MSU ‘ LIBRARIES n Vu— RETURNING MATERIALS: Place in book drop to remove this checkout from your record. FINES will be charged if book is returned after the date stamped below. $193355 EGO DEVELOPMENT AND COUNSELING: THE EFFECTS OF COUNSELORS' AND CLIENTS' EGO DEVELOPMENT LEVELS UPON THE EXPRESSED EMPATHY AND PREFERENCES OF COUNSELORS By Thomas Joseph McIntyre A DISSE RTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Counseling and Educational Psychology 1985 To Eddi, with my love ACKNOWLEDGEMENTS My thanks to John Powell for his faith, encouragement and goodwill. To Dick Johnson, Chris Clark and Ralph Kron for their understanding, suggestions, and mercy. To Joe Schroeder, Dirk Lorenzen, Joe and Marilyn Parke-Jeney, Rodger Rinnard, Jack Loynes and Chip West for their assorted invaluable aid in the initial planning, data collection, ratings, data analysis and housing. To the United Dissertation and Comprehensive Workers, Solidarity! To the Bluesmobile, for rising from the ashes to .make one last roadtrip again and again. To Brando, Ruby Falls Begonia, Ozzie and Harriet, and Tortillas for their diversion and sustenance. And one final acknowledgement and thanks to my family. for their encouragement. ABSTRACT EGO DEVELOPMENT AND COUNSELING: THE EFFECTS OF COUNSELORS' AND CLIENTS' EGO DEVELOPMENT LEVELS UPON THE EXPRESSED EMPATHY AND PREFERENCES OF COUNSELORS By Thomas Joseph McIntyre The study investigated the relationship between counselors' and clients' ego development levels and counselors' expressed empathy and expressed client preferences using a counseling analogue method. The research was based (on the theories of ego deve10pment (Loevinger, 1976) and empathy (Rogers, 1975). The hypotheses were a) counselorsl ego development is positively related to their expressed empathy, b) counselors' expressed empathy varies among deveIOpmental levels of analogues, c) an interaction exists between counselor and analogue developmental levels so that expressed empathy peaks when counselor and analogue develOpmental levels match, and d) counselors prefer analogues at one level higher and at the same level of ego development as the counselors. Hypotheses b, c and d were confirmed. Subjects, 42 master's level counseling students, completed the Washington University Sentence Completion Test of ego deve10pment, responded in writing to four typewritten analogues (representing clients at the Conformist, Self-Aware, Conscientious and Individualistic ego development levels), and rank ordered their preferences for the analogues. Subjects were assessed at the Self-Aware, Conscientious and Individualistic levels. Subjects' counseling responses were rated for expressed empathy (the dependent variable) with the ReSponse Empathy Scale. A 3 x 4 Analysis of Variance for repeated measures demonstrated a non-significant main effect for subject level, a significant main effect for analogue level (2 < .05, non-directional), and a significant interaction (3 < .05) between subject and analogue levels. A Chi sguare indicated the subjects preferred analogues whose developmental level was the same or one higher than the subjects'. The data showed the subjects' expressed empathy peaked in response to analogues at the same developmental level. Post hoc tests indicated a) lower (2 < .05) expressed empathy in response to the Conformist analogue than the Conscientious analogue, b) the Self-Aware subjects had lower expressed empathy (p < .05) with the Conformist and Individualistic analogues than with the Self-Aware and Conscientious analogues, c) the Individualistic subjects had higher expressed empathy (p < .05) with the Individualistic analogue than did the Self-Aware subjects, and d) the overall trend in the preference data was subjects across levels indicated lowered preference for the Conformist analogue. Implications and recommendations for counselor training, supervision, practice and research were discussed. TABLE OF CONTENTS List of Tables List of Figures Chapter 1: The Problem Purpose Theory Hypotheses Overview Chapter 2: Review of Literature Loevinger's Theory of Ego Development Stages of Ego Development Measurement of Ego Development Summary of Loevinger's Theory Empathy The Empathy Concept The Measurement of Empathy Response Empathy Scale The Use of Counseling Analogues in Research on Counseling Ego Deve10pment and Empathy The Theoretical Relationship Empirical Evidence of the Relationship of Ego Development to Empathy Summary of Ego Development and Empathy Summary of the Review of Literature iv vi 11 ll 11 16 19 25 30 30 33 37 38 Chapter 3: Design of the Study Sample Procedure Measurement of Ego Development Client Analogues Measurement of EXpressed Empathy Design Testable Hypotheses Statistical Analyses Summary of Method Chapter 4: Results Expressed Empathy Subjects' Preferences for Client Analogues Expressed Empathy by Sex of Subjects Chapter 5: Discussion The Parameters of Generalizability Review of Results and Discussion The Main Effect for Subject Level The Main Effect for Analogue Level The Interaction Between Subject and Analogue Levels of Ego Development Trends in Counselor Preferences Recommendations Appendix A: Washington University Sentence Completion Test Appendix B: Client Analogues Appendix C: Revised ReSponse Empathy Scale List of R eferences 41 41 41 42 43 44 45 46 47 48 49 49 S4 55 56 56 58 58 60 61 62 64 68 71 76 79 LIST OF FIGURES Figure 4.1: Mean RES Total Empathy Scores by Subject Level Figure 4.2: Mean RES Score Across Subjects by Analogue Level Figure 4.3: Mean RES Total Empathy Scores by Subject Levels vii SO 51 S3 Chapter 1: The Problem Since Loevinger's elucidation of her concept of ego deve10pment (1966) and the publication of an assessment technique for this construct (Loevinger, Wessler & Redmore, 1970) there have been a growing number of investigations involving ego deve10pment. These investigations have had various emphases. Some studies have sought evidence for the validity of ego development theory, some studies have examined correlations between measures of ego deve10pment and other constructs, and other studies have sought to induce deve10pmenta1 change in subjects by way of experimental interventions (Loevinger, 1979). Some studies have also related level of ego development to a counselor's level of empathy. Historically, a counselor's empathy has been considered a key ingredient leading to successful outcome of counseling for a client. Recently empathy theory has been broadened to consider empathy as a multifaceted process occurring within a relationship rather than a constant ability which a counselor exercised. Swensen (1980) stated that ego deve10pment theory may form a basis for a unified approach to the understanding of counseling. A study relating ego development of counselors to their general empathic ability is part of this basis (C arlozzi, Gaa & Liberman, 1983) as is an effort to use ego development theory as a way of understanding client dynamics (Young-Eisendrath, 1982). What has not been considered is how a counselor's level of expressed empathy may be a result of and vary according to both the counselor's and client's levels of ego development. This study investigates this question and therefore is an opportunity to understand the developmental level of counselors, client dynamics and counseling process simultaneously. Furthermore, this study seeks to establish whether the relationship between counselor and client level of ego development 1 2 may result in differences in a counselor's preference for a client. This asymmetry in preference was established in the case of moral development theory (Rest, Turiel & Kohlberg, 1969). The uncovering of an asymmetry in preference due to the relationship of counselor to client level of ego development would be an important contribution to the understanding of counseling and the client selection process. Both Loevinger's theory of ego development (1966, 1976) and Rogers' ideas about empathy (1957, 1975) deal with a person's internal frame of reference. A clarification of the relationship between empathy and ego'development may be of interest to those involved in counseling and the training and supervision of counselors since empathy is considered important in successful counseling. In addition, ego developmental theory is thought to present a potential addition to the understanding of the counseling process, but there is little empirical basis for this hope as of yet. Purpose The present study examines three interrelated tOpics. First, it is an attempt to confirm the relationship between ego developmental level and expressed empathy. Secondly, it focuses upon the interaction between ego deve10pmenta1 level of counselor and client and the counselor's expressed empathy through which he or she demonstrates understanding of the client's internal frame of reference. With this aspect the present study makes a unique contribution. Finally, an attempt is made to understand how a counselor's preference for a client relates to counselor and client level of ego development. Theory The frame of reference a person uses to bring coherence and meaning to his or her experience of self and others is the focus of the theory of ego development which Loevinger (1966, 1976) has constructed. It is this theory 3 which Swensen (1980) claims has the potential of providing a framework for the understanding of counseling. Yet there appears to be little research in existence which would bring substance to Swensen's claim (Loevinger, 1979; Carlozzi et a1., 1983). Carlozzi et a1. (1983) hypothesizes that there is a relationship between ego deve10pment and empathy. The construct of empathy (Rogers, 1957) is a thread which has run through much of counseling practice and research for decades. According to Rogers' definition, "Being empathic is to perceive the internal frame of reference of another with accuracy and with the emotional components and meanings" (1959, p. 210). It may be inferred that the individual's internal frame of reference is a point at which the relationship between ego deve10pment and counseling may be examined. Research has just begun to investigate the relationship between a counselor's ego deve10pment, which determines that counselor's frame of reference, and measures of the counselor's empathic ability which contributes to the counselor's understanding of a client's frame of reference (Carlozzi, et a1., 1983; Zielinski, 1973, cited in Loevinger, 1979). Of course, it was the client's internal cognitive and affective frame of reference which Rogers (1957, 1959, 1975) focused upon. What the literature has not addressed yet, except for one inconclusive example (Young-Eisendrath, 1981, cited in Young-Eisendrath, 1982), is the question of how frames ofreference of both counselor and client, as determined by ego development, effect the counselor's eXpressed empathy. Through this expressed empathy the counselor demonstrates an understanding of the client's frame of reference, but this understanding may be shaped according to the counselor's own frame of reference. The present study, then, may contribute to the literature by clarifying the relationship between counselor and client levels of ego development and the 4 counselor's expressed empathy. An understanding of this relationship might have implications for the content of counselor training, provide a potential focus for counselor supervision, and form one basis for optimal counselor and client matching. Hypotheses 1. A positive relationship exists between counselors' level of ego deve10pment and level of expressed empathy. 2. Counselors' expressed empathy will vary in relation to clients' ego development level. 3. Counselors will demonstrate a higher degree of eXpressed empathy in response to a client at the same level of ego development compared to clients at levels higher or lower than the counselors'. 4. Counselors will assign a higher preference to clients of equal ego developmental level than those at levels lower or more than one level higher than the counselors'. Overview Two important tapics remain before the reader can put the results of this study in perspective. In Chapter 2, the literature which is relevant to this study and its implementation is examined. This includes a look at the theories of ego development and empathy, the few studies which have so far examined their relationship, studies dealing with the possibly mediating characteristic of preference, and parallel studies dealing with the theories of conceptual level and moral deve10pment. In addition, areas of the literature which deal specifically with methodological aspects relevant here, such as the use of counseling analogues and of empathy measures, are reviewed. Next, in Chapter 3 the methodology is set forth. Characteristics of the sample, the instruments, methods of ratings and reliabilities obtained, 5 hypotheses, design and statistical analyses are described. It is hoped that the reader will arrive at the end of Chapters 2 and 3 with a thorough understanding of the foundation of previous research and theory and the subsequent method upon which the results of this investigation stand. In Chapter 4 the results of the statistical tests of this study's four hypotheses are presented. The results of post hoc tests which contribute to an overall understanding of this study's findings are also presented. The findings and their implications are discussed in Chapter 5 along with recommendations for practice, research and training. Chapter 2: Review of Literature Several areas of theory and research are relevant to the present investigation. Therefore, in the following review, the reader becomes acquainted first with the theory of ego development, its measurement, reliability and validity, and then with the theory of empathy, and its importance, variability and measurement. Recent re-conceptualizations of empathy and a focus on the uses of counseling analogues in empathy research are included here also. Finally these two theories are related to each other and the basis in previous literature for the present hypotheses discussed. Loevinger's Theory of Ego Development Jane Loevinger (1966, 1976) has developed a synthetic theory of ego development which, while having links to a broad spectrum of earlier theories of personality, can be seen as a unique influence in current psychological understanding. Swensen (1980) noted this theory as having the potential to provide a general framework for the understanding of counseling. Loevinger conceived of the ego as a unified framework which the individual uses to perceive and interpret the world. This framework included impulse control, character development, interpersonal style, conscious preoccupations and cognitive style (Loevinger, 1976). In addition, Loevinger postulated that this ego deve10ps through qualitatively different steps. Ego development is seen as a process parallel to, but separate from psychosexual, intellectual and physical development. Loevinger's...conception of ego development assumes that each person has a customary orientation to himself and to the world and that there is a continuum (ego deve10pment) along which these frames of reference can be arrayed. "In general, ego development is marked by a more differentiated perception of one's self, of the social world, and of the relations of one's feelings and thoughts to those of others" (Candee, 1974, p.621). It is the sequence of steps along this continuum of differentiation and complexity that Loevinger and her co-workers have delineated as stages of ego development. (Hauser, 1976, p.928-929) 7 These ideas have not remained solely in the realm of theory. Crucial to Loevinger's work has been the arrival at a method of assessment, which led to theory testing and modification. Loevinger, et a1. (1970) developed a projective sentence completion test which assesses an individual's ego development as a unified construct. This measure, the Washington University Sentence Completion Test (W USC T), repeatedly has been shown as reliable, and an accompanying self-training program produces raters equally reliable to those tutored or previously experienced ones (Loevinger, 1979; Hauser, 1976; Holt, 1980). The WUSCT has been shown to have sufficient validity for research purposes (Loevinger, 1979). _S_t_a_gg of Ego Develment Loevinger delineated .10 stages and transitional levels of ego development. Nine of these are measurable with the WUSC T, the first is not. The transitional levels are transitional only in the theoretical sense, as individuals may remain stable at these levels over time. The stages and associated symbols are: Presocial, I-l; Impulsive, I-2; Self-protective, Delta; Ritual-Traditional, Delta/3; Conformist, I-3; Self-aware, also known as the Conscientious/Conformist Transition, I-3/4; Conscientious, 1-4; Individualistic, I-4/ 5; Autonomous, I-5; and Integrated, I-6. Qmmon stfies for adolescents and adults. Holt (1980), in a national survey using a 12-item short form of the WUSCT, found the modal level for young adults to 'be the conscientious/conformist level, I-3/4. Other studies have agreed and have indicated I-3/4 to be the modal level for adults (Hauser, 1976). Additionally, Holt's data indicated that 832 of college men and 892 of college women were assessed at stage 1-3, I-3/4, I-4 or I-4/5. The corresponding stage names are the conformist, conscientious/conformist, conscientious and the individualistic sta ges. 8 A closer look at the common adolescent and adult stages of Loevinger's theory is useful as the interaction of these four stages of development were examined within this study. - ' Conformist stage. The conformist stage, I-3, is one which most people reach during childhood or adolescence. A person functioning at this stage is very concerned with following rules and with the disapproval and shame involved with breaking rules. Relatively little differentiation of internal states is apparent and the affect reported is generally banal. The focus of interactions with others is on actions and events instead of feelings and motives (Loevinger, 1976; Hauser, 1976). Self-aware level: The conscientious/conformist transition. Beginning at the self-aware level, I-3/4, an individual gains the capacity for introspection, psychological causation, self-awareness and self-criticism. It becomes apparent to the individual that the family or social groups can no longer provide workable, clearcut rules which corresponding to reality. Therefore the rightness of an action is judged according to relative aSpects of the situation, such as time and place. These contingencies are simplistic, however, and are to become more differentiated at higher levels (Loevinger, 1976; Hauser, 1976). Conscientious stage. Idealism and "an excessive and sometimes intrusive sense of responsibility for the welfare of other people", (Young-Eisendrath, 1982, p.332) are two notable characteristics of those persons at the conscientious stage, I-4, of ego development. Obligations, self-criticism and achievements measured by inner states and rules occupy consciousness. Feelings and motives rather than actions alone are the currency of this individual's view of social interaction. Non—stereotyped individual differences are perceived and "social interaction is experienced as more vivid and meaningful than in earlier stages" (Hauser, 1976, p.931). 9 Individualistic level. Greater conceptual complexity, a concern for emotional independence and an increased sense of individuality characterize those at the individualistic transition level, I-4/5. At this point persons are more aware of inner conflict as well as conflicts between one's own needs and those of others. An important change from the conscientious stage is that persons are more accepting of themselves and others and are more tolerant of differing viewpoints and Opinions of others. It is still difficult, however, for an individual to tolerate internal conflict which has increasingly entered awareness, or to tolerate the notion that conflict and paradox are a part of the human condition (Swensen, 1980; Loevinger, 1976). Measurement of Egg Development Washiflton University Sentence Completion Test. Loevinger's concept of development is assessed via the W080 T, a sentence completion test of 36 items, with separate forms for men and women (see Appendix A). The WUSCT "has been carefully constructed and standardized in terms of its form, administration, and scoring procedures" (Hauser, 1976). Since the publication of a scoring manual (Loevinger, et al.,1970) it has been extensively used for research purposes. The scoring manual consists of individual manuals for each sentence stem, a self-training program and the ogive rules by which the cumulative frequencies of the individually scored items are examined and converted into a single total protocol rating (TP R). The extensive self-training program produces raters as reliable as tutored or experienced ones (Loevinger, 1979; Hauser, 1976; Holt, 1980). Reliability of the scoring procedure. When the original WUSCT scoring manual was validated Loevinger reported a median product-moment correlation coefficient between experienced raters of E = .75 (Loevinger, 1979). The median correlation for TPR's was 5 = .86. Subsequent ratings of the same sample of 543 10 protocols of a diverse group of women and girls by four self-trained raters produced results indistinguishable from that of the experienced raters. Various pairs of experienced and self-trained raters produced a median interrater correlation on item ratings of 5 = .76 (Hauser, 1976). Holt (1980) reported reliabilities obtained in the scoring of a national sample of 966 persons between the ages of 16 and 25 years. The 12-item WUSCTs were scored using a team of graduate and advanced undergraduate students at New York University. Reliabilities computed with intraclass correlations were reported. Interrater reliability on individual items ranged from .72 to .90, median .825 (women), and .57 to .90, median .78 (men). This study provided important additional evidence for interrater reliability as scoring done by a group separate from Loevinger's and on a nationally representative sample verified consistency of results. Evidence for validity. The internal validity of the WUSCT rests on a number of points. This measure has face or content validity as its projective nature permits the subject to supply his or her own frame of reference. It is this frame of reference which the test tries to measure. The conception of ego deve10pment can be communicated coherently. The statistical homogeneity of the WUSCT has been shown and research aimed at breaking the items into subscales has failed. Thus, there is strong evidence that it measures a single construct (Loevinger, 1979). "Evidence for construct validity is substantial, but falls short of clear proof of sequentiality" (Loevinger, 1979). In any case this theory can be viewed as a series of character types. Evidence for sequentiality is supportive of construct validity, as Loevinger summarized: The profile or distribution of item ratings for protocols at adjacent stages are more alike than at non-adjacent stages. Mean ego level increases with age for cross-sectional samples during adolescence. Mean ego level increases with age for longitudional studies during ll adolescence. Test-retest correlations are significantly positive for longitudional studies, even over spans of 6 years. Attempts to raise ego level eXperimentally in a few weeks have not succeeded, but eXperiments that have lasted 6 to 9 months have had statistically significant success. Finally, peOple can lower their score more reliably and decisively than they can raise it. (1979, p. 305-306) In addition, Loevinger (1979) summarized evidence for external validity in the form of significant correlations of low to moderate magnitude between the WUSCT and a variety of other measures. These included interviews, objective tests, projective tests and behavioral measures. Summary of Loevinger's Theory Loevinger's theory of ego development (1966, 1976) has focused upon the evolving internal frame of reference which an individual uses to perceive and interpret the world. The location of individuals upon this common path of development is assessed using a projective sentence completion test._ Research has shown this test, the WUSCT, to be sufficiently reliable and valid for research purposes. my Several areas of theory and research focusing upon empathy are relevant to the present investigation. These areas are reviewed below. This review includes a look at the definition and importance of empathy in counseling, the commonly used measures of empathy and criticisms of these measures, the characteristics and validity of the Response Empathy Scale, and the strengths and limitations of counseling analogues employed in empathy research. The Empathy ConceLt Rogers' view. In Rogers' formal statement defining his concept of empathy were several key features. He referred to a "state of empathy" (1959, p.210) as if it were something constant. This state had to do with perceiving another's "internal frame of reference...with accuracy and with the emotional components" (1959, p.210), and the empathic person was to maintain an 'as if' 12 quality. It was this concept, along with Rogers' views on congruence and warmth (1957), which generated much research in following years (Gladstein, 1983). Rogers' (1975) re-examination of empathy underlined his Opinion that the continued use of empathy as a concept is valuable in training and research, and as a key part of human communication. Rogers also introduced a refinement to his views. No longer did Rogers (1975) refer to a state of empathy. Rather, Rogers presented empathy as a process with several facets. First, one must sense another's "private perceptual world" (Rogers, 1975, p.4). Next, one must communicate these sensings. Finally, this communication is received and one can be guided by the other's response. Reconceptualizations 1y Barrett-Lennard and Gladstein. Two subsequent viewpoints, Barrett-Lennard (1981) and Gladstein (1977, 1983; Feldstein 8 Gladstein, 1980) added depth and clarity to Rogers' (1975) definition. Barrett-Lennard examined the process of empathy Rogers refers to. Three distinct and temporally sequential phases are specified in what he called the empathy cycle. These are empathic resonation (Phase 1), expressed empathy (Phase 2), and received empathy (Phase 3). Barrett—Lennard (1981) went on to locate previous empathy measures in relation to the cycle. He predicted that only moderate correlations can be expected between measures of adjacent phases of the empathy cycle. One use of this cyclic conception may be to specify what part of the cycle is being investigated in a particular research effort. Gladstein (1983) reviewed the various theoretical models used in the definition and measurement of empathy. He noted two general elements in the many empathy definitions employed in the counseling, developmental and social l3 psychology literatures: cognitive and affective empathy. Cognitive empathy refers to intellectually taking the role or perSpective of another person. That is, seeing the world as the other person does. Affective empathy refers to responding with the same emotion to another person's emotion. That is, feeling the same way as another person does...Rogers' definition (1975, p.4) included both. (Gladstein, 1983, p.468) Gladstein's (1983) review was stimulated in part by the diversity of outcomes in the literature on empathy. His review concluded that empathy may or may not lead to helping behavior. Therefore, he said it is essential to determine when each element, cognitive and affective, is or is not important. He further noted that based on theoretical consensus it appeared that affective and cognitive empathy would be important in initiating and building a counseling relationship. Gladstein also mentioned cognitive role-taking empathy as important in the problem-identification and exploration periods of counseling. Gladstein (1983) stated that empathy is a multi-stage process agreeing with Rogers' (1975) and Barrett-Lennard's (1981) comments. He placed the affective component prior to the cognitive component temporally. It is apparent that both affective and cognitive empathy would occur within the counselor prior to Phase 2 of Barrett-Lennard's cycle. Resulting communication, expressed empathy, takes place in Phase 2. It is also important to note that Barrett-Lennard's definition of cognitive empathy included an understanding of the "client's feelings and thoughts fiom their frame of reference" (Feldstein & Gladstein, 19 80, p. 54). The importance of empathy. The above viewpoints concerned the nature of empathy, but did not address its importance in counseling. Rogers (1957) called empathy one of the "necessary and sufficient conditions for therapeutic personality change." Rogers (1975) concluded from his review of the literature that "empathy is clearly related to positive outcome" (p.6). He based his conclusion of positive outcome in part upon studies which have found a positive 14 relationship to exist between empathy measured from a counselor's or therapist's actual interview behavior and outcome measures (Kurtz & Grummon, 1972; Mullen & Abeles, 1971). He reasoned that this positive outcome is due to empathy's effect in dissolving alienation, and in increasing the client's sense of self-value and identity. Gladstein (1983) noted that challenges had been raised regarding the above conclusions and that research evidence for empathy's importance was mixed. He pointed out that mixed findings might have been due to the differing and sometimes obscure ways in which theorists and researchers had defined and measured empathy. He noted some conceptual refinements which could prove helpful to clear up the mixed results. Gladstein concluded, however, that empathy "can be helpful in certain stages, with certain clients, and for certain goals" (1983, p.467). He did note cognitive and affective empathy as important in the early stages of counseling, particularly when the counseling goals were self-exploration. He was less supportive of the importance of empathy, especially of affective empathy, when the goals involved Specific problem-solving or action-oriented decision making. Empathy's variability. Rogers (1959) originally spoke of a state of empathy. His more recent view (1975) and those of others (Barrett-Lennard, 1981; Gladstein, 1983) explained empathy occurring as a multi-phasic process. The magnitude of empathy may vary, particularly as expressed and received empathy. Several reports have confirmed that a counselor's level of eXpressed empathy may vary. Truax (1966) found that a counselor's empathic behavior varied in a manner contingent upon the patient's behavior. Curman (1973) measured the level of therapeutic conditions of empathy, warmth and genuineness of three high- and three low-functioning therapists. The levels of both groups were found to vary 15 across and within sessions. Changes across time within sessions were noted such that the therapists' empathic behavior and other conditions peaked in the middle and'end of the session. As noted by Rogers (1975), the hvel of a counselor's empathic behavior may vary positively according to the counselor's psychological adjustment. Bergin and Soloman (1970) compared the scores of 18 post-internship psychology graduate students on the MMPI and the Truax accurate empathy scale. A negative correlation was noted between the MMPI Depression and Psychasthenia scales and the scores for empathic behavior. Vesprani (1969) noted the same relationship with 33 untrained college student subjects. Summary of the empathy concept. Some specific points deserve emphasis. Empathy is a process not a state. Barrett-Lennard's Phase 1 empathic resonation is composed of both affective and cognitive elements. Empathy's behavioral component, Phase 2, expressed empathy, may vary. Empathy is important in counseling in that at the least it is sometimes helpful. The reconceptualizations reviewed above demand consideration in planning research and received it in the present study. Some questions need to be answered. Which element of Phase 1 empathy is involved, cognitive and/or affective; at what point is the cycle being examined, empathic resonation, expressed empathy or received empathy? Is the measurement one which can focus upon the particular element and phase involved; and, finally, is empathy being examined at a point in counseling at which it might be expected to matter? The answers to the above questions can help focus a research project, confirm the choice of measurement, and make sense of diverse research findings in the empathy literature. The present study, for instance, examined the result of cognitive role-taking empathic resonation evident in expressed empathy. This 16 study used an analogue representative of a time in counseling (initial counseling contact) when empathy is important, and used a measure which focused upon verbal Phase 2 expressed empathy. The Measurement of Empathy It has been noted above that empathy may vary. This is an important point, as the present study investigated the variability of counselors' expressed empathy relative to a client and counselor characteristic, ego development. In order to accurately measure this variability and thus provide a test of this study's hypotheses, it was essential that the subjects' expressed empathy be measured in a reliable and valid fashion. The review below, of the various measures of empathy available, resulted in the selection-of the needed measure. Commonly used measures. Following Rogers' (1957, 1959) conceptual advances, a number of different measures were deve10ped to examine empathy in counseling. The Barrett-Lennard Relationship Inventory (BL RI) (Barrett-Lennard, 1962) attempted to measure empathy via the perceptions of a relationship's participants. The Truax Relationship Inventory (TRI) (Truax & Carkhuff, 1967) attempted the same. Both measures are paper and pencil instruments completed by persons in a counseling relationship. Feldstein & Gladstein (1980) noted that both are subjective measures clearly linked to Rogers' empathy construct (Rogers, 1957, 1959). Both also appear to focus upon the empathy cycle at Phase 3 (Barrett-Lennard, 1981). Objective measures using external, independent judgments of counseling mtewfiw behavior were also deve10ped (Feldstein & Gladstein, 1980). These include Truax's Accurate Empathy Scale (AE)'(Truax & Carkhuff, 1967) and Carkhuff's (1969) Empathic Understanding in Interpersonal Process Scale (EU). A more recently developed objective measure is the Revised Response Empathy Scale (RES) (Elliott, Reimschuessel, et a1., 1982). All three are attempts to 17 measure what Barrett-Lennard (1981) referred to as Phase 2 empathy. In contrast, the Affective Sensitivity Scale (ASS) (Kagan, Krathwohl and Associates, 1967), examines processes identified with Phase 1 empathy (Barrett-Lennard, 1981), but it does not examine those processes within any one particular relationship. The ASS is administered by asking subjects to respond to multiple choice lists of affect descriptions following videotape segments of an interview. Criticisms of empathy measures. A number of criticisms have been leveled at the commonly used measures of empathy. Chinsky and Rappaport (1970) focused upon Truax's AE scale. They questioned the meaning of the scale. They noted that a validation study of the AE (Truax, 1966, cited in Chinsky & Rappaport, 1970) found no significant differences between ratings of therapist behavior with client responses present and ratings with client responses absent. Yet the defining points of the AE scale Specifically included reference to client responses. Chinsky and Rappaport (1970) concluded that the AE scale was measuring qualities other than those reflected in the definition of the scale. Ffidman and Stone (1978) also found that the Carkhuff EU scale was not sensitive to different rater stimulus conditions. They found ratings of counselor behavior from conditions with client responses present to be comparable to ratings with client responses absent. Gormally and Hill (1974) questioned whether Carkhuff's EU scale was measuring a global counselor characteristic beyond those defined by the scale. They based this criticism first upon factor analyses which showed one factor accounted for 792 of the variance in one study and 892 in another. Secondly, strong correlations had been found between Carkhuff scales measuring empathy, warmth and genuineness and global therapist characteristics. ‘ Chinsky and Rappaport (1970) also noted that Truax & Carkhuff's (1967) 18 indicated an inverse relationship between reliability obtained for the AE scale and the number of therapists rated in any particular study. They speculated that this too might have been due to the scale measuring therapist qualities other than those Specified. Finally, Chinsky and Rappaport pointed out that the Pearson 3's reported as reliability figures by Truax (1966) were spuriously inflated by the use of an 3 in the calculations based upon the number of repeated measures of the same therapists rather than the number of therapists itself. Thus the assumption of independent measures was violated. Their criticisms, then, questioned both the meaning and hence the construct validity, of the AE scale and its reliability. Kurtz and Grummon (1972) compared empathy ratings using six different measures of counselor empathy. They compared these counselor empathy ratings to ratings of client perceived empathy and a well constructed outcome measure. Their data were collected from actual therapy cases. They found no significant correlations between the empathy measures, including the BLRI, Carkhuff's EU scale and Kagan's ASS. They concluded from the results that the six measures were assessing six different variables. In addition, no relation between the measures and either client perceived empathy or outcome was established, except for a moderate correlation between Carkhuff's EU scale and client perceived empathy, (Barrett-Lennard's (1981) Phase 3 empathy). The EU scale was the only one of the six empathy measures to be based upon actual therapist interview behavior. The conclusions of Kurtz and Grummon (1972) questioned the construct validities of all six empathy measures used. Based on these findings need for an empathy measure with demonstrated construct validity is indicated. Feldstein and Gladstein (1980) derived from Rogers' (1975) theoretical discussion six criteria applicable for evaluating the construct validity of 19 empathy measures. The three criteria they considered the most important included a focus upon a) cognitive empathy, b) affective empathy, and c) the counselor's communication of his or her empathic experience. Of the four measures reviewed: Carkhuff's EU, Truax's AE and TRI, and Barrett-Lennard's BL RI, none were found satisfactory according to their criteria. Criticism is warranted regarding the Feldstein and Gladstein (1980) review in light of Barrett-Lennard's (1981) description of the empathy cycle. It may be too much to expect for any one instrument to be able to measure all three phases of the empathy cycle simultaneously. To accomplish the above, simultaneous access to the counselor's internal experience, the counselor'stand client's interview behavior, and the client's internal experience would be required. However, different measures of the three phases as separate entities might be sought. One or more measures could tap into the cognitive and/or affective components of the counselor's empathic resonation at Phase 1, another could focus upon the counselor's expressed empathy at Phase 2, and yet another could look at the client's perception of received empathy at Phase 3. An. important step in the validation of the measures would be determining what relationships exist between measures of the different phases (Barrett-Lennard, , 1981). The RES (Elliott, Reimschuessel, et a1., 1982) has been held for discussion until this point largely because its deve10pment and validation are too recent for it to have been reviewed and appear in the critical reviews above. However, it was apparent that the authors took into consideration the preceding research critiques and theoretical reconceptualizations. A review of the RES demonstrated its relative strengths and limitations. 20 Response Empathy Scale The RES was developed (Elliott, Filipovich, et a1., 1982) and revised (Elliott, Reimschuessel, et a1., 1982) partly in response to Barrett-Lennard (1981). Some features of this new measure are: a) its specific focus on Phase 2, expressed empathy, b) its subscales measure different aspects of expressed empathy, c) it is designed for use with individual counselor responses. The reliability and validity of the RES rests upon two studies conducted by Elliott and his associates. Initial validity study. Elliott, Filipovich, et a1. (1982) attempted to investigate three issues regarding the RES. a) Estimates of rater reliability were sought at the level of the total RES and its subscales; b) Using factor analysis, they sought to determine the core elements and underlying components of response empathy when measured in this way; and, finally c) evidence for the RES' validity was sought in its relationship to clients' evaluations of being understood. The data were sampled fiom 28 thirty-minute interviews in which 28 volunteer undergraduate clients were asked to discuss 'a personal problem. The counselors were 12 internship level graduate students and three faculty members in clinical psychology. Counselors' theoretical orientations were described as client-centered or psychodynamic. Three samples of four to seven counselor responses (episodes) were videotaped within each interview. Five raters used the RES to rate counselor responses, rating each on the nine subscales before going on to the next response. An adaptation of Kagan's (1975) videotape recall technique was used to obtain client perceptions of being understood after the interview. Clients were asked how understood they felt after each counselor response and rated their answers on a 6-point scale. Counselor responses were rated separately using audiotapes and transcripts with the RES. 21 Interrater reliability for total empathy scores of the RES was .91 using Cronbach's alpha. A high degree of internal consistency for the RES was indicated by an interitem reliability for total empathy of .82 (Elliott, Filipovich, et a1., 1982). A Factor analysis extracted two factors accounting for 622 of the total variance. Factor 1, described as a Depth/Expressiveness factor, had strong loadings for the Words, Centrality and Here and Now subscales. It accounted for 442 of the total variance. Factor‘ 2, described as an Empathic Exploration factor, had strong loadings for the Frame, Centrality, Manner and Impact subscales. It accounted for 182 of the total variance. The factor analysis also provided a measure of which components were most central to empathy as measured by the RES. The Centrality subscale, followed by the Frame and the Words subscales were most central; Voice and Impact appeared to be the most peripheral (Elliott, Filipovich, et a1., 1982). Validity of the RES was investigated using correlations between observer ratings (RES) and client's ratings of being understood. Correlations were calculated at three levels of analysis: a) the individual counselor response, b) episodes of four to seven counselor responses, and c) all counselor responses in the three episodes per session. The relationship at the individual reaponse level was small but significant, 5 = .26. A medium sized effect was apparent for total empathy at the episode level, 5 = .42. The association increased again when the level of the session was examined, _r; =- .53 (Elliott, Filipovich, et a1., 1982). The presence of this relationship is strong evidence for validity of the RES. The investigators considered the RES to be measuring what Barrett-Lennard (1981) referred to as Phase 2, expressed empathy, and considered the clients' ratings to be measuring Phase 3, received empathy. Therefore a. relationship between measures of these two phases in the empathy cycle was expected 22 (Barrett-Lennard , 1981 ). Description of the revised Response Empathy Scale. In order to improve reliability and clarity, some changes suggested by Elliott, Filipovich, et a1. (1982) were incorporated into the revised RES. Two of the original subscales were drapped: Accuracy and Here and Now. The EXploratory Manner subscale was divided into two new scales: Collaboration and EXploration. The Client Frame subscale was redefined to become Client Feelings. The Expressiveness subscale was defined from the previous Voice Quality and Choice of Words subscale. A Verbal Allowing subscale was added (Elliott, Reimschuessel, et a1., 1982). The revised RES has eight subscales: Client Feelings, Perceptual Inference and Clarification, Topic Centrality, Expressiveness, Collaboration, Verbal Allowing vs. Crowding, Exploration, and Impact on Exploration. Each subscale is a 5-point (0-4) rating scale with descriptive anchors referring to specific counselor behaviors. The Allowing and Impact subscales can be used only with taped interviews. The six other subscales require only the initial client statement or response and the counselor's reaponse, and can be used with writtenresponses (Elliott, et a1., 1981). Validity study for the revised Response Empathy Scale. Following revision of the RES a second validation study was undertaken to assess the reliability of the revised scale when applied to a general sample of outpatient counseling. The design‘of this study was also influenced by Barrett-Lennard's (1981) empathy cycle model. In addition to collecting data on expressed empathy (Phase 2) and client received empathy (Phase 3), this study also investigated counselor empathic resonation (Phase 1) (Barrett-Lennard, 1981). Sixteen pairs of clients and counselors in a variety of outpatient settings were used as subjects. The counselors were divided in experience. One half of 23 the counselors were practicing psychologists, psychiatrists and social workers, while the other half were psychiatric residents and psychology interns. The clients ranged in age from 18 to 45 years. Any with severe diagnoses were excluded. Half of the dyads were sampled after brief contact (6-12 sessions); the other half were sampled after extended contact (20-130 sessions) (Elliott, Reimschuessel, et a1., 1982). The methodology was as follows. Sessions were audiotaped in three 5-10 minute segments. The counseling dyads were asked to carry on their customary inteviews. RES and client received empathy data were collected as in the previous validity study (Elliott, Filipovich, et a1., 1982). This time counselors also participated separately in a modified audiotape recall technique (Kagan, 1975) in which they rated a 7-point scale following each of the first four counselor responses in an episode. This rating was in response to the question "'when you said that, did you feel you misunderstood or understood your client" (Elliott, Reimschuessel, et a1., 1982)? The revised RES was shown reliable by an alpha = .95 for interrater reliability for total empathy. A high degree of internal consistency for the RES was demonstrated by an interitem reliability of alpha = .81. Factor analysis revealed three factors accounting for 742 of the total variance. Factor 1, described as Depth/Centrality, accounted for 482 of the total variance. Factor 2, described as an Exploration factor, and Factor 3, described as a Collaboration factor, accounted for 262 of the total variance together. The communality values of the factor analysis showed the Centrality subscale to be the most central component; only the Allowing subscale appeared to be peripheral (Elliott, Reimschuessel, et a1., 1982). The association between RES scores (Phase 2 empathy) and client received empathy (Phase 3) was examined. At the response level a small but significant 24 relationship between the RES total empathy score and clients' ratings of _r_'_ = .18 was obtained. This relationship was positive but insignificant at the episode and session level. At all three levels the Collaboration subscale was positively and significantly correlated with clients' ratings, 3 = .32, .40, .64 (Elliott, Reimschuessel, et a1., 1982). The association between RES scores (Phase 2) and counselor ratings (Phase 1) was also investigated. This association was significant and positive at both the response and episode levels, E = .34 and _r_' 3 .52 respectively. Only the correlation between the Collaboration subscale and counselor ratings attained significance at the session level. The lack of significance in the correlations between the RES total empathy score and counselor ratings at the session level, and with client ratings at the episode and session levels may well be due to declining statistical power due to reduced 3. The p was lowest at the session level, three times greater at the episode level and greater still at the response level. Summary of ReSponse Empathy Scale characteristics. The revised RES has high interrater reliablity and appears to measure what it purports to, with the possible exception of the Allowing subscale. Expressed empathy as measured by the scale is related positively to measures of counselor empathic resonation and client received empathy. This evidence for external validity has been shown in both analogue and actual counseling settings. The RES' measure of expressed empathy appears to consist of three underlying factors. These core dimensions are a) "the counselor addressing feeling-laden core client issues...by means of interpretive inferences communicated in an expressive manner"(Elliott, Reimschuessel, et a1., 1982), b) counselor behaviors fostering client exploration, and c) the counselor's communication of a collaborative effort. 25 An important attribute of the RES is its focus on individual counselor responses in terms of specific counselor behaviors. Because of this and its construct validity as shown by factor analyses, it avoids some of the difficulties linked to the Truax AE scale. In addition, its definition and demonstrated validity in relation to Barrett-Lennard's recent empathy cycle model is an advantage. There are some limitations. The RES does appear to be largely a measure of verbally communicated expressed empathy; this limitation was acknowledged by Elliott and his associates. Because of the way in which the scale is used, the independence of the subscale ratings relative to each other for individual responses is also not clear. Thus, it may be advisable to use the total empathy score obtained from the RES as the chief unit of analysis when following their method. The construct validity of the Allowing subscale is in question. Also the association between the RES and client received empathy at the session level has not yet been firmly established. A larger 2 may be needed to establish this relationship. Further validation studies could help to decide these issues. At present, the RES appears to have sufficient reliability and validity for its use in measuring total expressed empathy in counseling research. A reasonable unit of analysis would be the independently rated response or the mean of response ratings in an independently rated episode of counselor responses. The Use of Counseligg Analogues in Research On Empathy Research investigations of empathy frequently have employed experimental . tasks which are analogues of actual counseling situations. The present study used typewritten analogues to simulate clients at varying levels of deve10pment. Several research reviews and empirical studies which have focused upon counseling analogues are discussed below. The objective of this discussion is to 26 delineate the strengths and limitations of this type of research approach. Strong'_s and Munley's reviews. Strong (1971) discussed the potential usefulness to counseling theory and practice of experimental laboratory research. He pointed out that this methodology gave the researcher more control over variables of interest. This control could lead to the formation of causal inferences not possible in correlational methods. This could "provide more direct, unambiguous answers than would have been possible using traditional correlational methods and 'real' clients" (Strong, 1971, p.109). Strong stated that some phenomena of interest to the counseling profession could be investigated in a laboratory setting which could not be examined practically or ethically in a counseling setting. One obvious limitation of this methodology was the difficulty in generalizing its results to a naturalistic counseling setting (Strong, 1971). Strong (1971) noted that the relevance of this method was found either in 'its implications being considered important in practice or in its direct application to counseling. The chances of direct application were increased when the conditions under which the research took place approximated the conditions of the counseling setting. Strong listed five boundary conditions to be considered in estimating the similiarity of the research setting to practice and hence its genaralizability. 1. Counseling is a conversation between or among) persons....2. Status differences between or among interactants constrain the conversation....3. The duration of the contact between interactants varies, and at times is extended....4. Many clients are motivated to change....5. Many clients are psychologically distressed and are heavily invested in the behaviors they seek to change. (1971, p.108) Generalizability of research increased if its conditions matched these boundary conditions (Strong, 1971). Munley (1974) agreed that the major limitation of the analogue method of counseling research was the uncertainty of generalizability. But Munley stated 27 its chief advantage was in the method's ability to investigate "Specifics of the counseling process, eSpecially interaction effects between counselor characteristics and behavior and client characteristics" (1974, p.328). Because of the methodological advantages of analogue research, Munley (1974) stated that the need for it to simulate closely the naturalistic setting was of less importance than initially thought. However, in order to facilitate generalizability, Munley (1974) made some recommendations. He advised that the independent variables should be measured and manipulated with techniques and within a range similiar to that of the actual counseling setting. He suggested that the boundary conditions set out by Strong (1971) be considered in deciding relevance of analogue method results. In addition he recommended that research be undertaken comparing the results obtained in different kinds of analogues as well as between analogues and naturalistic settings (Munley, 1974). An example of this would be research comparing results of the same measure obtained in written, audio and video analogues. Munley (1974) reviewed analogue research methods in studies appearing in the Journal of Counseling Psychology. He classified these studies according to the nature of the analogue and dependent variable. The category most relevant to the present investigation is that of audiovisual analogues with a dependent variable of counselor behavior. Munley described the general format of this kind of analogue. The general experimental model of audio-visual studies, where counselor behavior is the dependent variable, is to (a) present a counselor or group of counselors with a client stimulus tape or film, (b) have the counselor subjects imagine themselves in the role of counselor to the client shown, and (c) ask for certain response behaviors at different points during the tape. Client behaviors and/or characteristics may be employed as independent variables and investigated in relationship to counselor behavior. (1974, p.321) Munley (1974) noted that this analogue method has some advantages. Large numbers of subjects could view standardized stimulus materials with '28 standardized experimental manipulations. The effect of client characteristics upon a large number of subjects could be studied. Finally, research methods not feasible in the actual counseling setting could be used. Munley (1974) also analyzed this category of analogue study in terms of Strong's (1971) boundary conditions. He noted that none of the studies reviewed met the boundary conditions of actual counseling. Counselors in these studies participated only vicariously in the relationship. While status differences were present, Munley considered their effect to be minimal due to the short length of contact. He noted that this short contact time may also limit the generalizability of the findings to first session counselor behavior. Finally, counselor responses in all studies were written, and some responses were restricted further by a multiple choice format. Comparison of written with audio and audio/visual analggpe formats. One recommendation of Munley's (1974) review was that a comparison be made between results obtained using different types of analogue formats. As the present investigation employed a written analogue, a review of research comparing written with audio or audio/visual formats was appropriate. Two studies compared the levels of expressed empathy obtained from counselors when presented with different stimulus materials. Fridman and Stone (1978) compared expressed empathy ratings using Carkhuff's EU scale from subjects presented with video, audio and written client stimuli. No significant differences in empathy ratings were found between stimulus conditions. Melnick (1975) .also investigated the effects of differing methods of analogue presentation. Subjects were graduate level counseling students who were presented with role-played, videotaped and transcribed analogue clients. Responses were rated using Carkhuff's EU scale for expressed empathy and for affective response and exploration. The results from the videotaped and 29 transcript conditions were positively related. Significant Pearson _r_"s were obtained between the sets of dependent variable ratings of .90, .60 and .73 respectively. Melnick concluded that "these two methods are interchangeable as stimuli to elicit counseling responses for evaluation purposes" (1975, p.111). Melnick did find that significantly higher empathy scores were obtained from the role-played condition. However, no significant differences were found between the three conditions in the ratings for affective. and exploratory response. Melnick (1975) also noted differences in the dependent variable according to whether the problem discussed was social-personal or vocational-educational. Two studies also compared different methods of obtaining counselor responses. Carkhuff (1969) reported significant correlations between empathy ratings for written reSponses and for oral responses. The relationship was found for high scoring subjects only. Therrien and Fischer (1978) extended Carkhuff's (1969) findings to both high and low scoring subjects with a group of adult subjects mixed as to education, age and amount of previous human relations training. Using the Truax AE scale, they compared ratings of written versus oral responses. The two groups of ratings were highly related with 5 = .93 over both high and low scorers. Stokes and Tait (1978) compared written and oral measures using undergraduate and trained paraprofessional counselor subjects. Responses were rated for level of reflective skills, verbal following and general quality. The two conditions were a transcript of a client with written responses asked for and a five minute videotaped client with oral responses. Strong positive correlations were found between the sets of dependent variable ratings from the two conditions. These correlations ranged from 5 = .53 to E = .869. The relationships were greater for the paraprofessional subjects. 30 Hardin and Yanico (1981) compared naive subjects' perceptions of a male counselor's characteristics in a videotaped, audiotaped and transcribed analogue of the counselor interacting with a client. No significant main or interaction effects of mode of presentation upon four dependent variables were reported. In this way the different methods were shown to be comparable. These variables were counselor attractiveness, subject willingness to talk with the analogue counselor, interest in the interview and identification with the client. Some sex-related differences were noted although the study's design did not permit complete exploration of this effect. Male subjects rated the transcribed counselor as less trustworthy than the counselor in other conditions and lower than women subjects did. Additional research systematically varying the sex of counselor and client would be needed to explore this effect completely. One implication, though, is that raters of written counselor responses ought to be blind to the sex of the counselor in order to avoid this effect. Summary. The use of analogue methods in counseling research has the chief advantage of permitting experimental control not usually possible in practice. This is particularly true when investigating interaction effects between counselor and client characteristics (Munley, 1974). The method's chief limitation is the uncertain generalizability of results. This generalizability increases as the conditions under which the research is conducted approach the conditions of actual counseling practice. Different analogue formats have been compared and written methods of stimulus presentation and response recording have been shown to be comparable to audio and audio/visual methods. Ego Development and Empathy The Theoretical Relationship Carlozzi, et a1. (1983) stated that empathic ability might be expected to be related to level of ego development. They based this hypothesis upon the 31 structure of Loevinger's (1976) developmental framework. The hierarchical sequence of ego development suggests increasing social sensitivity, maturity, self-other differentiation, and decreasing self-centeredness as one progresses to higher stages. Furthermore, ability to empathize, which consists of suspending judgment and detecting and responding to the feeling expressed by another, is conceptually related to impulse control and interpersonal relations. (Carlozzi, et a1., 1983, p.113) Carlozzi, et a1. (1983) also noted that empathy was considered to be related to Piaget's theory of cognitive development and Rohlberg's theory of moral development. Both of these theories have been conceptually and empirically related to Loevinger's theory of ego development (Loevinger, 1976; 1979). Yet, few studies have been conducted in which an empirical relationship is examined between empathy and ego development (Carlozzi, et a1., 1983). Swensen (1980) viewed ego development theory as having the potential to provide a general model for counseling and psychotherapy. He noted, however, that the relationship between ego development level and any given piece of behavior is not likely to be simple. Swensen recommended that in seeking to apply ego development theory to counseling an approach be taken which takes into account both personal characteristics as exemplified by ego deve10pment level and characteristics of the enviornment. He found the origin of this approach in Lewin's B = f(P,E) formula. Recent research which has examined the interactions between person and enviornment in this way follows the aptitude x treatment interaction (AT1) approach advocated by Cronbach (1957). Swensen (1980) also stated that "another question...raised by the concept of ego development...is the optimal differences in ego stages between the therapist and client" (p.387). This was a question which Loevinger (1976) raised. She cited the work of Grant and Grant (1959) as indicating the differential effectiveness of therapists and juvenile delinquent combinations according to the ego level of those involved. Using their own concept of ego development, Grant and Grant (1959) found that: 32 subjects with high ego level responded best to teams of higher predicted- efficiency, but the relation was reversed for offenders of low ego 1evel...ego level was presumably one component in predicting effectiveness. (Loevinger, 1976, p.428) Swensen (1980) hypothesized that therapists of lower ego level would be less able to understand their clients than those therapists of higher level. This was because they "transform the concepts of more complex levels into concepts that are at their own (simpler) level" (Swensen, 1980, p.387). Swensen also speculated that the optimal combination might be for a therapist or counselor to be at a stage of ego development one level higher than the client, although he did not explain this. Presumably, he spoke of the theory of the pacer, common in cognitive deve10pmental literature, which Loevinger (1976) mentioned. Loevinger (1980) agreed with Swensen (1980) that therapists should not be at a lower ego level than the clients they work with. But she left open the question of optimal match or mismatch, with the therapist one stage higher, as Swensen suggested. Loevinger did see the need for therapists to "develop a deeper understanding of clients whose own level is quite far below their own" (1980, p.389). It appears from this viewpoint there may be something which gets in the way of a more advanced therapist deeply understanding a client at a stage of development which the therapist has long since passed through. Young-Eisendrath (1982) followed up on Loevinger's viewpoint in the development of a training program for social work students in which they learned about characteristics of persons below their own level of deve10pment. She reported anecdotally that post-conformist social work students "dislike the person who sees the world from the conformist point of view and resist being empathic with such a client" (Young-Eisendrath, 1982, p.332). Young-Eisendrath's observation focused not only on the change in empathy due to ego level mismatch but the factor of preference or liking as well. Greenson's writings (1967, 1978, cited in Gladstein, 1983) provided 33 additional justification for expecting a positive relationship between ego deve10pment and empathy. While he stated that emotional contagion or feeling like the client is part of being empathic, he also stated that a "capacity for controlled and reversible regressions" (Greenson, 1967, p.369, cited in Gladstein, 1983) was essential, and that without it empathy lost its effectiveness. Loevinger (1976) clearly placed regressions in the service of the ego as a characteristic of post-conformist levels of ego development. Therefore, if this regressive ability aids empathy, as Greenson (1967) believed, those persons at higher ego levels would be expected to be more empathic. Empirical Evidence for the Relationship ofigo DeveIOpment to Empathy Ego deve10pment studies. On the basis of theory an empirical link was suspected between ego development level and empathy. Such a link has been tentatively established. Zielinski (1973, cited in Loevinger, 1979) found a positive moderate correlation between the ego development level of graduate students in a beginning counselor education course and their ability to communicate empathic understanding as measured by Carkhuff's Index of Communication. Carlozzi, et al. (1983) found that dormitory advisors whose own ego development is at the I - 3/4 level and above scored higher on Kagan's ASS, a measure of empathic resonation (Barrett-Lennard, 1981), than advisors at the I-3 level or lower. Young-Eisendrath (1981, cited in Young-Eisendrath, 1982) investigated the relationship between counselor's ego level, client ego level and measures of empathy in a client analogue study. Subjects were 34 recent graduates and students in Master's level counseling and social work programs who participated in either empathy skills training or training in ego deve10pment. Subject ego level was assessed with the WUSC T. The client analogues were two 45-minute videotapes of clients. One client had been assessed via the WUSCT at the 34 conformist stage. The other had been assessed with the WUSCT as post-conscientious (I - 4/5 level or greater). The clients' videotapes were considered exemplary of their ego deve10pment functioning. After watching the videotapes, subjects were asked to respond to the Rokeach Value Survey and a forced choice questionnaire as the client might. They also wrote brief essays describing the clients. These essays were analyzed for agreement with the ego deve10pment scheme and for ability to synthesize into a whole picture. The only significant group difference found was a higher agreement with the conformist client on the value survey by the experimental group. Results indicated that post-conscientious subjects described the client analogue at the post-conscientious level better than did other subjects. Young-Eisendrath concluded that problems with the measures may have contributed to the lack of significant findings. This was not surprising as these measures were rarely used empathy measures. Kurtz and Grummon's (1972) results questioned the use of similiar techniques based on their lack of relationship to either counseling outcome or client-perceived empathy. In any case, Young-Eisendrath's measures could be considered measures of Phase 1, empathic resonation (Barrett-Lennard, 1981) and of comprehension of ego development stage-related functioning. Two studies have investigated subjects' comprehension of the ego functioning of the various stages. Redmore (1976) found that low-level subjects couldn't fake higher level ego development on the WUSCT even when asked following instruction on ego development concepts. Higher level subjects were able to respond to the WUSCT with lower level responses, however. Blasi (1976) also showed that lower level subjects could not comprehend roles belonging to higher ego levels. From this "asymmetry of comprehension" (Loevinger, 1979, p.292) it could be expected that'lower level counselors would show decreased empathy when interacting with higher level subjects, compared to their empathy 35 with equally low level subjects. These results were in keeping with those of investigations of comprehension of stages of moral reasoning. Rest, Turiel and Kohlberg. (1969) found that subjects had increasing difficulty comprehending moral reasoning indicative of stages above their own. Rest, et a1. (1969) also found that subjects tended to prefer the moral reasoning of the highest level which they could comprehend. These results might be extended to the realm of ego development with the hypothesis that a person's preference for another's ego functioning and hence empathy might vary according to the interaction of ego levels. Indeed, an incidental finding of the Young-Eisendrath (1981, cited in Young-Eisendrath, 1982) study of empathy and ego deve10pment was that subjects' preference for the post-conscientious client analogue and stereotyping of the conformist client analogue significantly affected the subjects' ability to infer the clients' frame of reference. Mullen and Abeles (1971) investigated the relationship between empathy, liking and therapy outcome. They found that liking as measured by the Nonpossesive Warmth Scale was unrelated to outcome. But they argued that the scale used indicated a general level of liking, instead of a therapist's liking for a specific client. It is obvious that measures indicating specific liking would be necessary in investigating liking or preference as a possibly mediating factor in empathy. Empathy and conceptual systems theory. The developmental theory articulated by Harvey, Hunt and Schroeder (1961) has spawned a large body of research studies (Hunt, 1977). Among these have been studies which have investigated the effects of a counselor's conceptual level upon his/her empathic behavior and other interview behavior. As there have been few reported studies on the relationship between ego deve10pment and empathy, a review of this parallel research from the viewpoint of a related developmental theory may be useful. 36 Goldberg (1974) used an analogue counseling task with master's level counseling students and demonstrated a significant difference in the subjects' response patterns as a function of their conceptual level. Focused in more clearly on expressed empathy, Lutwak and Hennessey (1982) showed that the empathic responding in interview behavior of undergraduate and graduate students was significantly and positively correlated to the subjects' level of conceptual functioning. Rimberlin and Friesin (1980) provided indirect evidence of an association between ego development and expressed empathy in a study involving conceptual level. They investigated subjects' empathic responses to ambivalent affect in analogue research. Ambivalent affect was considered by Loevinger (1976) to be indicative of post-conformist levels. In Kimberlin and Friesin (1980) undergraduate "helpers" responded to videotaped analogue clients which demonstrated ambivalent and non-ambivalent affect. Subjects' responses to the analogues were rated using Carkhuff's EU scale.~ Low conceptual level subjects showed a significantly lower level of expressed empathy in response to ambivalent affect. Another study addressed the matching and mismatching of the conceptual level of counselors and client analogues. Heck and Davis (1973) demonstrated that levels of empathic behavior could be significantly affected by the interaction of counselor and client analogue developmental levels. In this study, master's level counseling students responded to two analogues consisting of 12 client responses each. One analogue was of high conceptual level and one was of low conceptual level. The counselors were rated at high or low conceptual level independent of the experimental task. A two-way ANOVA for repeated measures yielded a significant main effect for counselor level and a significant interaction effect across counselors and analogues. The mean empathy ratings 37 revealed the nature of this interaction. High level counselors were rated high in expressed empathy with high level analogues, but were rated lower with low level analogues. Low level counselors were rated moderately for expressed empathy with low level analogues but were rated lower with high level analogues. This study showed both a main effect and an interaction effect related to counselor and client analogue developmental levels. As this deve10pmental construct is related to ego development (Loevinger, 1976, 1979) the results lead to the hypothesis that both a main effect and an interaction effect might be expected due to the impact of both counselor and client ego deve10pmental levels upon expressed empathy. Summary of‘ELo DevelOpment and Empathy Ego development has been related theoretically and empirically to empathy. There have been few studies in this area, however, and none focused Specifically on the interaction between ego levels of counselor and client and the resulting effect upon expressed empathy. Loevinger (1976) and Swensen (1980) cautioned against expecting a simple relationship between ego level and behavior. Swensen advocated a research strategy similiar to Cronbach's (1957) ATI approach. Hauser (1976) agreed with this strategy. Rather than predict correlations between ego deve10pment level and behavior, it may be more fruitful to look for behaviors which are generated by the interplay between certain situational cues and certain ego deve10pment levels....Links between ego development level and action can be effectively studied only by experimental conditions and dependent variables, which are based on specific predictions derived fi'om theoretically described characteristics of each stage. (p.939-940) Systematic variation in the expressed empathy of counselors due to the interaction between counselor and client level of ego development has been theoretically predicted. Studies which have investigated the relationships between ego deve10pment and empathy have not yet taken the form nor focus 38 with which to study this interaction empirically. It is this interaction between internal frames of reference of the participants in a counseling relationship and its resultant impact upon the counselor's ability to respond empathically which was the focus of this investigation. 7 Summary of the Review of Literature Several areas of theory and research formed the basis for this investigation. The theory of ego development has described the internal frame of reference through which an individual perceives and interprets the world. This theory has been complemented by an assessment technique with sufficient reliability and validity for research purposes. Theory and research was also reviewed above dealing with the empathy cycle. A counselor's behavioral manifestation of empathy, expressed empathy, occurs within this cycle following the counselor's internal empathic resonation. The content of the empathic resonation may be affective and/or cognitive in nature. Research into the empathy cycle has used a variety of measures. The RES, one such measure, has been shown to have sufficient reliability and validity for its use in research. Research into the empathy cycle also has used analogues of counseling interaction. The use of counseling analogues limit a study's generalizability, but also allow the investigation of interactions between variables not possible otherwise. The theories of ego development and empathy have been related conceptually because both deal with an individual's internal frame of reference. This investigation examined the effects of the interaction between counselor and client levels of ego deve10pment upon a counselor's expressed empathy and upon a counselor's expressed preference for a client. The client analogues used in this study allow the investigation of this interaction. The kind of expressed 39 empathy focused upon was verbally expressed, cognitive role-taking empathy. This investigation's hypotheses had their source in the literature reviewed above. A positive relationship was hypothesized between a counselor's level of ego development and a counselor's level of expressed empathy. Theoretically, this positive relationship was expected because Loevinger's (1976) description of the stages of ego deve10pment characterized individuals at higher stages as possessing greater openness, conceptual compleidty and ability to regress in the service of the ego. Empirically, this expectation was supported by the investigations of Carlozzi, et a1. (1983) and Heck and Davis (1973). These studies showed a positive relationship between empathic resonation and ego development, and between expressed empathy and conceptual level, respectively. The expressed empathy of a counselor also was hypothesized to vary among the ego development levels of the client analogues. These analogues were constructed so as to represent clients demonstrating internal frames of reference which varied qualitatively. It was expected that these four different stimuli, the client analogues, would elicit different behavior, expressed empathy, from the counselors. The hypothesis that an interaction between counselor and client levels of ego development would be evident in the levels of expressed empathy was based on the assumption that a counselor would demonstrate the highest level of expressed empathy when interacting with a client of equal ego development level. In part this hypothesis was based upon Heck and Davis' (1973) finding of a similiar interaction. between the conceptual levels of counselor and client analogue. This assumption also resulted from two expectations. A counselor was expected to manifest lower eXpressed empathy with a client of higher ego deve10pment level. This was because it would be difficult for an individual to comprehend the more complex frame of reference used by another person at a 40 higher level of ego deve10pment (Loevinger, 1976, 1979; Redmore, 1976; Blasi, 1976; Young-Eisendrath, 1982). A counselor was also expected to manifest a lower level of expressed empathy when interacting with a client at a lower level of ego development than the counselor's own. This would be perhaps because of the counselor's asymmetry of preference, that is, a counselor would prefer the highest level of ego deve10pment which that counselor could comprehend. The counselor's comprehension would be limited by the level of ego deve10pment which the counselor had reached. A similiar asymmetry of preference was observed in the case of moral development (Rest, et a1., 1969), a theory which has been related both conceptually and empirically to ego development theory (Loevinger, 1976; 1979). Evidence of an asymmetry of preference involving ego development was observed anecdotally (Young-Eisendrath, 1982) and as an incidental finding to an empirical investigation (Young-Eisendrath, 1981, cited in Young-Eisendrath, 1982). For these reasons, a counselor also was expected to indicate a lower preference for a client whose level of ego deve10pment was lower than or more than one level higher than the counselor's own level. In Chapter 3 the method by which these hypotheses were tested is described. The results of these tests are presented in Chapter 4 and discussed in C hapter 5. Chapter 3: Design of the Study This study investigated the relationship between counselor and client levels of ego development and the expressed empathy and expressed client preferences of counselors. The subjects, Master's level counseling students, completed a measure of ego deve10pment and responded in writing to four typewritten client analogues. These analogues were previously standardized at four sequential levels of ego deve10pment. Independent ratings by pairs of trained raters provided indices of both ego development and expressed empathy. The subjects also rank ordered their preferences for the client analogues. The eXpressed empathy data were analyzed with a 3 x 4 repeated measures Analysis of Variance; the preference data were analyzed with a Chi square. Sample Subjects were 42 masters level counseling students at a large midwestern university. The 31 women and 11 men ranged in age from 22 to 50 years. Median age was 27 and the mean age was 28.56 years. Subjects were initially contacted by way of a letter which was distributed in approximately six graduate counseling classes. The letter requested their participation in counseling psychology research but did not reveal the nature of the hypotheses. Subjects indicated interest by returning the letter with their name and phone number attached. It was apparent that a number of potential subjects who read this letter declined to participate. Those subjects indicating an interest were then contacted and they completed the instruments individually and in small groups as scheduling permitted. Procedure Subjects completed the instruments in two sessions, each lasting approximately 45 minutes. In the first session subjects completed the sentence completion test. In the second session subjects responded to the four client 41 42 analogues and rank ordered their preferences for the clients. Measurement of Ego Development Washinixon University Sentence Completion Test (WUSCT). Form 11-68 Women and Form 11-68 Men of the WUSCT were used to assess ego development level. This sentence completion test consisted of 36 sentence stems on three pages which the subjects were asked to complete (see Appendix A). Scoring of sentence stems. Two raters scored the WUSCT items separately. One rater was a post-internship doctoral student in counseling psychology; the other was a doctoral level psychologist. Both were eXperienced with projective tests. Both raters completed Loevinger's self-training program (Loevinger, et a1., 1970) before beginning the scorings. WUSCT items were prepared for scoring following Loevinger's (1970) instructions. All items from the same stem were typed in random order on the same page or two. Code numbers were removed during scoring to ensure independent ratings. Scoring manuals were consulted during the scoring process (Loevinger, et a1., 1970; Redmore, Loevinger & Tamashiro, 1978; Browning, 1978). Interrater reliability for the Washington University Sentence Completion E. Reliability for the total protocol ratings (TP R) obtained from the item ratings of each rater was calculated using intraclass correlations ggg) (Schrout & Fleiss, 1979). Although Loevinger, et a1. (1970) used product moment correlations for this index, Holt (1980) argued that the most appropriate method for determining interrater reliability for an ordinal scale such as Loevinger's ' was with an _I_C__C_. Loevinger, et a1. (1970) reported a median interrater reliability for TP R's of}: = .86. Interrater reliability for the TPR's in the present study was& = .773. 43 Total protocol rating. Following standard procedure, raters reached agreement on discrepant item scores before final TP R's for each subject were determined. Loevinger, et a1.'s (1970) automatic ogive rules were used to determine the TPR from a frequency distribution of item scores for each subject's WUSC T. The distribution of TP R's in the present study may be found in Table 3.1. TABLE 3.1 Distribution of Ego Development Levels of Subjects E go Development Level I - 3/4 . _I__-_4 I - 4/5 Subjects W omen 9 12 10 Men . 5 ' 2 4 Total 14 14 14 Client Analogues Development. Four client analogues (see Appendix B) were developed and revised in a pilot study following Loevinger's (1976) descriptions of ego development levels I - 3, I - 3/4, I - 4, and I - 4/5. Each analogue began as an intake report similiar in form to that used in some university counseling centers. Following lengthy quotations from the analogue client was a space for the first counselor response. Following this first response space were two more client statements and counselor response spaces. This second and third client statement were randomly sorted as a pair across the analogues which the subjects responded to. In this way the description and initial statements of the 44 client were the independent variable, and the last statements themselves were not. The analogues were written to be characteristic of the impulse control, character development, interpersonal style, conscious preoccupations and cognitive style of a woman at a particular level of ego development. There were, however, several common elements to all the analogues. Each client was a woman university student who discussed her interpersonal presenting problem, one other relationship, and her views about self, school and career. The analogues were approximately equal in length. Ratings of analpgue developmental levels. Client analogues were read and rated for ego development level prior to use by two raters other than the author. One rater was a counseling psychology doctoral student; the other was a doctoral level psychologist. Both raters had been instructed in ego development stage theory and made their ratings independently and in randomized order. Ratings were in 1002 agreement with each other and with the intended deve10pmental levels. Analogues were rated at the I - 3, I - 3/4, I - 4, and I - 4/5 levels of ego development. Presentation of analcfiues to subjects. Counselor subjects were given the four analogues together in randomized order and asked to read and respond to each client in the space provided as if he or she were that woman's counselor. Measurement of Expressed Empathy Revised Re3ponse EmLathLRatinLScale. The RES (see Appendix C) was used to obtain a measure of expressed empathy for each subject's responses to each analogue. The six subscales used were: Client Feelings, Perceptual Inference and Clarification, Tapic Centrality, Expressiveness, Collaboration, and Exploration. The Verbal Allowing vs. Crowding and Impact on Exploration subscales were not used as they necessitated a response to the counselor's 45 response from a 'live' client. Each subscale has a five'point (0-4) rating scale with descriptive anchors referring to specific counselor behaviors. The mean of the two raters' total scores for the six subscales across the three counselor responses to each analogue were used in the statistical analysis. Thus, each subject's score had a possible range of 0 - 72. Response Empathy Scale ratings. Subjects' responses were put in 16 groups according to the analogue level and the particular pair of client statements at the end. These responses were then randomly sorted within groups and the order of the groups randomized for each rater. Raters scored them independently and blind to the subject's developmental level. The two raters were doctoral counseling psychology students. Both had previously used a similiar scale (i.e., Carkhuff's EU scale) and were trained with the RES until familiar with it. Response Empatlm Scale interrater reliability. Elliott, Filipovich, et al. (1982) and Elliott, Reimschuessel, et a1. (1982) reported interrater reliabilities in the RES' validation studies of alphas .91 and alpha = .95 respectively for the total RES score. In the present investigation interrater reliabilities were calculated for the RES using a Pearson 5 for the ratings of responses to each analogue. Reliabilities were _r_ = .778 (I - 3 analogue), 5 = .832 (I - 3/4 analogue), 5 = .823 (I - 4 analogue), and 5 = .875 (I - 4/5 analogue). The mean reliability was E = .827. The study's design was comparable to that'of an aptitude x treatment interaction design (Cronbach, 1957). Both subject level and analogue level of deve10pment were considered fixed independent variables. The expressed empathy score from the RES was the dependent variable. The effect of subject level (at three levels) and analogue level (at four 46 levels) upon the outcome behavior of expressed empathy was examined. Subjects at all levels reaponded to client analogues at all levels. Testable Hypotheses Hypothesis 1. Null hypothesis: no difference will be found in expressed empathy of subject responses as measured by RES total empathy rating between the subject groups at levels I - 3/4, I - 4, and I - 4/5. Alternate hypothesis: The I - 4/5 group mean RES total empathy score will be greater than the I - 4 group mean; the I - 4 group mean will be greater than the I - 3/4 group mean. This was a directional hypothesis. Hypothesis 2. Null hypothesis: No difference will be found in the RES total empathy scores between subjects' responses across the four analogues. Alternate hypothesis: There will be differences in the RES total empathy scores for subjects' responses among the four analogues. This was a non-directional hypothesis. Hypothesis 3. Null hypothesis: The ego development level of the subjects will not have an interactive effect with the ego development level of the analogues upon the RES total empathy scores of the subjects' responses. Alternate hypothesis: The ego development level of the subjects will have an interactive effect with the ego deve10pment levels of the analogues upon the RES total empathy scores of the subjects' responses. Hypothesis 4. The subjects will not demonstrate a high or low preference for any analogue relative to the other analogues as demonstrated by the subjects' rank ordering of the analogues by preference. Alternative hypothesis: The subjects will rank higher those analogues at the subjects' own deve10pmental level or a level one step advanced compared to the other analogues. 47 Statistical Analyses Analysis of variance (ANOVA). A 3 x 4 ANOVA for repeated measures (Winer, 1962) was computed to test Hypotheses 1, 2, and 3. The F test for Factor A was the test for Hypothesis 1; the _I: test for Factor B was the test for Hypothesis 2; and, the _F_ test for Factor AB (interaction) was the test for Hypothesis 3. A Neuman-Keuls post hoc analysis was performed in order to ascertain the source of the variance within factors (Myers, 1979). The strategy of using an ANOVA and focusing upon the significance of the interaction between client analogue and counselor levels of ego development was selected to facilitate comparability with the one study most similiar to the present study (Heck 8 Davis, 1973). A multiple regression approach could have been adOpted for this analysis with total mathematical equivalence (Cohen 8 Cohen, 1975; Ward 8 Jenkins, 1973). If the ego development levels were of a ratio or interval quality and if the process of forming the levels of A and B in the ANOVA resulted in the collapsing of distinguishable classes into a class equatable to a level, then the linear model of multiple regression would have led to increased power and therefore greater parsimony. The theory and measurement technique of ego development, however, results in ordinal values at best and arguments that the scale is nominal are tenable (Allen 8 Yen, 1979). In addition the present study does not involve the loss of variability through the collapsing of data into classes. Therefore the ANOVA and multiple regression approaches are equivalent from a mathematical perspective, but the ANOVA approach has greater conceptual clarity given the structure of the data. Analysis of preference data. A Chi sguare was computed to test Hypothesis 4. An additional Chi sguare was computed to examine other overall trends in 48 the preference data. Summary of Method The hypothesis of chief interest was Hypothesis 3, regarding the interaction effect upon counselor behavior of a counselor and client characteristic. In order to allow a test of this hypothesis, an analogue method of research was selected. This method was expected and had previously been shown to allow for the test of similiar interactions. The client analogues were developed and standardized at a range of ego development levels which would commonly be found among late adolescents and adults seeking counseling. The counselors chosen as subjects would likewise often see clients at these levels in practice. The form of the analogues was meant to be comparable to an initial counseling contact. It is thought that in initial contact the dependent variable, expressed empathy, isian important ingredient in successful counseling for social/personal issues similiar to the analogues' presenting problems. Measures were used in a reliable fashion which have demonstrated reliability and validity. Statistical analyses were selected which provided powerful and conceptually meaningful tests of the hypotheses. The overall design of the study was one which allowed a test of the hypotheses with very high internal validity . Chapter 4: Results The results of the statistical analyses which allowed tests of this study's hypotheses are presented in this chapter. Initially, the values obtained from the 3 x 4 Analysis of Variance (AN OVA) are examined. The factors for main effects and interaction of the ANOVA are then discussed along with apprOpriate post hoc tests. The outcomes of the hypotheses dealing with expressed empathy are also decided here on the basis of the ANOVA. Finally, the results of the gh_1 Sguare are presented, along with a post hoc test, and the test of the fourth hypothesis decided. Expressed E m pathy A 3 x 4 fixed effects Analysis of Variance (AN OVA) for repeated measures tested Hypotheses 1', 2, and 3. The AN OVA summary can be seen in Table 4.1. Table 4.1 ANOVA Summary df MS F M A (subjects) 2 124.79 1.95, ns S/A 39 63.87 B (analogues) 3 68.05 3.61, p < .05 AB (interaction) 6 50.57 2.69, p < .05 SB/A 117 18.84 Test for Hypothesis 1. The F test for Factor A (subjects) did not reach significance at the p < .05 level. Therefore the alternative hypothesis was rejected in favor of the null hypothesis. The Response Empathy Scale (RES) 49 50 total empathy scores did not differ significantly between subject groups at the I - 3/4, I - 4, and I - 4/5 levels of ego development. The actual differences between the means were in the expected order, however, as seen in Figure 4.1. RES Scorei 44 43.5 43 43.05 42.5 42 42.08/ 41.5 41 40.5 40 40.12 39.5 Subject Ego Develoiment I - 3/4 I - 4 I - 4/5 Level Note: Non-significant differences in means (p > .05) Figure 4.1 Mean RES Total Empathy Scores by Subject Level Test for Hypothesis 2. The F test for Factor B, E = 3.61, was significant at the p < .05 level. Therefore, the null hypothesis was rejected and the alternative hypothesis was accepted. The RES total empathy scores for the subjects did differ across the analogue levels I - 3, I - 3/4, I - 4, and I - 4/5. Test for Hypothesis 3. The _F_ test for Factor AB (interaction), 2: = 2.69, was significant at the p < .05 level. Therefore the null hypothesis was rejected and the alternative hypothesis was accepted. The ego development level of subjects interacted with the ego deve10pment level of analogues. An examination of the cell means (Table 4.2, Figures 4.2, 4.3) and the results of the post hoc analysis shows the nature of the above results. 51 Table 4.2 Mean RES Scores For Subjects By Subject Level and Analogue Level Ego DeveIOpment Level of Analogue 1-3 1-3/4 1-4 1-4/5 Supject Level I - 4/5 40.21 42.57 44.95 44.47 I - 4 42.30 41.22 44.18 40.63 I - 3/4 38.61 42.24 41.00 38.63 All Subjects 40.37 42.01 43.38 41.24 RES Scorg 44 43.5 4 . 8 43 42.5 42 4 01 41.5 41 41.24 40.5 40 40.37 39.5 Ego Development I - 3 I - 3/4 I - 4 I - 4/5 Level of Client Analogue Note: Differences in means were significant (p < .05) Figure 4.2 Mean RES Score Across Subjects By Analogue Level 52 Post hoc analysis of Factor B (analogues). A visual examination of the data in Figure 4.2 indicates the sources of the statistically significant main effect B among analogue levels. The mean RES total empathy scores are most different betwen analogues I - 3 and I - 4. The means for analogues I - 3/4 and I - 4/5 are most similiar and fall between the means of the other two analogues. A Neuman—Keuls post-hoc analysis for Factor B showed that the mean RES total empathy scores for analogues I - 3 and I - 4 are significantly different from each other and thus contributed the most to this main effect. No other significant differences were revealed by this post hoc test. Post hoc analysis for Factor AB (interaction). A visual examination of the mean RES scores in Figure 4.3 and Table 4.2 reveals the nature of the significant interaction. For all subject groups the highest mean RES scores were in response to the analogue at the developmental level correSponding to that subject group's level. A post hoc test of simple main effects of the interaction provides a further breakdown of this interaction. First the results were examined by subject group. The only subject group showing significant differences between mean RES scores among analogues was the I - 3/4 group. This group had a significantly lower score with analogues I - 3 and I - 4/ 5 than with analogues I - 3/4 and I - 4. One other group approached significance with this test. This was the I -'4/5 subject group which had a lower mean RES total empathy score with the I - 3 analogue than with the three other analogues. These two differences can be seen as the chief contributors to the significant interaction for ANOVA Factor AB although all differences contributed to the overall interaction. A test of simple main effects was also performed by analogue. The only significant difference with this test was for analogue 4/ 5. With this analogue subjects at level I - 3/4 had significantly lower RES total empathy scores than 53 subjects at level I - 4/5. It is only with analogue I - 4/5 that a main effect similiar to that hypothesized for Factor A is apparent at a statistically significant level. The nature of the interaction prevents this effect from attaining significance overall in Factor A of the ANOVA. RES Score 45 44.5 44 43.5 43 42.5 42 41.5 41 40.5 4/5 .‘ 4o 39.5 39 38.5 3/4 374 38 Ego Development Level of I — 3 I - 3/4 I - 4 I-- 4/5 Analogue Note: Above interaction was significant (p < .05) Figure 4.3 Mean RES Total Empathy Scores by Subject Level 54 Subjects' Preferences for Client Analogues Test of Hypothesis 4. A Chi sguare (Table 4.3) was computed to examine the differences in preference of the counselor subjects for the client analogues at the correSponding ego level and for the level one step advanced compared to the other client analogues. The result was Elli. = 4.457 (if = 1; p < .05). Therefore the null hypothesis was rejected in favor of the alternative hypothesis. Subjects indicated a higher preference for those analogue clients at the level correSponding to the subjects own or at a level one step advanced compared to other levels. Due to ommissions by three subjects an 3 I 39 was used. Table 4.3 Chi Square 0 E. O E 39 32.5 26 32.5 39 45.5 52 45.5 Chi Sguare = 4.457, (d_f =- 1; p < .05) Post hoc analysis of the preference data. A second Chi sguare revealed the overall trend in this data. This was a 2 x 4 Chi Sguare (Table 4.4). The result was EL].- = 34.77 (d_f = 3; p < .001). This highly significant result showed a significant difference in preference overall. The main contributor, as determined by the cell contributions (Table 4.5), to this difference was that all subject groups ranked client analogue I - 3 significantly lower than the other client analogues. The conformist client analogue was significantly less preferred than the post-conformist client analogues by the post-conformist counselor subjects. 55 Table 4.4 Post Hoc Chi Square O E O E 4 19.5 35 19.5 28 19.5 11 19.5 24 19.5 15 19.5 22 19.5 17 19.5 Chi Sguare = 34.769 (d_f = 3; p < .05) Expressed Empathy by Sex of Subjects Malesubjects in the present study had a mean RES total empathy score across analogues of 41.15 (p = 11). Female subjects had a mean RES total empathy score across analogues of 41.97 (p = 31). Although this study's design did not permit a test of significance of this difference post hoc, the difference in means appeared small considering the magnitude of the scores and the size of other differences obtained. Table 4.5 Cell Contributions To Post Hoc Chi Scmare I - 3 12.32 12.32 Analogue I - 3/4 3.7 3.7 Level I - 4 1.04 1.04 Chapter Five: Discussion This chapter is concerned with the interpretation of the results of this study. Because of the nature of the design, an analogue counseling situation, the question of generalizability will be addressed first to prepare for a focus upon specific findings as viewed in the light of the generalizability issue. The findings will be reviewed one by one: the absence of a main effect for subject level; the main effect for analogue level; the interaction; and, the trends in subject preference. Reasons for these findings and implications Specific to individual findings will also be explored. Finally, the overall implications of this study for practice, training and research will be discussed. The Parameters of Generalizability The range of ego development levels. The subjects, Masters level counseling students were assessed to be at levels I - 3/4, I - 4 and I - 4/5 of ego develOpment. As these were the subject levels sought and expected, these levels are considered fixed by the design and therefore, not random effects. It is possible, although less likely, to find a counselor whose level falls outside of this range (e.g. Delta or I - 5). This study does not address the behavior of counselors outside of the fixed range chosen and it is a limitation to the study's generalizability. However, the fixed range chosen is also a strength, as a large percentage of counselors in training and in practice may be found at levels I - 3/4 through I - 4/5. Holt (1980) showed 70% of young adults with some college eXperience at one of these three levels. An additional limitation to the study's generalizability is the experience level of the subjects. The ego development levels of the analogue clients were fixed at I - 3, I - 3/4, I - 4 or I - 4/5 as it would be expected that a large percentage of persons seeking counseling would be at one of these levels. Holt (1980) showed approximately 852 of young adults with some college experience at one of these 56 57 four levels. This limits the study's generalizability to these specific levels of clients. The measures used to assess ego development and expressed empathy, the WUSCT and the RES, are measures which have been used with actual counseling clients. It is expected that this will be an advantage in comparing the present results with future research in a more naturalistic setting. Generalizability and the analgpe; This study has utilized analogue clients in the form of typewritten descriptions and quotations instead of taped, role-played, or actual clients. The major advantage of the analogue was that it allowed experimental control not usually found in a naturalistic setting. This strength was utilized in order to investigate an interaction requiring repeated matches and mismatches of counselor and client characteristics. The use of the analogue method also poses the major restraint upon the generalizability of this study's results. In effect, the situation is one in which very high internal validity was obtained at the expense of lowered external validity. Both Strong (1971) and Munley (1974) point to this as the major disadvantage Of the analogue method. In terms of Strong's (1971) boundary conditions the analogue used involved a constricted conversation of short duration in which the participants were of unequal status which differs significantly from actual counseling practice. Counseling practice involves unconstricted conversation where both participants are live and free to interact, and both the contact and the status differences would take place over a more extended and Open ended period of time. The analogue in this study does not meet Strong's (1971) boundary conditions and thus is not directly generalizable to a counseling situation. There is evidence, reviewed earlier, that a written analogue is comparable to a taped analogue. However, extensions of this study into a practice setting will be 58 necessary to establish fully the application of its findings. The nature of the analogue task. There are additional limitations inherent in the particular analogue used. This analogue correSponds but is not equivalent to an initial counseling contact with a female client who presents a personal/social problem. It does not correSpond to later phases Of counseling or to counseling for other presenting problems. It is not expected that the results would have been different for a male analogue client. However, the sex of the analogue was fixed and thus the results do not generalize specifically to male clients. Extensions of this study should focus upon initial contacts with persons presenting personal/social presenting problems to maintain comparability. Review of Results and Discussion The Main Effect for Subject Level The finding. The hypothesis that the expressed empathy of the subjects would relate positively and significantly with the their level of ego deve10pment was not supported. The group means of the RES scores were in the expected order but this did not reach significance. There are two possible explanations for this lack of a significant effect. It may be that either the study's conditions were not powerful enough to elicit this effect or that it is unrealistic to eXpect a simple positive relationship to exist across experimental conditions. Discussion and implications. Support for the lack of power as the cause of the lack of significance is found in the actual group means assuming the expected order. That is, subjects at I - 4/5 had higher RES scores than did subjects at I- 4; subjects' at I - 4 had higher RES scores that subjects at I - 3/4. One might argue that simply increasing the p of the sample would elicit this effect. However, the main effect for subjects did not reach significance at the p < .05 level and the means may logically be explained as a chance occurrence. Furthermore, the main effect for analogues and the interaction did 59 reach significance with the same p, and SO the explanation concerning lack of statistical power is weakened. A different reason for the absence of a significant main effect for subjects may provide a better explanation. Loevinger (1976), Hauser (1976) and Swensen (1980) argue that it is not expected that a construct as complex as ego development would covary positively with any set of behaviors. Instead, curvilinear relationships and interactions are expected. Using this reasoning, the absence of the main effect for subjects is expected. This reasoning is reinforced by the findings regarding the interaction. First, a more complex relationship, an interaction, was found instead of the simpler main effect. Secondly, a post hoc analysis found that when the analogue level was held constant at I - 4, I - 4/ 5 subjects had significantly higher RES scores than I - 3/4 subjects, and I - 4 subjects' scores appeared in an intermediate position. This demonstrates a main effect for subjects evident only when the variable which the subject encountered was Specified and held constant. The four RES score means across analogues for each subject group assumed curvilinear patterns. When these three curvilinear relationships were analyzed together, the main effect was eliminated whereas the interaction was evident. This result may eXplain why Zielinski (1973, cited in Loevinger, 1979) and Carlozzi, et a1. (1983) found positive relationships between WUSCT levels and measures of empathy. In both studies the ego deve10pment level of the stimulus materials was not established or systematically varied. One may presume that the ego deve10pment level of those materials was constant if it was evident at all. In the present study it is apparent that the introduction and systematic variation of an important client characteristic, ego development, resulted in the absence of a' main effect for subjects. The cautions of Loevinger (1976), Hauser (1976) and Swensen (1980) are 60 supported. It is advisable that future research investigating the relationship between ego deve10pment and the empathy cycle either a) focus upon interactions or curvilinear relationships, or b) hold constant and report the ego development level of the experimental condition when investigating a main effect. The Main Effect for Analpgue Level The finding. The subjects' RES scores did vary significantly among the analogue levels. The post hoc analysis showed that the chief contributor to this effect was that RES scores in response to the I - 4 analogue were significantly higher than the RES scores in reSponse to the I - 3 analogue. The RES scores for both the I - 3/4 and the I - 4/5 analogues were in an intermediate position. This finding shows both that the analogue levels were a significant variable, and that the post-conformist subjects had significantly higher RES scores with the post-conformist conscientious client analogue than with the conformist client analogue. Discussion and implications. This finding is a validation of the use of the client analogues as the main effect shows that significant variance resulted fiom their use. Because the length of the analogues, the sex, and the topics discussed were held constant, and the second and third statements which the subjects responded to were randomly sorted across analogues, it is reasonable to conclude that it was the ego development levels of these analogues which created this variance. This finding may also be taken as an indication that clients in an initial counseling contact may be afforded varying amounts of expressed empathy by counselors in relation to the ego deve10pment level of the clients. This may be eSpecially true in the case of conformist clients being shown less expressed empathy. 61 This relation between client ego development level and counselor expressed empathy will need to be replicated in a naturalistic setting so as to ascertain if it exists in such settings. To determine this, further research in a counseling setting could be undertaken in which the relationship between the ego development levels of clients and the expressed empathy of counselors in initial contacts is investigated. Research might be unobtrusively conducted at agencies where tape recording of sessions is often used and where written information from clients is routinely obtained. The Interaction Between Subject and Analogue Levels of ELO DeveIOpment The finding. The significant interaction factor in the ANOVA showed that the RES scores were determined by both subject and analogue levels Of ego deve10pment. The post hoc analysis revealed that two additional statements may be made. A) I- 3/4 subjects had significantly lower RES scores with analogues lower or more than one step higher than I - 3/4 (I - 3 and I - 4/5 analogues) when compared to their RES scores with analogues at or one Step higher that I - 3/4 (I - 3/4 and I - 4). B) Subjects at I - 4/5 had significantly higher RES scores with analogue I - 4/5 than did I - 3/4 subjects. I - 4 subjects' RES scores were at an intermediate position. This result was discussed earlier in relation to the main effect for subjects. Discussion and implications. This finding confirms the hypothesis of chief interest in this study and as such this study can be considered a success. Each group of subjects showed the greatest expressed empathy as measured by RES scores with analogue clients of equal ego development level. The level Of , eXpressed empathy for every subject group declined with analogue clients higher or lower than the subjects' own ego development level. This finding suggests that counselors will demonstrate considerable less empathy when dealing with clients developmentally different from themselves. 62 An examination of the means in Figure 4.3 also suggests that this effect of decreased expressed empathy may be more pronounced when the counselor is dealing with a client whose ego development is more than one level different from the counselor's own, at least for counselors at I - 3/4 and I - 4/5. Research confirming this interaction in a naturalistic setting may be difficult because of the number of matches and mismatches of levels required. However, the results from this study indicate that this sort of research may be worthwhile. An alternate to a design including the range of levels examined here, may be a study to investigate and compare specific matches and mismatches rather than the entire range possible. For instance, the post hoc analysis of the interaction Showed that I - 3/4 subjects had significantly higher expressed empathy with analogues at the same developmental level and one level higher than with other levels. Trends in Counselor Preferences The findings. The fourth hypothesis was confirmed as subjects indicated a higher preference for client analogues at their own level or at one level higher compared to client analogues at a lower level or more than one level higher than the counselor's own. An additional analysis revealed that the overall trend in the subjects' preferences was that the conformist client analogue was less preferred by all the subject groups. All subjects were post-conformist. The significance of this second finding (p < .001) when compared to that of the first analysis (p < .05) suggests that this incidental result may overshadow the formal test of hypothesis four in importance. Discussion and implications. There are three studies which Specifically . relate to the present study's investigation of preference. Rest, et a1. (1969) found that subjects preferred the moral reasoning at the highest level which they could comprehend. This level was often one level higher that the subject's 63 own. This is the primary source for the exact nature of hypothesis four. Therefore, this study has confirmed this asymmetry of preference using ego development theory. Mullen and Abeles (1971) investigated the relationship between empathy, liking and counseling outcome. Whereas they found no relationship between liking and outcome, it was concluded that this might have been because the measure of liking, the Nonpossesive Warmth Scale, tended to measure an overall warmth of the therapist rather than a therapist's particular liking for a specific client. The present study obtained a measure of the subjects' Specific preference for specific client analogues which may have aided in obtaining a Significant finding. An additional difference between the two studies is that the present one does not relate preference to outcome. Young-Eisendrath (1981, cited in Young-Eisendrath, 1982) obtained the result most similiar to this finding regarding preference. Young-Eisendrath (1981, cited in Young-Eisendrath, 1982) reported an incidental finding that the subjects, also Masters level counseling students, preferred the post-conscientious client analogue to the conformist client analogue. Young-Eisendrath reported that this difference in preference affected the subjects' ability to infer the client analogue's frame of reference. The present study confirms Young-Eisendrath's (1981, cited in Young-Eisendrath, 1982) finding. The subjects, post-conformist Masters level counseling students, demonstrated a significantly lower preference for the conformist client analogue relative to the post-conformist client analogues. In addition, this study showed that subject preference was a function of both subject and analogue levels Of ego development. These findings establish a relationship between ego deve10pment and counselor preferences for clients based upon the counselor's initial knowledge of 64 the client. These findings are more directly applicable to counseling practice than the findings involving expressed empathy. This generalizability is based upon the nature of the task by which the subjects' preferences were established. Subjects read descriptions of four clients and then rank ordered their preferences for these clients. The subjects were asked to rank their preferences as if they were counselors with schedules which did not allow them to see all the clients. This task directly resembles client selection in many organizations in which counselors are allowed to choose their caseload on the basis Of the contents of an intake report. Therefore, these findings regarding counselor preferences and its generalizability to client selection raises an issue previously unestablished. If post-conformist counselor subjects in this study preferred post-conformist clients significantly more than the conformist client, it is reasonable to expect that this difference in preference leading to differential selection of clients would occur in an organization with a similiar client selection procedure, with a similiar mix of client developmental levels, and with counselors of similiar experience and ego deve10pment level. Investigations confirming this effect in an organization providing counseling could be done using counselor WUSCT ratings, client selection data normally kept by the organization and independent ratings of intake reports for client ego deve10pment level. Recommendations The application of aptitude x treatment interaction (ATI) findingp. This study most resembles an ATI approach in its research design. The subjects' WUSCT levels are the aptitude, the four analogues are the treatments, and the dependent variable, RES scores, provides evidence of interaction. Therefore, the views of Cronbach (1957, 1975) are helpful to consider before seeking general implications and recommendations of this Study. 65 Cronbach (1975) stated that ATI findings have two general applications. The first is the use of the knowledge of the empirical relationships in new situations. Cronbach (1957) stated that ATI results can be used empirically in order to achieve maximally beneficial matchups between persons and treatments. Cronbach (1975) moderated this recommendation. He pointed out that an empirical relationship established in one situation between an aptitude and treatment must be re-established in new situations before being used reliably for matching. ATI results are not automatically generalizable. There is a second use for ATI results. "The other reasonable aspiration is to develop explanatory concepts, concepts that will help people use their heads" (Cronbach, 1975, p.126). Recommendations for research. (A) Research is needed to confirm the existence of this study's findings in a counseling practice setting. This is necessary because of the artificial nature of the analogue task. (B) Research is needed to extend the present findings to a greater range of client ego development levels, counselor ego deve10pment levels and counselor experience. (C) Research is needed to ascertain the impact of the interaction of counselor and client ego development levels upon client perceived empathy and upon counseling outcome. (D) Research investigating components of the empathy cycle Should hold constant and report the ego development levels of counselor and clients or include these levels as a variable. Recommendations for practicye, traininj and supervision. If the generalizability of the present findings were established, the use of this empirical relationship fOr optimal matching of counselors and clients might be undertaken. In each setting, however, evidence would be needed as to the presence of the empirical relationship. The waiting lists in organizations providing counseling services might be examined by their Staffs to see if a 66 disproportinately high number of conformist clients were waiting for services due to differential selection. Counselor training programs using empathy training might systematically include and vary the ego development levels of the materials used in this training. Counseling students could be made aware of ego development theory and its possible impact upon counseling process in a systematic fashion. This might help the counseling students to become more aware of ego deve10pment levels as a source of counter-transference and the misunderstanding of clients when and if this occurs. Young-Eisendrath (1982) has begun such a program within a social work school. Ego deve10pment levels Of client and counselor might be a fruitful issue to consider in the supervision of counselors. Implications of the interaction between counselor and client levels which this study uncovered can be usefully looked at from two additional viewpoints. These are the viewpoints Of high-level and low-level counselors. Counselor training including the ItOpic of ego development could help to raise the consciousness of high-level counselors to the possibility of an important counter-transference issue. The present study showed that counselors demonstrated lowered expressed empathy and lowered preference toward clients whose level was below the counselors'. In practice, a Conscientious or Individualistic level counselor dealing with a Conformist level client might dislike this person's viewpoint and be less empathic and less helpful as a result. This same counselor, dealing with a Self-Protective level client might resist aiding this person to attain a Conformist level of deve10pment (the next stage) even when this might be an apprOpriate treatment goal. A counselor who had previously been alerted to this possibility and to his or her personal reaction to low level clients might consider a) referral, b) supervision, c) practice in viewing the client's world through this more primitive 67 viewpoint in order to understand better the client's outlook. The other side of this issue is the case of a low level counselor. A counselor at the Conscientious/Conformist Transitional level may have Significant difficulty comprehending the frame of reference of a high level client. Thherefore, referral, supervision and participation in life experiences designed to enhance the counselor's personal deve10pment are options here. This same counselor, however, would not experience this difficulty comprehending clients at the same level or below, and may have a higher preference for these clients than a high level counselor. This could be advantageous. The use of peer-counselors from the same group as a client population (e.g. high school students) would be a case in which the counseling outcome could possibly be improved through the use of peer counselors or paraprofessionals. This would be because peer counselors, for example, may prefer and be very fluent in the frame of reference used by the bulk of the client population compared to trained and possibly older and more deve10ped professionals. APPENDICES APPENDIX A Washigton University Sentence Completion Test Sentence Completion For Women (Form 11-68) Code #-—--— Date Instructions: Complete the following sentences. 1. Raising a family 2. A girl has a right to 3. When they avoided me 4. If my mother 5. Being with other people 6. The thing I like about myself is 7. My mother and I 8. What gets me into trouble is 9. Education 10. When people are helpless 11. Women are lucky because 12. My father 13. A pregnant woman 14. When my mother spanked me, I 15. A wife Should 16. I feel sorry 17. Rules are 18. When I get mad 19. When a child will not join in group activities 20. Men are lucky because 21. When they talked about sex, I 22. At times she worried about 23. I am 24. A woman feels good when 25. My main problem is 26. My husband and I will 27. The worst thing about being a woman 28. A good mother 29. Sometimes she wished that 30. When I am with a man 31. When she thought of her mother, she 32. If I can't get what I want 33. Usually she felt that sex 34. For a woman a career is 35. My conscience bothers me if 36. A woman should always 68 69 Sentence Completion For Men (Form 11-68) Code #-—-—- Date Instructions: Complete the following sentences. 1. 2. 3. Raising a family When a child will not join in group activities When they avoided me 4. A man's job 5. Being with other peOple 6. The thing I like about myself is 7. If my mother 8. Crime and delinquency could be halted if 9. When I am with a woman 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. Education When people are helpless Women are lucky because What gets me into trouble is A good father A man feels good when A wife Should I feel sorry A man should always Rules are When they talked about sex, I Men are lucky because My father and I When his wife asked him to help with the housework Usually he felt that sex At times he worried about If I can't get what I want My main problem is When I am criticized Sometimes he wished that A husband has a right to When he thought of his mother, he The worst thing about being a man If I had more money I just can't stand people who My conscience bothers me if He felt proud that he 70 Examples of WUSCT sentence completions scored at DELTA, I - 3, and I - 4 levels of ego development (Hauser, 1976). DELTA, Self- Protective Stage: When peOple are helpless....."I don't like to be bothered with them." When they talked about sex, L...."listep, 'cause I won't 83': in trouble, also I'll know what I'll be doing; I - 3, Conformist Stage: When peOple are helpless....."I try to help them." When they talked about sex, I....."joined riLht in." I - 4, Conscientious Stage When people are helpless....."I sympathize, unless they are unwflfinLto help themselves." When they talked about sex, I....."was interested and respected what they said." APPENDIX B Client Analogues The following analogues were developed for the present study by the author. They were constructed so as to simulate a initial contact with a client at a varying level of ego deve10pment. Each client analogue discussed, in typewritten quotations, her presenting problem (a troubling interpersonal relationship, one other relationship, her attitudes toward school at the university, her views about career and about herself. All information included was varied according to the ego deve10pment stage descriptions of Loevinger (1976). After construction, these analogues were independently rated by two trained raters, in random order. The ratings were in 100 Z agreement with each other and with the intended stages. The ego development levels of the following analogues are: Ann B. I - 3, Conformist Stage Maria R. I - 3/4, Conscientious/Conformist Transition Susan G. I - 4, Conscientious Stage Julie K. I - 4/ 5, Individualistic Level The instructions were: My self selected code number is Instructions For Part II I would like you to read descriptions of four counseling clients. After each are 3 client statements followed by spaces marked "Counselor Response." Imagine that you are the counselor and respond, in writing, in a way you feel would be helpful. Also, try to respond to each client based upon everything you have learned about her, and not just on the basis of a single statement. After the subjects read the instructions and completed the analogues (in random order), they were asked to indicate their preferences for the client analogues in the following way: Assume that you are a counselor with a busy schedule and that you only have time to see 2313 Of these clients for counseling. Please rank order these clients based upon whom you would most like to see, not necessarily who needs counseling the most, but rather whom you'd prefer to counsel with your last available time. Please Rank Order Most Preferred 2nd Most Preferred 3rd Most Preferred Least Preferred (Client's names) Thank you. 71 72 Client: Ann B. Presenting Problem "...feeling sad about breaking up with my boyfriend." Since the breakup Ann has been quite upset and has had low motivation. Notes: Recently Ann broke up with her boyfriend. This was a "long story", she said, which she felt "...really sad" about. Ann Stated that last summer, when he wasn't working or in school, her boyfriend started wearing his hair differently. "...It looked so weird, but he didn't seem to care. I mean, no one I've ever known has had hair. like that—and none of the guys who date my girlfriends, either. I just didn't feel its right! Finally her boyfriend broke off the relationship. Ann said she feels lousy about this and that now on weekends she's the only one of her girlfriends without a date. She feels ashamed at these times. A girlfiend of Ann's had suggested that she speak with a counselor and that this would make her feel better. She and this woman are quite close. Ann said "...we're really best friends. I mean, we always see things the same way and everything. And when school or the classes we're in are confusing we can talk together and figure it out." Ann is not quite sure about her career direction, but she says that she knows she'll find the right one with a bit more time. ‘ Ann also said that she is the kind of person who "...likes to be part of something, my family, my group of friends, and I like to be friendly and nice with all my friends. I'm really a social type of person, you know, I always try to look good whatever I'm doing. That's why its been so hard breaking up with my boyfriend." Ann said that our interview was like some of her courses at the university, where "...the prof's ask different questions and get you to look at different areas, almost like they're uncertain, until you, I mean we find the solution. Its like the saying my best friend has 'you have to go down a rough road before you find the answer.‘ Do you know what I mean?" (C ounselor R eSponse) "The way I'm feeling is really bothering the hell out of me!" (C ounselor R eSponse) "Well, that's the situation, what do you think?" (C ounselor R eSponse) 73 Client: M aria R. PresentipiProblem She's been feeling generally uncomfortable and self-conscious around her fi'iends lately, especially about her opinion relative to theirs. Sometimes this leads to feelings of panic. Notes During our interview Maria seemed always concerned about what she would say. It seemed that she would become uncertain after she began speaking, almost embarrassed about what she was saying or the way in which she was saying it. Maria's uncomfortable feelings have to do with herself in relation to her friends, although she's vague about this. Since the time when she entered the university, she has become aware that she didn't always follow what she used to think were the "right" rules of conduct in school and around her friends. For instance "...I was taking a course that was pretty easy but very boring, and I just stopped doing a lot of the work. I figured that I'd get by and I did. It was different, though, not applying myself to school like I used to. And in other courses, I've started seeing that there's a lot of different ways to look at things, which is different, too. It's like the way I grew up isn't the only way." "When I came here I thought that I'd find the right guy and marry, and find a career that fit, but now...I just don't know what to think, there's so many different ways." "And then this guy I'm seeing wants me to make up my mind about marriage and everything, but I just feel confused. I guess it's alright for him right now, 'cause he's older. He really gets frustrated about this, but I just end up feeling lonely (she became tearful at this point). I'm still seeing him, even though I don't intend to marry him...why, even that's a change!" She looked fearful after she said this. I've talked about this with my friend, Jill, but we just end up with the same confusion. It's good to have her to talk with, though. We've always had a rule: if one Of us needs to talk out something, the other listens, even if she's involved in something else. I guess that seems like a really feminine thing to do." Right now, you could say, I'm the kind of person who's trying to sort out which way to go and how I feel about it. Does it seem that way to you, too?" (Counselor Response) "This feeling I have...I just can't stand it!" (Counselor ReSponse) "I keep wondering...what do you think of all this?" (Counselor ReSponse) 74 Client: Susan G. Presenting Problem She's depressed and feels responsible because a close friend had a "nervous breakdown" recently. Since then she's been eating much less and has had a great deal of trouble getting to sleep at night. Notes Susan took an active part in our interview and tried to tell me the things she felt were important to know about her and her situation. She was quite tearful throughout our meeting. She started by telling me about who she was as a person. "I've set up my own standards in my life. I still try to take others into account, of course, even if they have a different point of view. But, overall, I use my own perspective. This helps, like in coursework here at the university. A lot of subjects I study don't give definite answers, because that's how the world is. Even then, I've found school useful. Now I'm able to view situations in a relative sort of way and then I decide how to act. My boyfriend is a lot like this too—that's one thing that attracts us to each other—that we share a similiar way of ‘ understanding the world. We both have some long term goals for our careers and lives and, for now, we find we can support each other in our different goals." "Well, one of the standards I have is to help Others if I'm able. This may sound crazy, but I sort of decided that I'd always try to do this. It was like that with my roommate. She was real depressed and having a hard time of it. I found that I was able to help her see a different perspective about things. That felt really satisfying to me, real good, being able to help." "But then things in her life got tougher—its complicated because some things for her go back to her family and all. She dropped out and I didn't hear from her. Then last week I found out that she was in the hospital. (Susan began sobbing at this point and looked away.) She had a nervous breakdown. I felt horrible! I just keep saying to myself that somehow I could have reached out after she left school. Maybe that could have changed things. I feel so responsible. I don't know what to do." (C ounselor R esponse) "My feelings about this just keep eating away at me!" (C ounselor R esponse) "...uhm, how do you think I should approach this?" (C ounselor R eSponse) 75 Client: Julie K . PresentingyProblem She's been involved in a relationship which is conflictual for her. She decided to seek counseling in order to more fully understand her feelings and the situation before acting. Notes Although she was visibly distressed, Julie seemed to be actively struggling with the issues she spoke about. She also seemed to carefully consider and respond to the interviewer's comments. Julie said, "I wish my life were as straightforward as some of my courses seem. You know, applying different principles to situations depending upon the circumstances, like in physics. But its not. Julie sounded like she had deve10ped a number of strong relationships with those around her. But one was particularly troubling. In the past year Julie's boyfriend has apparently been experiencing a personal career upheaval. She said, "I could tell that he was going through some changes because his appearance had changed. There were some feelings inside of him which conflicted. That was OK in a way, because I've experienced that sort of conflict, too. So I've tried to understand what he's going through and support him, but I've only got so much energy. You see, there's some things that I've committed myself to achieving as an individual. But I'm just not sure that I'll be able to work toward these goals I've set, and be there for him as much as he needs..." She seemed to have some really complex feelings about this. "It's not so much a practical matter, since he supports himself and has his own apartment, but more an emotional one. That's partly why it's so frustrating to me. Sure, things might be easier if, say, my academic program at school were less pressured or something, but this is still a conflict I feel inside me. I want a deep relationship, one in which the two of us can develop a real closeness yet develOp ourselves as individuals, too. "But develOping myself as an individual takes a lot of determination and endurance. I know, because its a commitment I made to myself some time back, to live life as fully as I can while keeping to my own values. Still, it's really been feeling like a bind, since I'm not sure if I can manage both areas as one person. And that hurts, thinking that I might have to give up one. Its a paradox to me; as Ilearn more about life, it only seems to get more complex. Does all of this make sense?" (Counselor ReSponse) "I hate feeling this way...I hate it!" (Counselor ReSponse) "I figured it would be helpful to find out your reaction." (Counselor Response) APPENDIX C Revised Response Empathy Scale Elliott, R., Reimschuessel, C., Filipovich, H., Zapadka, J., Harrigan, L., and Gaynor, J. (1981). Department of Psychology, University of Toledo, OH 43606 Subscales L 2J 3, 4, 54 7. Score each helper response on each of the following six dimensions. Leave E blank items. 1. Client Feeling: To what extent does the helper address the client's feelings? 4 - Refers to specific client feelings (e.g., hurt, depressed, angry, anxious). 3 - Reference is to generalized or vague client feelings (e.g., bad, upset, disturbed, okay) or to motivations (e.g., wants or desires). 2 - Reference is to client thoughts, cognitions or Opinions (e.g., "Do you think that was the right way to handle it?") 1 - Reference is to client behaviors or actions (e.g., "Why did you do that?") 0 - Reference is to other people's feelings, cognitions and behaviors (e.g., including therapist's, significant others', etc.). Note: For long responses, identify all ratable aSpects of the response and average; use judgement for rounding up or down. 2. Percgptual Inference 8 Clarification. Does the helper make inferences to tell the client something the client hasn't said yet to add to the client's frame of reference or to bring out other implications? 4 - Significant inference or clarification present (e.g., interpretations). 3 - Moderate inference or clarification present (e.g., reflections of implications, significant summaries). 2 - Slight inference or clarification present (e.g., paraphrase or nonverbal reflection). 0 - No inference or clarification present (e.g., echoic reflections, information questions, advisement, guesses about factual Situations, etc.). 3. Centrality of TOpic: Does the helper refer to what is important to the client? Does the helper's response relate to the helpseeker's basic complaint or problem? This dimension is to be rated on the basis of content. 76 77 (3. Centrality of Topic, continued) 4 - Definite, clear reference to client's feelings and perceptions connected with the client's major complaints; refers to conflicts, central issues, basic relationships. 3 - Topic moving toward major concern. 2 - TOpic not clearly central; background facts, possibly related to major concerns. 1 - Points of clarification of meaning; repairing mishearings, misunderstandings. 0 - Topic definitely pp_t central (e.g., "side tracks," talk about the helping process; third-party talk, general talk, "mhm.") 4. Expressiveness: Does the helper communicate in an appropriately expressive manner, using language and voice, including: a) language which evokes imagery in the rater, b) creative, novel language (words and/or syntax), c) modifiers (adjectives/adverbs) which intensify or upgrade (vs. attenuate) meaning of message, d) interjections (!), e.g., Wow! Gee!, e) inflected voice (i.e., with variations in pitch and/or volume). f) voice which communicates a specific emotion, e.g., sadness, excitement, prizing, caring, surprise, anger. Except - where language or voice distracts or evokes confusion in the rater. 4 - Greatly expressive; multiple or strong cues. 3 - Moderately expressive; singular or less-strong cues. 2 - Ordinary language or voice. 1 - Mildly distracting or confusing; uhuh; echoic reflections, repetitions. 0 - Moderately or extremely distracting or confusing; moderate or exu'eme repetition. 5. Collaboration: Does the helper's manner communicate a sense of working together in a collaborative process? This item is to be rated on the basis of intention; do not use impact to rate this. 4 - Clearly communicates a sense of collaboration, mutuality, a "working together" or "we're in this together attitude"; personalizing reSponseS that communicate a sense of sharing and give support and strength to the client (e.g., collaborative responses, completers); possible references to the T-C relationship. 3 - Perception check in self-disclosure format; personal, peer informal helping equalizer, friendship (e.g., "reallyl", "Let's talk"); reference to shared history, themes. 78 2 - Possibly communicates collaboration (e.g., sympathizers - "That's hard", "mhm", laughter intended to match client, ordinary reflections). 0 - Fails to communicate a sense of collaboration (e.g., session ending, disagreement, unsolicited advice, cutting client off). 7. Exploration: Does the helper actively encourage the client's exploration by the content of what the helper says? DO not use impact to rate this. 4 - Clear, active push to encourage client to explore (e.g., exploratory questions, "How do you feel about that?", process advisement in the form of "tell me" questions, tag questions at the end of significant summary reflections and interpretations. 3 - Exploratory questions at beginning of session to gather information. 2 - Possibly encourages client to explore; passive facilitation of exploration (e.g., pausing; tentativeness; ordinary content reflections; using self-disclosure or question format or frame for something else; "mhm".) 0 - Discourages exploration (e.g., closed questions or questions seeking specific information; "topic jumping;" session ending). LIST OF REFERENCES LIST OF REFERENCES Allen, M. J., 8 Yen, W. M. (1979). Introduction to measurement theory. Monterey, CA: Brooks/C ole. Barrett-Lennard, G. T. (1981). The empathy cycle: Refinement of a nuclear concept. Journal of CounselinyPsychology, 28, 91-100. Bergin, A. E., 8 Soloman, B. (1970). Personality and performance correlates of empathic understanding in psychotherapy. In T. Tomlinson 8 J. Hart (Eds.), New directions in client-centered therapy. Boston: Houghton-Mifflin. Blasi, A. (1976). Personal responsibility and ego development. In R. deCharms (Ed.), Enhancinimotivation: Change in the classroom. New York: Irvington. Browning, D. (1978). Preliminary manual for WUSCT items. Unpublished manuscript, New York University. Candee, D. (1974). Ego development aspects of new left ideology. Journal of Personality and Social Psychology, 32, 620-630. Carkhuff, R. R. (1969). Helping and human relations(Vol. 1). New York: Holt, Rinehart and Winston. Carlozzi, A. E., Gaa, J. P., 8 Liberman, D. B. (1983). Empathy and ego development. Journal of CounseliniPsycholojgy, 3_0, 113-116. Chinsky, J. M., 8 Rappaport, J. (1970). Brief critique of the meaning and reliability of "accurate empathy" ratings. Psycholoiical Bulletin, 13, 379-382. Cohen, J., 8 Cohen, P. (1975). Applied multiple regryession/correlation analysis for the behavioral sciences. Hillsdale, NJ: Lawrence Erlbaum. Cox, L. N. (1974). Prior help, ego development and helping behavior. Child Development, 42, 594-603. Cronbach, L. J. (1957). The two disciplines of scientific psychology. American Psychologist, 2, 671-384 Cronbach, L. J. (1975). Beyond the two disciplines of scientific psychology. American Psychologist, _3_Q, 116-127. Elliott, R., Filipovich, H., Harrigan, L., Gaynor, J., Reimschuessel, C., 8 Zapadka, J. (1982). Measuring response empathy: The development of a multicomponent rating scale. Journal of CounseliniPsychology, 2_9_, 379-387. 79 80 Elliott, R., Reimschuessel, C., Filipovich, H., Zapadka, J., Harigan, L., 8 Gaynor, J. (1981). Revised Response Empathy Rating Scale. Unpublished manuscript, Department of Psychology, University of Toledo, Toledo, OH. Elliott, R., Reimschussel, C., Filipovich, H., Zapadka, J., Harrigan, L., 8 Gaynor, J. (1982). Validation of the Revised response empathy scale: Comparison with client and counselor perceptions. Unpublished manuscript, University of Toledo, Toledo, OH. English, R. W., 8 Jelenevsky, S. (1971). Counselor behavior as judged under audio, visual, and audiovisual communication conditions. Journal of Counselinj Psycholm, L8, 509-513. Feldstein, J. C., 8 Gladstein, G. A. (1980). A comparison of the construct validities of four measures of empathy. Measurement and Evaluation in Guidance, fl, 49-57. Fridman, M. S., 8 Stone, 8. C. (1978). Effect of training, stimulus context, and mode of stimulus presentation on empathy ratings. Journal of Counseling Psychology, 22, 131-136. Gladstein, G. A. (1977). Empathy and counseling outcome: An empirical and conceptual review. Counseling thologist, _6_, 70-79. Gladstein, G. A. (1983). Understanding empathy: Integrating counseling, deve10pmental, and social psychology perspectives. Journal of Counselinj Psychology, fl, 467-482. Goldberg, A. D. (1974). Conceptual system as predisposition toward therapeutic communication. Journal of Counseling—Psychology, _2_1, 364-368. Gormally, J., 8 Hill, M. (1974). Guidelines for research on Carkhuff's training model. Journal of Counselifi Psycholcgy, 2;, 539-547. Grant, J. D., 8 Grant, M. Q. (1959). A group dynamics approach to the treatment of non-conformists in the Navy. Annals of the American Academy of Political and Social Science, 322, 126-135. Greenson, R. R. (1967). The technique and practice of psychoanalLsisWol. 1). New York: International Universities Press. Greenson, R. R. (1978). Explorations'in psychoanalysis. New York: International Universities Press. Gurman, A. S. (1973). Instability of therapeutic conditions in psychotherapy. Journal of Counseliri Psychology, A), 16-24. Hardin, S. L, 8 Yanico, B. J. (1981). A comparison of modes of presentation in vicarious participation counseling analogues. Journal of Counseling Psychology, a, 540-543. Harvey, O. J., Hunt, D., 8 Schroeder, H. (1961). Conceptual systems and personality organization. New York: Wiley. 81 Hauser, S. T. (1976). Loevinger's model and measure of ego deve10pment: A critical review. Psycholcgtical Bulletin, _8_3_, 928-955. Heck, E. J., 8 Davis, C. S. ( 1973). Differential expression of empathy in a counseling analogue. Journal of Counseling Psycholggy, 20, 101-104. Hicks, S. W., Cormier, W. H., 8 Bounds, W. G. (1974). Readirgstatistics and research. New York: Harper 8 Row. Holt, R. R. (1980). Loevinger's measure of ego development: Reliability and national norms for male and female short forms. Journal of Personality and Social Psychology, 32, 909-920. Hunt, D. E. (1977). Conceptual level theory and research as guides ~to educational practice. Interchange, 8(4), 78-90. Kagan, N. (1975). Interpersonal process recall: A method of influencinLhuman interaction. Mason, MI: Mason Media. Kagan, N., Krathwohl, J., 8 Associates (1967). Affective Sensitivity Scale. Mason, MI: Mason Media. Kimberlin, C. L., 8 Friesen, D. D. (1980). Sex and conceptual level empathic responses to affect. Counselor education and supervision, l_9_, 252-258. Kohlberg, L. (1969). Stage and sequence: The cognitive-developmental approach to socialization. In D. A. Goslin (Ed.), Handbook of socialization theory and research. Chicago: Rand McNally. Kurtz, R. R., 8 Grummon, D. L. (1972). Different approaches to the measurement of therapist empathy and their relationship to therapy outcomes. Journal of Consulting and Clinical Psychology, 3_9, 109-115. Loevinger, J. (1966). The meaning and measurement of ego develOpment. American Psychologist, a, 195-206. Loevinger, J. (1976). Ego deve10pment: Conceptions and theories. San Francisco: Jossey-Bass. Loevinger, J. (1979). Construct validity of the sentence completion test of ego development. Applied Psychological Measurement, 3, 281-311. Loevinger, J. (1980). Some thoughts on ego development and counseling. Personnel and Guidance Journal, 28, 389-390. Loevinger, J., Wessler, R., 8 Redmore, C. (1970). Measu'pg egp develgpment(Vols. 1-2). San Francisco: Jossey-Bass. Lutwak, N., 8 Hennessey, J. J. (1982). Conceptual systems functioning as a mediating factor in the development of counseling skills. Journal Of Counseling Psychology, a, 256-260. Melnick, R. R. (1975). Counseling response as a function of method of problem presentation and type of problem. Journal of Counselifl Psychology, g2, 108-112. 82 Mullen, J., 8 Abeles, N. (1971). Relationship of liking, empathy, and therapist's ‘experience to outcome of therapy. Journal of Counseling Psycholm, l8, 39-43. Munley, P. H. (1974). A review of counseling analogue research methods. Journal of CounseliflPsychology, 2_1, 320-330. Myers, J. L. (1979). Fundamentals of experimental desigp(3rd ed.). Boston: Allen 8 Bacon. Redmore, C. (1976). Susceptibility to faking of a sentence completion test of ego development. Journal of Personality Assessment, 39, 607-616. Redmore, C., Loevinger, J., 8 Tamashiro, R. T. (1978). Measurig ego deve10pment: Scoriananual for men and boys. Unpublished manuscript, Washington University, St. Louis, MO. Rest, J. R. (1973). The heirarchical, nature of stages of moral development. Journal of Personality, fl, 86-109. Rest, J. R. (1980). Development in judgipgymoral issues. Minneapolis: University of Minnesota Press. Rest, J. R., Turiel, E., 8 Kohlberg, L. (1969). Level of moral deve10pment as a determinant of preference and comprehension of moral judgments made by others. Journal of Personality, 31, 225-252. Rogers, C. R. (1957). The necessary and sufficient conditions of therapeutic personality change. Journal of Consulting Psychology, 2i, 95-103. Rogers, C. R. (1959). A theory of therapy, personality and interpersonal relationships as developed in the client-centered framework. In S. Koch (Ed.), Psycholpgy: A study of a science: formulations of the person and the social context(Vol.3). New York: McGraw-Hill. Rogers, C. R. (1975). Empathic: An unappreciated way of being. Counseling Psychologist, 5(2), 2-10. Schrout, P. E., 8 Fleiss, J. L. (1979). Intraclass correlations: Uses in assessing rater reliability. Psychological Bulletin, 8_6_, 420-428. Stokes, J. P., 8 Tait, R. C. (1978). Comparability of written and oral measures in evaluating counselor training. Counselor Education and Supervision, 11, 222-229. Strong, S. R. (1971). Experimental laboratory research in counseling. Journal of Counseling Psychology, l_8_, 106-110. . Sutton, P. M., 8 Swensen, C. H. (1983). Reliability and concurrent validity of alternative methods for assessing ego development. Journal of Personality Assessment, 4_7, 468-476. Swensen, C. H. (1980). Ego deve10pment and a general model for counseling and psychotherapy. Personnel and Guidance Journal, 58, 382-388. 83 Therrien, M., 8 Fischer, J. (1978). Written indicators of empathy in human- relations fi'aining: A validational study. Counselor Education and Supervision, _1_7_, 272-27 7. Truax, C. B. (1966a). Influence of patient statements on judgements of therapist statements during psychotherapy. Journal of Counseling Psychology, a, 335-337. Truax, C. B. (1966b). Reinforcement and nonreinforcement in Rogerian psychotherapy. Journal of Abnormal Psychology, A, 1-9. Truax, C. B., 8 Carkhuff, R. R. (1967). Toward effective counseling and psychotherapy. Chicago: Aldine. Vesprani, G. J. (1969). Personality correlates of accurate empathy in a college companion program. Journal of Consultipg and Clinical Psychology, a, 722-727. Ward, J. H., 8 Jennings, E. (1973). Introduction to linear models. Englewood Cliffs, NJ: Prentice-Hall. Winer, B. J. (1962). Statistical principles in experimental desigp. New York: McGraw-Hill. Young-Eisendrath, P. (1981). Comparison of brief training in ego development and in empathy skills for beginning psychotherapists. (Doctoral dissertation, Washington University, 1980). Dissertation Abstracts International, 41B, 3159. Young-Eisendrath, P. (1982). Ego development: Inferring the client's frame of reference. Social Casework, {33, 323-332. Zielinski, C. E. (1973). Stages of ego development as a correlate of ability in discrimination and communication of empathic understanding. (Doctoral dissertation, University of Houston, 1973). Dissertation Abstracts International, 34A, 1635. ll! 1011111111101