An application of Kelly's personal construct theory to counselling : a philosophical and empirical study ; a dissertation presented in partial fulfilment of the requirements for the degree of Doctor of Philosophy at Massey University

Counselling is a process of interpersonal interaction in which clients increase their understanding of themselves in relation to others. Investigations of counselling and psychotherapeutic practice have traditionally focused either on the process of counselling or its outcomes (Greenberg, 1986a). More recently, however, researchers have argued that counselling research should focus upon relationships between counselling process and therapeutic outcomes (Greenberg, 1986a, 1986b; Garfield, 1990). The theory of personal constructs, elaborated by Kelly (1955), provides a theoretical framework for the study of both counselling process and its relation to in-session outcomes. This study demonstrates an application of personal construct theory to an investigation of therapeutic process and its relationship to therapeutic outcomes. In this dissertation, the metatheoretical assumptions of Kellian theory are discussed and fundamental theoretical concepts are elaborated. In addition, the theoretical relationship between personal constructs, common constructs and social constructs in the context of counselling is described. This study also extends the work of Proctor (1985a, 1985b, 1987) to suggest a possible theoretical relationship between people's constructs and their verbal behaviour. The establishment of role relationships through verbal interaction is a further subject of discussion. The investigation described in this dissertation involved an application of the fundamental principles and concepts of Kellian theory to a study of counselling and psychotherapeutic practice. Research participants were four therapists and eight clients. The therapists were chosen from those who described their practice orientation as largely consistent with a humanistic-existential approach to therapy. The average age of the therapists was 48.25 years and the average number of years of practice experience was 13.75 years. The average age of the clients who participated in the study was 35.25 years. For the purpose of this investigation, each therapist engaged in a single audio-taped therapeutic interview with each of two clients. Prior to engagement in therapy each therapist and each client completed a personal character sketch. Subsequent to therapy, each client was asked to indicate therapeutic interactions in the preceding counselling session which may have had personal significance. Therapists were also asked to identify in-session therapeutic events which may have had personal significance for their clients. "Laddering" (Hinkle, 1965) and "pyramiding" (Landfield, 1971) techniques were then used to elicit constructs which may have been associated with constructs indicated by participants' verbal behaviour during the course of psychotherapy. Data analysis followed the order of data collection. Guidelines were established for the identification of role construct poles, based upon Davis, Stroud and Green (1989). In addition, a list of categories was developed for the classification of role construct pole expressions derived from the self-characterisations completed by therapists and their clients. Analysis of the therapeutic interviews involved the identification of verbal expressions of therapist and client constructions. In the third, and final stage of analysis, associations between constructs indicated during counselling and constructs indicated in the post-therapy interviews were discussed. The outcomes of this study suggest that therapists may construe themselves with a greater diversity of role constructs than the range of role constructs used by clients in their construal of themselves. Nevertheless, there is evidence to suggest that construct commonality often exists between therapists and their clients. However, there was little similarity between constructs indicated by the content of therapists' self-characterisations, and the constructs indicated by the verbal behaviour of therapists during counselling. In contrast, the verbal behaviour of five clients during counselling indicated constructs which were similar to those indicated by the content of their self-characterisations. A further outcome of this investigation was the identification of six levels ot construing: a) client constructions, b) client superordinate constructions, c) therapist constructions, d) therapist constructions of client constructions, e) therapist constructions of client superordinate constructions, and f) therapist superordinate constructions. This study indicates that therapists' superordinate constructs, which govern their subordinate constructions, have a significant influence upon the counselling process. In particular, client construct system change may follow when therapists and their clients do not share similar superordinate constructs. However, this study suggests that therapists must have constructions of their clients' constructions which are similar to the constructions which their clients have expressed. Only under such circumstances may therapists be able to predict possible changes in clients' construct systems which may occur during counselling. Implications are indicated for counselling and psychotherapy research, therapeutic practice and therapist education and training. This investigation provides further evidence that the character sketch, first proposed by Kelly (1955), can be used in research contexts. Moreover, when character sketches are used in conjunction with a relevant list of categories, they may provide evidence of the constructs which govern the constructions which clients express during counselling. This outcome may have particular relevance to researchers and practitioners. A significant implication of this study is that personal construct theory may be used as a framework for the analysis of counselling and psychotherapeutic interactions in which the therapist adopts an approach to practice which is largely inconsistent with personal construct therapy. Notably, this study demonstrates that an application of personal construct theory to therapy process research enables the identification of links between the overt verbal behaviour of therapists and their clients, and the usually inaccessible psychological processes which govern that behaviour. Interactions in the context of single counselling sessions may be described and occasions of psychological change identified. In addition, apparent links between in-session interactions and therapeutic outcomes may be demonstrated. Thus, applications of personal construct theory to investigations of counselling enable researchers to meet the objective of contemporary research: therapeutic process can be demonstrably linked to therapeutic outcomes. Importantly too, therapist educators may be able to relate student practice to possible client outcomes, thereby enhancing the educational outcomes of counsellor and psychotherapist education and training