The efficacy of supervision in training of psychotherapists is discussed in relation to (a) trainees' attitudes, beliefs, and skills, (b) trainee's performance in the therapist role, (c) interactional process events in supervision and psychotherapy, and, (d) client change. Although investigations to date suggest the potential of supervision for teaching advanced skills of psychotherapy, few studies exist that examine directly the relation of therapist performance and client change to supervision. There are virtually no studies that compare the efficacy of supervision to other training methods. If supervision is to remain an integral part of training, then standardized training manuals, analogous to those in psychotherapy, need to be developed.

Science and practice cannot continue together without a major attitudinal shiA a broadening perspective of science and practice and how these two activities can be integrated to strengthen each other. This article represents the culmination of a 2-year project that examined the roles of science and practice within counseling psychology. The central goal of the article is to present (a) specific recommendations for promoting the integration of science and practice within Division 17 and (b) suggest specific strategies for implementing the recommendations.

Developmental models of supervision have become the Zeitgeist of supervision thinking and research. Eighteen different models of supervision that refer to developmental principles have been cited in the psychiatric, psychological, and social work disciplines (Worthington, 1984). Numerous empirical studies have been designed to investigate developmental paradigms (Roehlke, 1984). If these models are to continue to influence research and practice, it seems warranted that the underlying assumptions and principles of the proposed models be critically examined. This article contains such a critical examination, and the conceptual and methodological approaches currently applied in the research of this topic are discussed from a developmental perspective. 209 This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

This study investigated the relationship between verbal patterns of behavior and supervisor and trainee judgments of satisfaction in the supervision interview. Blumberg's Interactional Analysis System was used to code the interviews and a sequential analysis methodology was employed to characterize interactional patterns. Three independent areas of satisfaction were assessed: (a) supervisor's (or trainee's) evaluation of other; (b) supervisor's (or trainee's) evaluation of self; and (c) supervisor's (or trainee's) level of comfort in the interview. Multiple regression was used to identify patterns of verbal behavior that predict satisfaction in each of the three cases. The findings reflect some generally held beliefs about dyadic .communication and challenge some of the basic assumptions concerning the use of supportive communication in the supervision interview.

The group supervision literature in counselor training from 1960 to 1983 was reviewed. Only those articles pertaining to supervised groups of counselor trainees participating in individual counseling practica were included. Descriptions of group supervision included elements of case presentation, didactic information, and interpersonal process. Because of the paucity of research in group supervision, practices are based on intuitive common sense approaches.

Thirty-seven counseling students completed the Paragraph Completion Method, which is used to determine conceptual level. All students then viewed a videotape of a client in a counseling session and completed the Clinical Assessment Questionnaire, which is designed to elicit information regarding counselors' hypotheses about a client's problem. The Clinical Assessment Questionnaire responses were scored on seven categories relevant to clinical judgment and information processing. Two factors were considered in a multiple regression analysis of the data: (a) conceptual level and (b) clinical experience. The results indicated that experience level was not significantly related to the characteristics of the students' clinical hypothesis formation. Conceptual level was significantly related to the overall quality and clarity of expression in clinical hypotheses (p < . 002) and the number of divergent questions asked about the client's problems (p < .004). Methodological and counselor-training implications of these results are discussed.

Conceptual systems theory (CST) has been used to predict the relation between a counselor's conceptual level (CL) and the counseling process. A meta-analysis was used to review 24 studies that applied CST in one of the following two ways: (a) to investigate the effect of CL on the counselor's performance and (b) to investigate the matching of CL and environmental structure. The results corroborated the predictive power of CST in the context of the counseling situation. First, the prediction that counselors who were matched with a compatible environment (i.e., low-CL individuals with high structure and high-CL individuals with low structure) would perform better than those who were mismatched was corroborated. Second, it was found that low-CL individuals generally performed significantly better in more highly structured environments, whereas high-CL individuals showed only a slight improvement in their matched low-CL environment. Due to the lack of standardization in the task stimulus, it was unclear whether higher CL individuals generally performed better than low-CL individuals. Methodological, conceptual, and training issues are discussed. CL Interactionist ModelThe central principles of the CL interactionist model include (a) CL as a cognitive structure in information processing, (b) environmental structure, and (c) matching of person A portion of this study was conducted while we were at the

Audiotaped supervision sessions from 40 master's level counselor trainees and 40 field placement supervisors were used in a content analytic study of the relation of supervisor and trainee gender to supervisory discourse. The Penman Classification Scheme was used to rate the middle 15 min of supervision on power and involvement dimensions. Classification cells were aggregated to yield high-power, low-power, and high-involvement categories. Transformed proportional data and kappa scores from sequential analysis were used in two 2 × 2 (Gender × Role) multivariate analyses of variance to compare dyad types on speaker's use of message categories and patterns of discourse between speakers. Findings indicated that male and female supervisors reinforced female trainees' high-power messages with low-power, encouraging messages significantly less often than for male trainees. Female trainees responded to supervisor low-power, encouraging messages with high-power messages significantly less often than male trainees.