. S., & Ickes, W. J. Correlates of self-consciousness. Journal of Personality Assessment, 1978, ~. 285-289.

Von Evra, J.P., & Rosenberg, B. G. Ego strength and ego disjunction in primary and secondary psychopaths. Journal of Clinical Psychol ~, 1963, ~. 61-63.

Watson, C. G. Prediction of length of hospital stay from the MMPI scales. Journal of Clinical Psychology, 1968, 24, 444-447.

Webster's new collegiate dictionary. Springfield, Massachusetts: Mer rian, 1977.

Weiss, J. M. A., & Schaie, K. W. Factors in patients' failure tore turn to clinic. Diseases of the Nervous System, 1958, ~. 429-430.

Welkowitz, J. Behavior patterns in group psychotherapy sessions in two Veterans Administration hospitals. Dissertation Abstracts Inter national, 1960, 20, 4202-4203.

Wicklund, R. A. Discrepancy reduction or attempted distraction? a reply to Liebling, Seiler, and Shaver. Journal of Experimental Social Psychology, 1975a, 1!, 78-81. 247

Wicklund, R. A. Objective self-awareness. In L. Berkowitz (Ed.), Ad vances in social psychology. New York: Academic Press, 1975b, 233-275.

Wicklund, R. A., & Duval, S. Opinion change and performance facilita tion as a result of objective self-awareness. In S. Duval & R. Wicklund, Theory of objective self-awareness. New York: Academic Press, 1972, 168-171.

Wilkins, W. Expectancy and therapeutic gain: an empirical and concep tual critique. Journal of Consulting and Clinical Psychology, 1973, 40, 69-77.

Williams, H. L., & Lawrence, J. F. Comparison of the Rorschach and MMPI by means of factor analysis. Journal of Consulting Psychol ~· 1954, ~. 193-197.

Williams, R., & Pollack, R. R. Some nonpsychological variables in therapy defection in a child guidance clinic. Journal of Psy chology, 1964, 58, 145-155.

Winder, A. E., & Hersko, M. Effect of social class on the length and type of psychotherapy in a Veterans Administration mental hygiene clinic. Journal of Clinical Psychology, 1955, _!_!_, 77-79.

Winder, C. L., Ahmad, F. Z., Bandura, A., & Rou, L. C. Dependency of patients, psychotherapists' responses and aspects of psychotherapy. Journal of Consulting Psychology, 1962, ~. 129-134.

Wine, J. Test anxiety and direction of attention. Psychological Bulle t in, 19 71 , 1!!._, 92-104 •

Winters, W. D., & Stortroen, M. Comparison of several MMPI indices to differentiate psychotics from normals. Journal of Clinical Psych ology, 1963, ~. 220-223.

Wirt, R. D. Further validation of the Ego Strength scale. Journal of Consulting Psychology, 1955, ~. 444.

Wirt, R. D. Actuarial prediction. Journal of Consulting Psychology_, 1956, 20, 123-124.

Wolfe, G. R. Role of objective self-awareness and field independence dependence in the induction of elation and depression. Disserta tion Abstracts International, 1975, ~. 4205B.

Yalom, I. D. Study of group therapy dropouts. Archives of General Psychiatry, 1966, ~' 393-414. 248

Yalom, I. D. Theory and practice of group psychotherapy. New York: Basic Books, 1975.

Yamamoto, J., James, Q. C., Bloombaum, M., & Hattem, J. Social factors in patient selection. American Journal of Psychiatry, 1967, 124, 84-90.

Zajonc, R. B. Social facilitation. Science, 1965, 149, 269-274.

Zubin, J. Classification of the behavior disorders. Annual Review of Psychology, 1967, ~· APPENDIX A

249 250

Following is a list of statements about personal and social behavior. Please read each statement, thinking about how true for you each statement is. If the statement is "extremely unlike" you, then circle the number- "1" next to that statement. If it is "somewhat unlike" you, then circle the number "2''. If you are uncertain or undecided, circle "3"; if the statement is "somewhat like" you, circle "4"; and circle "5" for those statements that are "extremely like" you. Take as much time as you like on each question. Remember, there are no right or wrong answers. What matters is how you are, or how you see yourself to be. Please complete the questionnaire before leaving. When finished, turn it in to the secretary or the staff member that gave it to you.

1 2 3 4 5 Extremely Somewhat Undecided Somewhat Extremely Unlike Me Unlike Me or Unsure Like Me Like Me

1. I'm always trying to figure 1 2 3 4 5 myself out 2. I'm concerned about my style 1 2 3 5 of doing things 3. Generally I'm not very aware 1 2 3 4 5 of myself 4. It takes me time to overcome my 1 2 3 4 5 shyness in new situations 5. I ret1ect (or think carefully) about 1 2 3 4 5 myself a lot 6. I'm concerned about the way I 1 2 3 5 present myself (or come across to others)

Extremely Somewhat Undecided Somewhat Extremely Unlike :Yle Unlike ~le or Unsure Like :Yie Like Me

7. I'm often the subject (or center) 1 2 3 4 5 of my own fantasies s. I have trouble working when 1 2 3 5 someone is watching me 9. I never scrutinize myself (or 1 3 4 5 examine or look at myself closely) 10. I get embarrassed very easily 1 2 5

11. I'm self-conscious about the way 1 2 3 I look

Continue on the next page. 251

1 2 3 4 5 Extremely Somewhat TJndecided Somewhat Extremely Unlike Me Unlike Me or Unsure Like Me Like Me

12. I don't find it hard to t:llk. to 1 2 3 4 5 str:mgers 13. I'm generally attentive to (or 1 2 3 4 5 aware of) my inner_feelings 14. I usually worry about makin!$ 1 2 3 4 5 a good impression 15. I'm constantly examining my 1 2 3 4 5 motives (or looking at my reasons for acting as I do) 16. I feel anxious when I speak in 1 2 3 4 5 front of a group 17. One of the last things I do before 1 2 3 4 5 I 1e ave my house is look in the mi.F!'or

Extremely Somewhat Undecided Somewhat Extremely Unlike Me Unlike Me or Unsure Like iVIe Like Me

18. I sometimes have the feeling 1 2 3 4 5 that I'm off somewhere watching myself 19. I'm concerned about what other 1 2 3 4 5 people think of me 20. I'm alert to changes in my mood 1 2 3 4 5

21. I'm usually aware of my appearance 1 2 3 4 5

22. I'm aware of the way my mind works 1 2 3 4 5 - when I work through a problem 23. Large groups make me nervous 1 2 3 4 5

You may be contacted by telephone for a brief, follow up interview about your reactions to your treatment at Hines. If there are times of the day or days of the week that are particularly convenient, please indicate them below: days ______times ------

Thank you for your cooperation 252

This inventory consists of numbered statements. Read each statement and decide whether it is true as applied to you or false as applied to you.

If a statement is TRUE or MOSTLY TRUE, as applied to you, place a check mark V) in the column marked True. If a statement is FALSE or NOT USUALLY TRUE, as applied to you, place a ckeck mark in the column marked false. If a statement does not apply to you or if it is something that you don't know about, make no mark.

Remember t~ give YOUR OWN opinion of yourself. Do not leave any blank spaces if you can avoid it. Blacken only one response for each statement. Erase completely any answer you wish ;o change. Remember, try to make~ answer to each statement. True False 1. I have a good appetite. 2. I have diarrhea once a month or more. 3. At times I have fits of laughing and crying that I cannot control. 4. I find it hard to keep my mind on a task or job. 5. I have had very peculiar and strange experiences. 6. I have a cough most of the time. 7. I seldom worry about my health. 8. My sleep is fitful and disturbed. 9. When I am with people I am bothered by hearing very queer things. 10. I am in just as good physical health as most of my friends.

True False 11. Everything is turning out just like the prophets in the Bible said it would. 12. Parts of my body often have feelings like burning, tingling, crawling, or like "going to sleep." 13. I am easily downed in an argument. 1-L I do many things that I regret afterwards (I regret things more or more often than others seem to). 15. I go to church almost every week. · 16. I have met problems so full of possibilities that I have been unable to make up my mind about them. 17. Some people are so bossy that I feel like doing the op posite of what they request, even though I know they are right.

Continue on the other side. 2S3

True False 18. I like collecting flowers or growing house plants. 19. I like to cook. 20. During the past few years I have been well most of the time. 21. I have never had a fainting spell. 22. When I get bored I like to stir up some excitement. 23. My hands have tiot become clumsy or awkward. 24. I feel weak all over much of the. time. 25. I have no difficulty in keeping my balance in walking. 26. I like to flirt. 27. I believe my sins are unpardonable. 28. I frequently find myself worrying about something. 29. I like science. 30. I like to talk about sex. 31. I get mad easily and then get over it soon.

True False 32. I brood a great deal. 33. I dream frequently about things that are best kept to myself. 34. My way of doing things is apt to be misunderstood by others. 35. I have had blank spells in which my activities were interrupted and I did not know what was going on around me. 36. I can be friendly with people who do things which I consider wrong. 37. If I were an artist I would like to draw flowers. 38. When I leave home I do not worry about whether the door is locked and the windows closed. 39. At times I hear so well it bothers me. 40. I often cross the street in order not to meet someone I see. H. I have strange and peculiar thoughts.

Continue on the next page. 254

True False 42. Sometimes I enjoy hurting persons I love. 43. Sometimes some unimportant thought will run through my mind and bother me for days. 44. I am not afraid of fire. 45. I do not like to see women smoke. 46. When someo.ne says silly or ignorant things about something I know about, I try to set him right. 47. I feel unable to tell anyone all about myself. 48. My plans have frequently seemed so full of difficulties that I have had to give them up. 49. I would certainly enjoy beating a crook at his own game. 50. I have had some very unusual religious experiences. 51. One or more members of my family is very nervous. 52. I am attracted by members of the opposite sex. 53. The man who had most to do with me when I was a child {such as my father, stepfather, etc.) was very strict with me. 54. Christ performed miracles such as changing water into wine.

True False

55. I pray several times each week. 56. I feel sympathetic towards people who tend to hang on to their griefs and troubles. 57. I am afraid of finding myself in a closet or small closed space. 58. Dirt frightens or disgusts me. 59. I think Lincoln was greater than Washington. 60. In my home we have always had the ordinary necessities (such as enough food, clothing, etc.). 61. I am made nervous by certain animals. 62. My skin seems to be unusually sensitive to touch. 63. I feel tired a good deal of the time. 64. I never attend a sexy show if I can avoid it.

Continue on the next page. 255

True False 65. If I were an artist I would like to draw children. 66. I sometimes feel that I am about to go to pieces. 67. I have often been frightened in the middle of the night. 68. I very much like horseback riding.

Thank you for your cooperation. 256

Patient's Name ------Refused Dare ______Before Afrer

Referred to MHC from ------Preresred

Diagnostic Impression and Manifestations------

Personality Organization (mark '1' next to the most appropriare description; place "2" next to the secondary care gory, if needed) Hysrerical: overuse of repression; likely presents with somatic concerns or the feeling of being used in personal relationships; exhibitionistic, dramatic; relares so that the therapist has a sense of engagement with the patient despire a focus on external events. Paranoid: projection prominent; presents with concerns about being isolated or deliberarely abused by other people; relares vaguely and guardedly; therapist feels a strong pull to confirm the patient's perspective or else be confronred with the patient's underlying anger. Obsessive: isolation and inrellectualization prominent; problems relared to feared loss of control or, if loss of control has been experienced, feelings of depression, anger, and guilt; high expectations for personal productivity or morality; insight into thoughts and motives without affect; patient is controlling, and the inrerview may be so wordy as to become boring. Passive: schizoid and avoidant; may present with multiple fears and feelings of helplessness; inward focus on fantasy; history suggests lack of social skills and contacts; therapist may still feel detached afrer the inrerview; include: passive dependent, passive aggressive, and narcissistic characters. Sociopathic: history of acting out rensions; minimizes problems, and likely attends the inrerview because of pressure from family, courts, or other outside source; may be problems with alcohol or drugs; appears socially skilled and comfortable, and yet the contact remains superficial and the relationship feels tenuous.

Prognosis (choose among: excellent, good, fair, guarded, poor)------

Patient's Motivation for Tr:eatment (excellent, good, fair, poor)------

Treatment Modality Recommendation Patient's Preference (rank order those (mark"+", lf0", or "-" considered) for those recommended) Individual Therapy Group Therapy (individual or couples) lV1arital or Family Therapy Hypnotherapy lVIedication (exclusively) Discharge: with referral elsewhere Discharde: no treatment desired Other: ______257

Checkliat for Intake Interviewers

Dorothy and Dagmar will administer the questionnaire to all patients who have nor. been given the scale as inpatients or on screening.

Ask all intakes to turn the questionnuire in at the beginning of your interview. They may have already completed the questionnaire and turned it in at the receptionists desk. If the'y have not finished it, remind them to complete it before leaving the clinic. They can turn the form in to Dorothy or Dagmar. If completed, the form can be turned in to you and included in the intake folder. Thank the patient for his cooperation on behalf of the research team.

After the interview, complete the Intake Rating Sheet: copy from the intake sheet: name, date, referra 1 source, diagnosis, and prognosis; rank the relevant personality typologies (remembering that this is not to indicate degree of pathology, only personality structure), rate motivation for treatment, and rank recommendacions (noting the patient's reactions with+, 0, or -).

Place the Inta~e Rating Form in the iutake folder.

Thank you for your continued help. 258

CLINICAL RECORD Report on or------Continuation of s. F.-:-:-:--:----:---:--:---:--:-----

(Sifln and date)

INFORl\1ATION ABOUT: FACTORS L~FLUENCING THE DECISION TO BEGIN OR CONTINUE IN OUTPATIENT TREATI.IENT

The purpose of this research is to better understand what factors a veteran considers in malting a decision about beginning or continuing in tre stmevt at the :O.lental Hygiene CUnic. We will be interested 1 n characteristics of the veteran and in hospital policies that affect this decision. You will be asked a sel·ies of questions on the following pa:;es about yourself and your experieuces. f.t some time in the next four months you may receive a follow up telephone call irom a research staff person, who will ask you some brief questions about your re~ctions to your treat ment in the Psychiatry Service at Hines. In addition, the intal;;e worker will auswer several qut:~stions about your interview.

In this research, our goal is to discover facts about people and programs so r.hat the !VIental Hygiene Clinic can best meet veteran:;' r1eeds. In this kind of study, we are less interested in a particular person'3 score on the paper and pencil questionnaire than in the responses of veterans as a group. So no individttal case is ever identified. Your test results are stricti:• confidrmtial. While the forms you complete wi.ll remain in your clinical record during the coursA of the study, they will not be interpreted by staff. Your reSiJOnses to e1e q~,;astionnaire will not affect the type of treatment you receive at the ~rental Hygiene Clinic. So the be~1ifits of this research will help those who come after you, a3 your parth.:it:~ation will hopefully make it possible for veterans in the future to more efficiently get the help they desire from the r..Iental Hygiene Clinic.

Our hope is that your participation m this re:;earch will be ::.. comfo1:table experience for ycu. Many others have filled out the CJUestionnail'e with no problema. !f you do have questions or concerns about this reseru:ch, you can feel free to discuss them with a ward psychologist or ;\!ental Hygiene Clinii.: staff member. If you sustain any physical injury related to your participation in this study you will be entitled to medical c&e and treatment. Compensation may also be payable under 38 USC 351, or in some circumstances, undel' the Federal Tort Cl'lim~ Act.

(Contin'!Je on :ereu~ !ic!e)

TAi:i.::NT"S iDENTIFICATION (Fat' typ~:l or writreon lftttri~u AH•e: Nam-!:..s!, forst. WAftDt«::. rn1 ddJc; ~r•d•; d;,jtr:; hospitttl or tn~du:.·li (tlc.ili~Ti I

P~PO~T ON--- or COMTIH~A'fiOM I)F ----- ~tunf.l.:\l"d .Form :urr a.t .. flt.-1. St.rvlC!S ·U'IIIUhtt••ttOiol ~ItO I!OITI:~.Io'~tlriiCT (:OIUIIITTtl C'l't WIO"t:Af. ltt:CIWIU 259

CLINICAL RECORD Report on or------Continuation of S. F •.-:-:;-;:-~----=-~::-.:---:--:------

(Sitn and dat~)

INFORlV!ATION ABOUT: FACTORS INFLUENCING THE DECISION TO BEGIN OR CONTINUE IN OUTPATIE:t\T TREATMENT

Your participation in this testing project is strictly voluntary. You may withdraw your consent' at any time by stating tha~ you will not complete t!m questionnaire. Your decision to participate or not will not affe::t the treatment that rou wculd normally receive at the Mental Hygiene Clinic now or in the future. Your veteran's benefits will not be affe·:;ted by yC'ur test results or your decision to particif)&te or not.

I •.mderstand that this questionnaire is for reseaJ.•ch purposes. I agree to fill out this form and, perhaps, to have a short telephone interview in the future. I uncte1·stand that the reults will not affect my curreut treatment at Hiue~ and thc>t tl:wse results will be kept stl·ictly confidential by the research staff aad by all other staff at Hines Veterans Administratic::n 1Heclical Centc:·.

FJ'incipal Investigator's Signature Patient's 'Signature

Responsible Therapist's Signature \Vitness

Date Time

Continue on the next page.

(Cor tlnue on r~vorae tu'd,}

~A ;.~NT'S ICZNT:I"~ICATION (ror ryp~d or written •nt'i~s ~rv.ft: •'"'f'!'O I•H. first. I RF.:QiST:!R NO. tnid.J,"e; 'radt~: date: hc~pital or f'r.r!!d•~allacility) IWAI!l)NQ. I RE?OHT 011 ----- or Cvh"Tllltl.\T!O!t OF -----

&t)ttiJ.Il 5i•v•.:u •o .. '"t.sr.u,p,., A~~ro ,._;r"•C~>tCT co••rnrt ()..,. '~r~iCAt. •t:c•~:s ,.., 'tl It' ·I; :.~I t APPENDIX B

260 ~~~2~~~3~~4-,~5~~6-,~7~~8_,~9~~10~,_1~1~~1~2~~1~3~~14 ~1=5-,-:.1~6~~17~~1~8~~1~9~~2~0- _A ______~---+---+---+-~-~~l--4'-3--~--~--~--+-4--~--4--4-l-----t--'~~ j:__j'---~-3---+-4---+_s __ _

: : j : 2 : 4 I : 4 < 1 - -++-_1 ]-r'--4--1 --4:_-F-+2'_ ·----11,1-44--+-l--t-)·---ti·_2_5_ 1 n 5 2 J ,, i 2 -+=- 3 3 3 5 1 _ ~--+-5--+-~--+ --t----+-1--+_s_4 _2__ _ : -:r=i I' I, ~: : 1: : _:__~_:__ ---j,_:__ +-_: __ -+:----11-:--+_:_-+ __: __ +_:__ c s 4 2 4 4 J 3 1 Is 2 1 js 1 4 1 4 2 3 ,, 1 --+---+---+---~-----t-·---+'--1---l----+---·+ ll 4 1 4 2 2 4 1 1 -J------;--;-Jt; 14 1 , 4 I, 1 1 ~t'-4---t_4_:>_+_s_4_+-14__ 3-+4 ---1-~-4-S-!-l-5-l-2 J-+-2--+2----1-4-·

K s tjt, ; s Is 1. _ __j 3 4 s I ~-L..l' __ji _ ___j__ _j _ _,__~ _ __,\c___

* l • Hysteric, 2 ~ Paranoid, 3 "' ObBessive, 4 ~ Passiv<>, 5 = Sociopathic .!'ltl!CENT Ar.REENENT flN APPLICARILITY OF EACH GIVEN _!'(?Z PRIMARY .TIFE ONLY PAT~'{

2 3 4 5 6 1 8 9 10 ll li 13 14 !.5 16 17 18 1.9 20 % Ar.ltEE~IENT

Hysteri 100 l1oa '10o !10o too 71.4 ?S 100 71.4100 62.5100. 62.5 75 87.5100 100 T62.s1oo 100 89.3 -----~--+--f---+-·--1·-- --f-· . Paranoi teo I 10 6o j1oo 85.1 1oc 100 100 100 100 100 100 100 87 .s 75 I 75 S7 .s 100 100 62.5 89.3 1 OhSPSS ~---;0-;;-~ 70 ~ 100 71.'~ 75 185.7 71.'• 100 62.5 75 100--100 87.5 75 87.5 75 87.5100 - 86.2

P.:tssive . 10± ;~ _1_6~- ~~ 100 ! 85.7r-;-~~oo 7'!1 75 87.5 62.5 100 100 75 B7.5 75 100 ·- !l/1, 9 Soe iop::t ;;--~J!-~_j -;;-r~~~~~-~;0 t 1~~- _10~ 1~-;· 100 10~5 75 100 100 _ 100 100 6:. 5 62.5 91.2

!'!~ l'!ARY AND 88.2 ~~:_s;_y~p_ARY TYPE

2 3 4 'i 6 7 8 9 10 11 12 13 111 15 16 17 !8 19 20 --- 87.5,100 100 62.5,100 100 86.8 ,-;,-i-;-oo to-;, 8o jwo_ J 7l -: J 75 57._: 1lC)."__ . 62 · 5 too '""'"-,_~= !_ f j I"~ ;·;--r;;-r--· -- Pm~td 100 160 60 lwo 1 85.7~~ :oo '._~ 85.7 too , too 10_' J85.7J 1 J 87.5 1100 1I 87.s 62.5 37.4 Obsec:sive 90 llO . 70 ~c 100 I 57.1 100 100. 1 57.1 100 1 50 75 100 poo 75 ::.._j 87. s j 75 1 ai. ~too__ _ 83.0 ''"";;,_-- lOO _ ~~f· _--;_,__j_ ~;;;- ~':.. __;;:~r;-;-:-: 85.,1 , -;5 .,.5 ... 5 100 .100 ~ 75 I 87 •.:1._1s ..1!~ 83.0 Sociopath 100 lOG 70 90 100 1~ ltoo jtoo 100 62.5 62.5 100 ·llVG 1100 1100162.51 87.5 91.2 ------·-- ______._

86.3

NUMHER OF' l :.' 3 4 56 7 f, 9 10 11 Ji 13 14 15 16 J 1 18 19 20 !U.T WGS -10 ~LiQiiJ 10 1 7 1 7 1. 4 T-7 CCI , 1 8 1 a 1 8 1 a 1 a Ls 'l'OThl. RATINGS ~ 159 lilsJ=8 I a N 0'1 N APPENDIX C

263 VARIMP...X RC'!'ATED FACTOR MATRIX FOR AU_, FACTORS WITH AN EIGENV .ALUE GREATER THAN ONE

VARIABLE FACTOR 1 FACTOR 2 F.ACTOR 3 FACTOR 4 FACTOR 5 FACTOR 6 FACTOR 7

1 .125 .709 .llO ~- .062 .121 .025 .280 2 .171 .353 .085 -.021 .153 .018 .083 3 -.113 -.027 -.017 .381 -.286 .351 .207 4 .643 .234 .083 -.005 -.149 -.548 .234 5 .186 .545 -.075 .265 .121 -.056 .112 6 .112 .416 .245 .139 .102 -.052 -.076 7 .268 .275 .224 .036 .359 .083 .011 8 .606 .164 .118 -.053 .229 -.004 -.017 9 .060 .089 -.041 .196 -.039 .486 -.018 10 .526 .231 .282 -.073 .212 -.160 .053 11 .153 .123 660 .079 .146 -.073 .145 12 .100 .027 .047 -.221 -.058 -.025 .1+39 13 .104 ,141 .036 .525 -o068 .054 -.093 14 .303 .325 .508 .011 -.148 -.044 -.243 15 .152 .593 .195 .085 .070 .124 -.104 16 .391 .182 .151 -.045 -"015 .068 -.053 17 .139 .158 .194 .087 .2l7 -.067 .000 18 .185 .253 .::>31 .063 .553 .008 -.060 19 .311 ,338 .454 .023 -.008 .094 .066 20 .025 .042 .101 .545 .267 .132 .030 21 -.099 --. 026 .381 .382 .199 -.019 -.042 22 -.096 .041 .023 .549 .011 .089 -.182 23 .730 .116 .003 .llO • 2.02 .069 .172

"PERCENT OF VARIANCE N INCORPORATED 9,79 8,97 6.10 6.10 4,20 3.40 2 • .50 0\ .P- APPENDIX D

265 266

DSM II DIAGNOSES OF RESEARCH SUBJECTS

Code Diagnosis Frequency

290.00 Senile dementia 1 291.50 Alcoholic deterioration 1 295.00 Simple schizophrenia 1 295.30 Paranoid schizophrenia 31 295.40 Acute schizophrenic episode 7 295.60 Residual schizophrenia 2 295.70 Schizo-affective schizophrenia 3 295.90 Chronic undifferentiated schizophrenia 3 295.99 Other schizophrenia 2 296.00 Involutional melancholia 5 296.20 Manic-depressive illness, depressed 1 296.30 Manic-depressive illness, circular 11 297.00 Paranoia 2 298.00 Psychotic depressive reaction 1 300.00 Anxiety neurosis 20 300.10 Hysterical neurosis 1 300.20 Phobic neurosis 3 300.40 Depressive neurosis 43 300.70 Hypochondriacal neurosis 1 301.00 Paranoid personality 2 301.10 Cyclothymic personality 1 301.20 Schizoid personality 2 301.30 Explosive personality 5 301.40 Obsessive compulsive personality 1 301.70 Antisocial personality 1 301.89 Borderline personality 1 302.80 Sexual deviation 1 303.10 Habitual excessive drinking 1 303.20 Alcohol addiction 7 304.80 Other drug dependence 1 305.50 Gastro-intestinal psychophysiologic disorder 2 307.30 Adjustment reaction to adult life 25 307.40 Adjustment reaction to late life 9 309.20 Non-psychotic organic brain syndrome due to trauma 1 309.90 Non-psychotic organic brain syndrome due to other physical condition 316.00 Marital maladjustment 4 318.00 No mental disorder 1 APPENDIX E

267 268

FORMAT OF RAW DATA Card Columns Content

1 1- 3 Patient identification number (Inpatients 000-106, outpatients 300-417, screenings 500-510, other out patients 600-611, ratings only 700-795) 1 4 Card number 1 6-28 Self-Consciousness scale items (on a scale of 1-5)a 1 30-31 Private self-consciousness 1 33-34 Public self-consciousness 1 36-37 Social anxiety 1 39-41 Total self-consciousness score 1 43 Data combination available (Self-Consciousness scale only 1, all instruments 8, etc.) 2 1- 3 Patient identification number 2 4 Card number 2 6- 8 Days between referral and scheduled intake appointment 2 9-11 Days between referral and first therapy appointment 2 12 Number of sessions before beginning therapy 2 13 Same therapist for treatment as for intake (no 1, yes 2) 2 14 Same therapist throughout first six therapy sessions (no 1, yes 2) 2 16-18 Patient age (as of January 1, 1980) 2 18-20 Distance between residence and clinic 2 21-23 Percent disabilitya 2 24 Previous therapy at the clinic (yes, all of 5 or less sessions 1; never began, although referred 2; never referred 3; one or more episodes of at least 6 sessions 4) 2 25 Sessions attended (6 represents 6 or more sessions) 2 26-27 Therapist 2 28 Therapist discipline (psychologist 1, scoial worker 2, nurse 3, physician 4) 2 29 Staff/student status (staff 1, student 2) 2 30-31 Years of experience of the therapist 3 1- 3 Patient identification number 3 4 Card number 3 6-11 Date of intake (year, month, day) 3 12 When the Self-Consciousness scale was completed (before interview 1, after interview 2, in waiting room prior to interview 3)

Note. Blanks or zeros are typically considered to be missing data. aBlanks or nines are considered to be missing data. 269

Card Columns Content

3 13 Referral source (Admitting 1, Screening/Admitting 2, Inpatient Psychiatry Units 3-7, Inpatient General Medicine 8, Outpatient General Medicine 9) 3 15-19 Diagnosis (DSM II) 3 20 Global diagnosis (psychotic 1, neurotic 2, personality disorder 3, organic brain syndrome 4, other 5) 3 21 Personality type (hysteric 1, paranoid 2, obsessive 3, passive 4, sociopath 5) 3 22 Personality type (secondary) 3 23 Prognosis (poor 1, guarded 2, fair 3, good 4, excel lent 5) 3 24 Motivation for treatment (poor 1, fair 2, good 3, excellent 4) 3 25 Recommended modality of treatment (individual therapy 1, group therapy 2, family therapy 3, hypnotherapy 4, medication only 5, referral 6, discharge without re ferral 7, other 8) 3 26 Recommended modality of treatment (secondary) 3 27 Actual therapy modality delivered 3 28 Patient's reaction to the actual modality of therapy (negative 1, neutral or no response 2, positive 3) 4 1- 3 Patient identification number 4 4 Card number 4 5 Inpatient unit 4 6-73 Ego Strength scale items (evidence of ego strength 1) 4 75-76 Total Ego Strength scale score 4 77 Test administered first (Ego Strength scale 1, Self Consciousness scale 2) 270 COOl ~~~~~512~252254522~22~~ 3! 27 ~9 077 4 2 3 0004lllOlOlOOlOl01lulOl00100l01lOlllll01ClOlllOOlOl010lOlOOllllOlll011ll! 411 0011 55555554555153555555535 48 33 25 lOb o 0012 017000000 ~oOll~OO~OCJOOOO 3 0014311JC00010Cl10~COllllCClll00010000000100100l010001ClC10000l1101000001 251 0021 44535,45535155555155,5~ 44 34 21 U99 6 0022 Ol902310205coososo3b012llO 3 00244110lll0lllllll0l000llllll0101110011101lll010111lll011vCOlll1lllll0lO 491 0041 555155515151555555~5551 50 35 10 09~ 6 00~2 0070141020330100CC4602ll13 3 0044411111101111101C101101111111111JllC101llllOl0011111llOll0ll0111110lll 531 0051 551~5525545255445455~55 41 35 24 100 4 2 3 00544llll0llllOO~lC'Ol0000lllOll0lll'lOCCOC1gl1C000000011000000l101110011 311 00~1 22320140222512543122443 27 15 19 061 6 0062 008000000 2701700C30COvOOO 3 0064oulll01100010lCC10l00Cv00110011l00110lC1ll010l1ll019101ll1lCOl01l1111 401 007! 53525553535153~54545!42 49 29 16 094 0 0072 007023202 3101000023031115 3 00744ll110111109Cll0101011.lvli90ll00111100llC1110101911110011lC1000llOll ~51 COol ~~444512~5525552215~~44 3o 3J 19 085 6 0082 oo7007022 340290lC~c042ll5 3 0084401011lu1l001011001000011001110001vCOOOlll010ll1l011110Cll1ClllllCll0 381 0091 4135444225~~44234455554 37 28 23 u8~ 6 0092 OOoOOOOCO 750C8CC020COOOOO 3 00944lllll1lll1001lCOlCOlll1100ill01100lOOC0111001111lOllOOC0100101100010 391 0101 44354535442234551144454 36 24 25 oe: 6 0102 Ol~OOOOOO 5499900C200CvOCO 3 OlC421ll00101191100vlll00011010101Jl0lOOCOCl010000100991110100110010110Cl 331 0121 4~4'5'22515425222022253 ~4 2o 14 074 6 0122 02oOCC100 2tOClOOC30000000 3 Ol24llOCOOllOOCOOOClOOCC~COlOl100lll00101000ll0l0011l1CCllCCOOCllvll00CvO 252 0141 55255541155535;55~5~43' 3~ 34 Zo 098 6 C1~2 021COOOOO ZSC1302020~00000 3 Ol44311010~v~OCClllGCOCC~OC111010ll00~1vOlll01l010l~1000llCCl111ll1100000 322 0!51 44~955il~l42111~1ll4451 30 20 Od 056 4 2 3 01!43lll00101111lllC10ClO~OlllOl01~01101CO~lll019lOll01010010lllOllOl01lO 402 Cl61 '55144515141515111~5;55 45 24 lJ J79 6 Cloi u~Jo~vouo 4300700G3ooooooo 3 Ol64210ll!lClllCllOCJ01CllllllllOJil000llllll10lvllOCC1l11010Clll010ll011 44Z 271 0172 033000200 2401200020000000 0173 7~071eo3 2954C15vl170C2 Ol74ll0ll1100lllll~JU01ullllll0100~11111010111010111110010001010101101110 43~ 0181 33153~15115515151151515 14 29 2b 069 t 0182 OllCOOOOO 370C80002000v000 3 Ol8421ll~Oll011llOOOl01100llv10lOv1lOv01001lllOlllOOOllOllOlOllllll000010 382 0191 44442414544232444141414 Zb 2b 22 074 6 0192 032000000 2502800030000000 3 - Ol942ll00011011011000010vlOlOOlOOl0l0Cll01l0lCOOl1l00001l101ll1011000000l 322 0201 ;44445445431445531545~2 39 29 20 0~8 6 0202 007Jl4l02 520ubOOC3202lll3 3 0204111lll10011lOlO~OOC010ll1ll10!000111lJ1111001J1100lll1010llll01Cll011 432 0211 43044433244~34444334434 33 2b 21 08G 4 2 3 02l431lllvl10llOllll1100lll~OOllulJlOOOll00lll0100ll100lOlll0991101010110 412 0221 ~15J~55J:3~5~3~5434~555 44 27 28 0~~ 4 2 3 JZ241~~1Cl~C~l~ll!.:~l~lCOJall~dC~JJil~l~l01~l~~JlllCl:ll0J1Jll~llllllU11 402 0231 44345541,3454;?12343255 41 2~ 19 OBe 4 2 3 0234ll1001100111COCOvllClOOv00010100000001010~000lOCOl11COllllllllllOlv01 302 0251 ~5224511:141!54~4155554 36 33 13 082 4 2 3 025421111110111111C000001l0lll011100111110010l0110110110100llllll01100ll0 442 0271 i551141152!1l2521214541 29 ~3 J8 060 6 C272 J~4000QOO 4Z00300020000000 3 027431Jl111Clll1lll0100010lll10~l0v1l01100llllv1l911110110100ll0101100100 442 0281 44544243434243455422555 40 25 22 087 b ~252 012000200 3b003~0C3~CCOOOO 3 02842llllllllllllllCll011lll1llllCOll1ll110Cll01011ll1lOll1~lOlll0lll1010 522 0291 ll4l2l214LlZ41•13112591 25 13 01 045 e 0292 Ol~0001CC 4'CC200030COOCOO 3 C2942ll?1011ll11lll010101llllll111llll0ll01l010l1lll111Cl0llllllll0110ll0 552 0301 14521411315254114134551 27 29 08 064 b 0302 ~03003102 250219994b062ll4 3 03042l1ll0101llCll101011l0Clll01l00011ll01Clll01Cllll1l01~001l11ll1lll1l0 492 0311 12514412514144121142541 28 24 J8 ObO 4 2 3 03144111100111lCl1lC101101001111lllCllOlllC1ll0ll1l10lllllJllllllOlll0llC 512 0321 555125455255~555il51521 3. Jl l9 081 4 2 3 03244lOOlJl001U1ClCC011000~1Clllll1:0000llCCCOOOllllOl001l01l010lCCOl0001 322 0331 54455:21454444444i~l~43 ~~ 29 22 085 0 0332 J4~~0COOC 490C:00010000v00 272 3 0334311!10ll1011110010111101110l0100011lC0101C010010101001111100011110010 402 0351 53552551345545141152411 29 28 20 077 7 0352 048000000 2904400030000000 0353 79092904 2915C430006Cb3 035421111001111C1111lll01011lll00110010~110ll~Ol!ll~Gl0lllv00llll0C011010 432 03ol 544444ZZ53lZ4~2;4443223 35 23 19 077 4 2 3 0364110101101111111COOll1lOll01111000111llOlll0101l100101l1llll0ll1llJ110 482 0371 44444444344244444344444 3d 2d 22 068 b 0372 03oOOOCOO 4i0270003C00jj000 3 03742111111011l01ll90010111111011110100111Cl00010111100ll100000ll01001l00 421 0381 15555515555544445145455 36 32 29 097 b 0382 04800UOOO 000~205040000~00 3 038431lllll0110910C010110lOOl10100vll111ClC1010l0000011110910109lll0011ll 402 0391 445125025241544455;;54~ 43 3' 11 ~ob 4 2 3 039421111000ll0lOlllll001011111l010l1l0llOCC1C01l01111001100011011lllOOOO 412 0401 5415~32014553i345351521 22 28 23 073 b 0402 ~270COOCO 2t00400030000000 3 04044110100101llOOCOOll1000010lCC0100010110010U011100111l11~1001luOll1000 322 0~11 5~4~5542,223454523~3255 42 24 21 087 8 0412 001007102 3301003036071201 0413 79061314 3073C530CC31C3 041429l401~1!10910990lll100lllvl0l001100110llluvl9110011100111llC011lOOll 431 0431 15515112;151541•115455~ 33 26 J7 Ooc 4 2 3 043411111l11l119lllll0101111!1Jll101111l100ll~C10011101ClCOOOl10101100lll 491 0441 54544444145155454255555 40 32 23 095 6 0442 02~00JOOO 370l4999300uOOOu 3 04442111111000111110ullv0lll10ll0001Jl00110001001111111110ll0011100100011 411 0451 25311522211154~23143231 Zb 24 J8 056 6 0452 038000000 300l800000COOOOO 3 04544111lll01l00110100llC11110110llOl1000lCOlC011ll00010100001ll111110010 401 0461 45544542424444424254422 37 31 16 084 c 0462 043000000 2802200030000000 3 04b421110111111111~COl10llll11110001110l1lll1100llllCC1lll11CllllOlll1100 511 0471 15211521214152251124551'25 24 10 059 4 2 3 C474210ll1ll111C0111101011Cll00110C01l0110111:0llll0ll0001JOOllllll1001ll 451 0481 22512432424244334144432 31 26 LZ Obq b 0482 045oooooo l~00999930000vuo 3 04842111lllllllC001Cl010110l11011lJ0111111Dll1011Q01111Cll01ll101.l1ll110 501 0501 55144543224244434334443 35 29 11 oa1 6 0502 J4900COOO 230lOOC~20~COv00 273

050~4011lllOl1l01111co1110lOll10010v1001lllOOoooll900lOllOOl0100100111019 381 os11 55555555455459555252555 43 35 29 1oe o 0512 wZlOZ1102 5200600006081119 3 051~31110010000101CClOlOC101111l01110000COC011000111001111011ll0100010001 331 0521 55425511435515515111551 32 31 13 076 4 2 3 05244111111010010l00010011lll1010001l10lOlOOlCOl0110011l110000111l1l00001 381 0531 325434343442444~~343444 35 l6 22 083 6 0532 014000000 4300400~30000000 3 053441100010100111000110110111000111000011CC1100001001ll1l00000110l011001 331 05~1 55324112515430231233243 32 23 15 070 4 2 3 a544301110111010111C01101010u10000111011001111000111011001011111111010000 391 0551 51454421111155515115553 37 22 12 071 6 0552 015000000 319q905020000000 3 055411111010101011C00110001lull111001010101C~l0110101l0ll10001010vl00011u 361 0561 55415541115155551155551 40 31 10 061 6 0562 025000000 3305000040000000 3 05642l10101011101110110011111111l1111001lOCCOOOOlC10110l1C101100110000ll0 401 0571 11422~34444244344242444 29 25 20 074 6 0572 007014102 4901700046012110 3 057421l100111101011010110100110110101111Cl01l10111Cl10lCCOl1111Clll111lll 471 osa1 15535514,15152551125453 3a z~ 11 o79 4 2 3 0584l01111100101Cl11001l110l110l01000111C10ll10l100100llll000101111000110 401 0591 55155555155511111533313 28 23 24 075 4 2 3 05944l11lullllCC11CClull1l11110~l1001001lllOl0111000010010011019ll1l0011l 441 ~601 544~4445545~45555554425 ~l 32 lo lOU 6 0602 0060!3102 3801108043081119 3 06044l0000l0000000000l~OOC1C010100110000100000l000000101l1000C01010000000 lo1 0611 44414422222~43324334~43 34 24 15 073 6 0612 001000000 3701100030000000 3 06144llll010l11111COl011C0011001000100l01lOllOOCOll0001CC001C11Clll01111l 372 0621 45524451545245513154532 40 32 12 084 6 0622 007000000 2100500030COOOCO 3 0624211l1ll01119l10l0011lC1111u11l1111111lOl110100l11110110011l110ll1l1lO 532 0631 45455535145254~11145551 37 29 id ~84 4 2 3 063420lll010010lC1lOCOl11ll0011lOOl110001001ll101Sll10101100010llll01001C ~92 0641 15241455315135125144:51 26 33 14 073 6 0642 114024212 ~:OC400036092104 0643 79102610 29630130325113 C6442111l~ll1l111llClOCOl1u1llllll1ll001ll~Oll010lll11111100JlOll01011110 ;02 274 0651 24222222222222222242422 20 20 l2 052 o 0652 012021102 4400300006081119 3 0654211111111119111Cl0101000110lll~llll111011101~111001011010119101110lll 0661 42143542144224554453~22 29 29 19 077 8 0662 004011122 3103000036103110 06o3 79100934 3C32031C433133 Ob6421l01l111ClC1001000011111101001001111000l0000001001111000110100100010 332 0671 44145445344245452244445 35 27 25 087 ~ 2 3 0674411101100110l000001100lll10101000100COC011000100011110000001100090010 272 0681 45544444344144441134444 37 25 21 083 b 0682 005~12122 5900500036112201 3 068411111011011100010110101vll110101100110CC1100ll1001011C0001010C1100010 362 0701 54451411224444525152421 27 30 15 072 4 2 3 0704211100110111110100111101111100100001110111000111001011000111111011011 0711 452224~5243444442342345 30 25 24 079 4 2 3 0714011010101C0~010000000llll0110lll01Cll00111000000101110910011100000001 312 0731 44444444254244444454442 38 29 21 Ob8 e 0732 013000000 2501100030000000 3 073431111111011Cll101010110010101110110110001C01101l10111100110111ll00110 442 0741 3524~434344233343344333 31 zo 21 078 4 2 3 07444110l1l10001COC10100001110010110001011ll1100100G011111000011100000111 342 0751 41254512415154142144212 27 23 15 085 b 0752 010011122 ~90C2030460921~4 3 0754211lll101l10111110010l0111011100110110011100000l00ll00010111001100110 402 0761 ~lZ;4l~l255255421154545 34 23 20 077 8 0762 012019122 6200300036124202 0763 79110503 29630130431613 07o411ll1ll0110111COOOl00001110101011001Cl01ll0l001100110000000r1011011lO 362 0771 555~5544544:55545455554 48 34 25 107 4 2 ~ 0774411110l0ll1Cl1C1001101011l010001010ll101110001l1001010000110111100010 382 0761 45541512522144;22122552 33 25 13 071 4 2 3 07B4llll1110111Cllll40C0100111111100110110010101llUlll1ll00011l1011100110 462 0791 55123~22455115444444531 31 33 15 u79 b 0792 ~17ulo~02 4300900031134201 3 07942l.OOvC100ll~lC001lCOCl0111lOll~~ll11000100001110llll0900l0900101001l 342 cao1 31154.11541554451111421 21 13 !1 oe1 4 2 3 08044011lll001010lCOOl1luOOC90Gl010JlllCC10101CCOOlOOOllCOl1~1101lil10Cll 332 0811 5~155433322311~15555555 32 l2 ~q 075 4 z 275 3 081~~110C01091110ll01000111000000110100110001C000110010111001l01100010011 311 0821 1551151151115~1~3115551 3u 2~ 09 Oo3 ~ 2 3 0824311110101111lll110100001111111l01111110l1l0100111l01l100ll1011110lll0 501 0831 ~~~~~~3443443434~2~3434 33 28 23 08~ 4 z 3 08342111011111111910l0ll1C011101llv01lll01011l01010ll0101000100000l11011l ~41 08~1 5~355535~~5555~5435351~ 30 33 lB 097 4 2 3 08442lOOOOOl010901000l10001l09lOOlOOOOOOOl1100001110011l11990000l0101oooo 2o1 0851 4512~5332~3422333323022 27 23 18 068 b 0852 0210vOOOO 5700100040000000 3 085~2101090l0010C0010100009100UOU11100101000001011100ll001000001100000000 221 0861 345145413122312222524~~ 32 23 11 ooo ~ 2 3 08o421i1l01111111ll011101olulOll1lllllOll1Clu10101110111100lOl1ll111010lO 501 0871 4~25451~411244444152~4~ 30 27 20 077 7 0872 008008122 330COOOC301~ll13 0873 Bv013lOo 30l4u33Z34141J 08744000l9199l000900901lOlOOOOol090l1l011101ll001109011110910199109191011 3o1 0881 5553~43252~252424145532 ~0 28 13 081 4 2 3 0884110110ll0ll00ll110l00lll1l011llll1011lC1110110111llC1100110llOOOOOll1 461 0891 452544442~41~4422~44454 37 21 z~ oa4 ~ 2 3 089~4111l01010lOlOOC100ll01lOl110101000lCOOOOOOOl010010111000111110100010 321 0901 ~44455111221414~5111141 30 l~ 13 062 6 0902 J2102812l o801C0003oOl2110 3 0904~lll111111111111ll011Clll00ll1ll100010ll01001010011Cl001J00110101Jll0 441 0911 3~5155115•1155115155551 3o 30 oo 012 7 0912 008036102 4901400031132201 0913 soo20103 3032055000o5o3 09l~1llll1ll1ll~l1lll00Glll1ll1l1110111110lll10111111010110001l01110l0110 521 0921 5251555151~1545222324~1 43 2~ 07 074 ~ 2 3 0924210100110l1C110C1ll~lll0111100100ll1019lll00191111111000lll01110101l0 451 C931 543455553545154442~4425 3~ 30 26 092 4 2 3 093~2110~110011010COCll0010101010000010001010~10110101lC01010100ll10ll00l 311 ~941 ;5355535~~~33!5532:5535 39 33 27 099 8 0942 00d041202 3t02b~OC361512Cl 0943 80010300 30130350003103 09444l1000010llCOOC000100Cl1CCOOOl010000~1~CO~OOlll009l119Cl00lCC0100l00l 231 0961 532355~323412531532342~ 32 28 15 07! 6 0962 005000000 25Clbv0v2JOOC000 276 l 09o~410ll111001ll~C00ll001llll0100110010llll01111000110101000010101111000 361 0971 11515531512151151515551 ~0 lb 10 Obb b 0972 024000000 24C2200C20000000 3 09742101001101101lCOla100110111111010000110C00010110110llC011000111110000 351 0981 44344431234544512352245 34 25 19 076 8 0982 005020102 Z9C~~OJ02o1o1Z01 0983 30010904 300~0240221202 09842lll000111lll10000001Ull111100101110011l1101~1100llll0000lll0llll00l0 411 1001 1243~312422433321123111 24 14 14 052 4 z 3 100421111011010919101000010llll10190ll0ll0ll110lll0llllll0001009lllll0lll 4b1 1011 4455252252515~424144~54 37 29 lo oa2 4 2 3 10144llll01100101100010llllllllu0111100111100011111000011011110C011101010 422 1021 34222555~~5154215425553 37 30 1b 083 ~ 2 3 102~111100101110llll000100111101010l9019101Cll001011100011000119ll0100110 1041 2414~455214141445142425 27 2b 2C 073 4 2 3 1044210111111110111001101011100111101111000111010001011111000111101000011 432 1051 11542411345154244113554 28 24 18 070 4 2 3 105421lll1101111111110001101111111111111100l11011019l1101000011911ll00ll9 512 l0b1 45155555541355354553535 39 30 27 096 4 2 3 l06~11100011100COOOOOOOC0000lllll0011000100001010110011010000C11101l00000 252 3001 55545545544155554455455 48 32 24 104 5 3002 00b012102 31008C003b071201 3003 79061921 30130350433133 4 3011 51515~11115551~41155511 37 22 13 072 5 3012 030055202 4700200036021113 3013 79072o11 29o30130332023 4 3021 43442452445234022145432 32 2b 18 074 5 3022 042000000 3200100~30000000 3023 79~73~31 30040230333033 4 3031 54343443453232553452345 37 24 24 085 5 3032 005021102 oooo90003e092104 3033 79062519 3GC4CZ40433233 4 3041 55545514554455451154445 39 29 27 095 5 3042 019027102 000l2J003o0Zlll3 3043 79051429 3CCOC230334043 4 3051 12431525554234313415444 32 23 20 075 2 2 3053 79051700 3C32035011b262 277 4 3061 45554554214234355443455 38 22 21 090 5 3062 016000000 4o01300C30oooooo 3063 790ou809 OOOOOC50034043 4 3071 44~~~534324255515153454 3d 32 17 087 5 3072 021000000 5600499~200~0000 3073 79051715 29800130317073 4 3081 5443444~434443455144434 37 28 23 088 2 2 3053 79051824 Z9000430ooooo3 4 3091 55535434554144554145434 41 29 22 092 3 3092 009021202 280ll0003o17120l 3093 79052921 3CGOC200~31212 4 3101 34244434343433342234433 30 23 23 07o 5 3!02 Vl90492CZ ~20Cv00o360l2110 4 3103 79092319 30040240212122 5 4 3111 45334333343235554455353 38 28 20 086 5 3112 015150302 6COC5000361d2201 3113 79052221 30040210323032 4 3121 4513!112545145351355535 30 27 20 077 2 2 3123 79052114 29!30120232553 4 3131 34323332422422222323432 29 19 14 062 5 3132 021000000 5!02000030000000 3133 79051506 30040240323233 4 3141 53215431125152514345421 3o 26 01 069 5 3142 OJ6013102 3800200046031115 3143 79051421 30000230443033 4 3151 34234432234344443234443 32 26 18 07e 5 3152 J13020202 450060~036012110 31~3 79051511 30040230332022 4 3161 41555541515554541155435 39 28 21 088 2 2 3163 79061421 30730530003032 4 3171 3324454443!155454154443 35 31 20 086 5 3172 018024202 630C500C4303lll' 3173 79052931 30040230302022 4 3181 33311151515531115115555 32 21 14 067 5 3182 001000000 650C,03040000uOO 3183 79061819 30040210222523 4 3l9l 35331434355355433354553 34 32 21 087 5 3192 ~15~68202 560~400036092104 3193 79070228 30u40240321212 4 278 3201 524~441232243~254225433 32 22 20 074 5 3202 001000000 2902300020000000 3203 79080331 30000240222102 4 3211 54355521554105554141551 36 31 18 08~ 5 3212 010129201 300C600036l71201 3213 79081~19 30040230~4211l 4 3221 44342412531234422153435 29 24 18 071 5 3222 039052202 ooooe04036092lu4 3223 79081011 31600540423033 4 3241 4441~~13512151211144441 35 zu va 063 5 3242 0340~1301 6101C00036191201 3243 79071911 30040230323133 4 3251 42424332424243434244442 37 24 13 074 2 2 3253 79073021 29630120325253 4 3261 25445555444234551152352 36 27 22 085 5 3262 011061402 CC0140003607l201 3263 79052222 29530132223133 4 3271 43324542114235441132433 29 25 14 068 5 3272 035086302 5701600024012110 3273 79ij72401 29530130332~23 4 3281 11511014544141114155454 29 22 15 066 5 3282 029000000 6700600030COOOOO 3283 79060921 29600130005202 4 3291 45155541524154224255554 38 32 15 085 5 3292 OU10,o202 3000406036092104 3293 79091329 30920430431013 4 3301 55112555555555555550215 38 32 26 096 2 2 3303 79091611 29530120118502 4 3311 14Q44411314544215124541 24 30 13 067 3 3312 001061302 3101COOC32031115 3313 79082911 30020200434113 4 3321 325111115343511~1154445 30 1d 17 06~ 3 3322 uOlvOOOOO 34004vOC300000uO 33l3 79072,10 30Cu020003101Z 4 3331 45445534225452225145552 36 31 1s oe~ ; 3332 006013102 570C500006202201 3333 79082021 3C730540332012 4 3341 24~3254522~444242155444 30 28 22 080 5 3342 v3800v000 54CU40003000COCO 3343 79081611 3C74C525312622 4 3351 4424444~2~4243443333343 34 24 21 C79 5 3352 ~~l0184~2 62004~1036191201 279 3353 79081012 30000240333233 4 3361 5445445343324445524345; 40 28 23 091 5 3362 0350COCCO 3tC030003000CUCO 3363 00000000 00000040000000 4 3371 11521411144454121125542 25 21 15 061 5 3372 001027202 25008050460~2104 3373 79061512 29530124112112 4 3381 55255515505245555153535 34 35 26 095 5 3382 021027102 2400800031031115 3383 79080321 30189350321022 ~ 3391 34253434242234231142244 25 21 22 068 2 2 3393 00000000 29500140322023 4 3401 24534531455344335445544 38 32 19 089 2 2 3403 79083111 30040230436063 4 3411 21454521125453242155433 29 25 19 073 5 3412 006021202 2102600036031115 3413 79100331 29570142441223 4 3421 42423243324232233333424 31 20 16 067 5 3422 018v26102 3400800036021!13 3423 79061114 2~530120222023 4 3431 44324212524442344245451 36 24 15 075 5 3432 034u41102 s9ooeooo36031115 3433 79101511 30000230441413 4 3441 14554554454544542255445 39 Z8 2a 095 3 3442 013033202 2b0~399936092104 3443 7910i911 29530120325221 4 3451 44314211411241242212442 30 15 11 056 5 3452 052063102 4400400036092104 3453 79101811 30730520421213 4 3461 45045415244154452545545 39 2d 24 091 5 3462 043057202 5200800042191201 3463 79101125 29530140332112 4 3471 224252324412453~;315534 36 21 19 c7c 5 3472 014000000 270D60CD2DOOOOOO 3473 79l0Cll5 2S5301202l25Z3 4 3481 44134545535324454455415 34 31 24 0~9 2 2 3483 79092131 29700l52005C53 4 2491 4534555525424324,354445 3b 31 25 092 2 2 280 3493 79092!31 3C~4C230C06113 4 3501 21534232311134432C54542 36 20 12 06e 5 35(12 UO~Ol1102 4201300036213~1~ 350"3 79100931 30730513431013 4 3511 15201404255154154555541 29 32 19 080 2 2 3513 79100931 30320350126063 4 3521 4424;44424424~44434444% 36 29 20 085 5 3522 044000000 3201610020000000 3523 79091733 29530120212523 4 353l 15145545544555544155455 37 32 27 1)96 2 2 3533 79091721 29530120111253 4 3541 44151542544255554443424 34 30 22 086 5 3542 001000000 5906200030000000 3543 79091431 30740530433031 4 3551 15224115514145151141432 23 2't 16 Ce3 2 2 3553 79082411 29530120317253 4 3561 425241~5221241451315434 35 u 15 co 1 z 2 3563 79092021 30000234325053 4 3571 34354544434424423243444 33 28 22 083 5 3572 029000000 ZCU6400030000000 3573 79092021 31600540433033 4 3591 11322241322152323125442 31 16 10 057 5 3592 J49053102 3200300036031115 3593 79101231 30730530321413 4 3601 43443532244234422233422 30 ,, 16 071 5 3602 037101202 4100300C31v31ll5 3603 79101236 29530120432023 4 3611 1454Z111214154151144452 27 22 14 063 5 3612 J4d00v000 61C0300C4COOOOOO 3613 79103031 30740530310000 4 JoZl 14~54155355214115112415 zz 24 z;; Oo9 ",J 3622 037051102 Z70l1U0036012110 3623 7906~321 3vlJ03523Z23l~ 4 3631 34343144312!21224121323 24 17 15 05o z 2 3633 79110931 29590124227072 4 3641 3404444434424~432443344 36 25 21 082 !:; 3642 JJlJ43302 e40130C03c~311l5 3643 79071332 29600140323221 281 4 3651 4534453224424535235~544 34 31 21 086 5 3652 035oooooo 5302v0~~3oooouoo 3653 79082412 30740530431013 4 3661 1~14151ll4l351131l1l33l 16 17 16 049 2 2 3663 79111339 30040240114243 4 3671 13311111411133331113331 23 13 08 044 2 2 3673 79071611 29530120225260 4 3681 44424434423244441244444 37 24 18 079 5 3682 021047102 6902600026021113 3683 79062226 29620130425333 4 3691 11214551411151511515521 38 15 06 059 2 2 3693 79072518 29530120215752 4 3701 43343435243344444234333 32 24 23 079 5 3702 001045202 3100300034012110 3703 79071211 30730540433033 4 3711 3554545152215455554543j 46 28 18 092 5 3712 001028lJl 3300300036l9.l01 3713 79062231 30000230431213 4 3721 44214442424144522445441 40 26 09 075 5 3722 014028102 4600210036092104 3723 79121919 30040230433033 4 3731 45545524544444342244444 38 28 24 090 5 3732 011013102 5100600046021113 3733 79112637 29530120432023 4 3741 55234532454225444344433 34 31 19 084 5 3742 017018202 4300900031132201 3743 79112634 30730530433033 4 3751 55553115511551115154415 31 22 22 075 5 3752 oo2ooa2o2 2o0030v02o031115 3753 79110938 30480340112123 4 3761 55344535454254412344435 38 28 22 088 2 3762 v04012102 330l20C0350ooovo 3763 79100228 3C0~~250321Cl~ 4 3801 44544312311421234123121 27 16 15 056 2 3802 009037202 6200300C33012ll0 3803 800l3C3Q 30740530321322 4 38ll 45544522124244442444442 36 28 16 ObO ~ 3812 007014122 37C5~00C3103lli5 3813 o0~2u431 oooooo32441313 4 282 3821 3~2~3~22~~~23~~422~342~ 27 26 20 073 5 3622 oo1o1~oo2 3voosooo36132201 3823 80020131 300402~3423132 4 3831 345445544~!14535415352~ 34 33 22 089 5 3832 001006202 3901200036231201 38B 60020132 3~730530531013 4 3841 3221151~425112215115~~4 2~ 25 13 062 2 2 3843 79051701 3017035011606Z 4 3851 44~245~23144322~4324442 35 25 15 075 5 3852 02402710l 4601100046021113 3853 79060422051600530431303 4 3861 4412234212511445~242422 23 28 14 06 5 5 3862 015022102 3600900033191201 3863 79052~16 29630130431213 4 3871 35445555555155555553535 ~3 35 25 103 5 3872 020025102 460~D002~031115 3873 7906u825 30040240201012 4 3881 ~41~25~523~425434152412 23 3l 21 075 2 3882 013023101 3401400030092104 3883 79070221 30730330433013 4 3891 54525453525154454445455 48 29 18 095 5 3892 007029202 3101000023031115 3893 79052916 OCOOC030332113 ~ 3901 ~4224425511142251124244 32 1b 18 066 3 3902 015020102 3600200035081119 3903 79052408 3180050044~0~2 ~ 3911 22~225~4~12125~25151522 27 29 12 068 2 2 3913 79051622 29530120212153 ~ 3921 5555555554~25~5252~5~55 47 31 23 101 5 3'122 001000000 3401300030000000 3923 79101908 295901~2211201 ~ 3931 55535525~51155551555235 ~5 24 2~ 09 3 5 3932 000000000 CCOOCQ99COCCOOOO 3933 7911Z6vZ 29530120221013 4 3941 5435553515~555~12555535 39 31 26 09!: 5 39~2 015000000 3301300C20000000 3943 79120711 3CC4C250211Cl3

3951" 3l41221141l5~2ll1133131 26 ll 10 04 7 5 39:)2 ooav~4l03 o700899931w31115 3953 79112728 3004~230431~11 4 3'161 3~150~~424445435~5~3~35 32 28 27 067 :. 283 3962 01~036102 5300~99~31012110 3963 79052100 29599112222012 ~ 397l 3531121151~515~5~15~551 ~ 29 30 H 073 "' 3972 0010~7200 !90C40l030012110 3973.. 00000000 307305lOC01~13 3981 5~5454255~225455~225~45 ~3 2ft 25 092 5 3982 0100~6202 3800900033182201 3983 79110631 30000230431213 4 3991 52515511555155515154541 ~0 32 10 082 5 3992 017032202 3300300036151201 399~ 79112631 30130350433133 4 4001 ~542~422424244~44444~42 38 29 1~ 081 2 2 ~003 80012401 29599140221313 4 ltOll 15515511511111315515533 34 23 08 085 5 4012 01~028102 2302~00046213112 4013 79122036 29530120332123 4 40Zl 2542211251~522542125554 32 21 18 071 5 4022 001036202 3100400036231201 4023 79121039 30040210431013 4 4031 45545552525351544455555 48 30 20 098 5 4032 033044200 3100500000191201 4033 79080233 30100320223213 4 4041 32413542415135455354551 37 32 11 080 1 2 3 4 4051 55255555255155545155555 40 35 25 100 5 4052 022029122 5401200016213112 4053 80013031 30000330332023 4 4061 45354~5335532~355151415 27 32 26 08 5 5 lt062 0070L4122 390lo01036031115 4063 80012939 30189340441412 4 4071 54145545254435454154535 32 32 2 8 09 2 5 4072 001011112 35011v1u31112201 4073 80010331 30730520321013 4 4081 2~2~3~4325533~2~2~~5323 29 27 23 oJH 5 4082 001028212 330C5~0C36151201 ~083 80011131 30730530~41013 ~ 4091 1234515111~1441~~111554 27 22 lt; Oe5 2 2 ~093 aJ~llo31 31o005lOOv3'01 4 4101 555~5~4323434443~235541 4v 29 11 ost: 5 4102 010000320 2701000G30COOOOO 284 ~103 79121331 3t00031032~00~ ~ ~111 44544131221434~15142443 33 23 15 071 2 2 4113 80013031 30280314221013 4 4121 54544414145344251445545 35 27 25 087 3 4122 008015212 230v4COC36021113 4123 80012931 3011v300Z22023 4 4131 54514411522542211125551 37 20 11 068 2 2 4133 00000031 30730~20231012 4 ~14l 215124215112~442~455455 38 23 12 Q73 5 4142 001100212 3101600031081119 4143 79070611 30000210321013 4 4151 44422422224244444244444 32 26 16 07t 5 4152 032000030 5401o00030000uoo 4153 79071719 30040230433232 4 4161 55155555555155515515555 46 31 22 099 2 2 4163 79102630 2953C1201lo062 4 4171 54235555555135534455555 44 33 22 099 2 2 4173 80020431 29560120432503 4 5001 55555544555555555355555 47 35 29 111 l 2 3 4 5011 55555555555l55555355515 44 35 26 105 1 2 3 4 5021 44432542524244424243441 36 29 12 077 1 2 3 4 5031 55235444255355525552451 40 33 18 091 l 2 3 4 5041 55255555555555555~55555 47 35 30 112 l 2 3 4 5051 54525545545553541553555 47 28 25 100 l 2 3 4 5061 44454555125245552254555 38 31 24 093 l 2 3 4 285 5071 53425354524054501105050 45 21 lb 082 1 2 3 4 5081 1151111141212C111121441 21 12 Ob 039 1 2 3 4 5091 44344404244~44444043444 35 28 22 085 l 2 3 4 5101 54555545554444452244425 40 27 29 09~ 1 2 3 4 6001 24422422411151121124343 29 16 11 056 1 2 3 4 6011 55225535444145543353455 39 31 21 091 1 2 3 4 6021 44145442425154555044442 40 30 16 086 1 2 3 4 6031 44245255444254544144424 37 26 23 086 1 2 3 4 6041 44452422444444251124223 29 21 23 073 1 2 3 4 6051 22552321234425254144224 24 24 22 07C 1 2 3 4 o061 15115511111154511115551 30 22 Oo 058 l 2 3 4 6071 42522542424445444154442 36 29 16 081 1 2 3 4 6081 54244444244244434444443 37 28 20 085 1 2 3 4 6091 445i4431514;4453214~54l 39 27 12 078 1 2 3 4 286 6101 43454344344544334144434 34 26 25 085 1 2 3 4 blll 542545411552454214~4525 34 30 zu 084 1 2 3 4 1 2 7003 80011732 3CC4C22C002lu3 4 1 2 7013 79100134 30040240431003 4 1 2 7023 79122721 30C4C230331303 4 l 2 7033 7~102531 3C73u54043lCv3 4 1 2 7043 OOOCOuOO OOOOCC203220u2 4 1 z 7053 OOuOuuOL 29530120010000 4 1 2 7063 a001C831 CC000050210COO 4

2 7073 80012221 3C04C24J32350C 4 1 2 7083 79112600 30120340432003 4 •' 2 7093 7~082401 2953C1201l3003 4 1 2 7103 79lv1c3C 29:90140114201 4 l 2 7113 79101228 3C~CC240ll5lC3 4 l 287 2 7123 79122138 30040240433i03 4 1 2 7133 79092431 30040234321203 4 1 2 7143 80012031 30C402433210v3 4 1 2 7153 80011832 29630150445003 4 l 2 7163 80012831 00000045211201 4 1 2 7173 79080809 30730~40~33203 4 l 2 7183 80013130 29530120321303 4 l 2 7193 79082215 30310340222003 4 1 2 7203 79110539 30730530331203 4 1 2 7213 79122731 30730510302103 4 1 2 7223 79092o18 29530120232003 4 l 2 7233 90011031 O~COOC40221203 4 1 2 7243 79J9l001 2954C1~022bl03 4 1 2 7253 79092831 30730~30431003 4 1 2 288 72b3 80010431 3C55C!l303213v3 't 1 2 7273 00000010 29bOC130322303 It 1 2 726 3 79092131 3004023043 B(.3 4 1 2 7293 800.1.2301 2954\..l.'t2123003 4 1 2 7303 7909173() 30990410222003 It 1 2 7313 7910303~ 3022C.340ZZ2lu3 It l 2 7323 80012202 30120340231203 It 1 2 7333 79ll1b04 3004v23032bCv3 It 1 2 731t3 791013 3't 3C000230321~02 4 1 2 7353 79121931 3(.(.40225331203 4 l 2 73b3 79091232 30000240322003 4 1 2 7373 7'H..L2b3.i 3C.Ct4C.250433l03 4 1 2 738 3 7909 2030 2957014C322

The dissertation submitted by JohnS. Crandell, M.S., has been read and approved by the following committee:

Dr. Robert C. Nicolay, Director Professor, Psychology, Loyola

Dr. Alan S. DeWolfe Professor, Psychology, Loyola

Dr. John R. Shack Associate Professor, Psychology, Loyola

The final copies have been examined by the director of the dissertation and the signature which appears below verifies the fact that any neces sary changes have been incorporated and that the dissertation is now given final approval by the Committee with reference to content and form.

The dissertation is therefore accepted in partial fulfillment of the requirements for the degree of doctor of philosophy.