Sequential Analysis of Topics in Psychotherapy Discourse: Single-Case Study

Abstract

While content analysis of discourse has a rich history, relatively little research has been aimed at identrfying topics by their content. In this paper we test the assumption that identrfying the content and sequences of topics reveals clinically useful information about a person's mental structure, by studying change in the discourse structure over the course of a single therapy and by demonstrating the relevance the discourse structure and content have for the case. A new method is presented for idiographically identrfying topics of discourse by content and applied to the entire therapy of a single case of prolonged and severe bereavement. Through sequential analysis of the structure of patient discourse, sequential patterns of associated topics were revealed that converged with clinical case formulations and with clinicians' and patient's ratings of the importance of the identified topics to the case. Patterns of changing associations among topics from the first to the second half of therapy are interpreted as indicating an increased flexibility in the accessibility of clinically relevant topics.When a patient discusses a topic in therapy we expect the expressed content to indicate important aspects of how a patient thinks about ideas and felt emotions. Manifest in these communications is a patient's current state of organizing knowledge. In pathological states the organization of knowledge structures can include contradictory contents (Strauman & Higgins, 1988) as well as rigidly repeating, rather than flexible, ideational sequences (Horowitz, 1988;Segal, 1988). Important aspects of knowledge organization-healthy or pathological-can be examined by studying the patterns of associations among presented topics. This paper describes research which identifies topic content in the psy-

Thematic segmentation of psychotherapy transcripts for convergent analyses.

“…These have included a study measuring information production and avoidance of both individual and small groups, and demonstrating differences by topic and distress (Horowitz, Milbrath, Reidbord & Stinson, 1993); a study relating these same variables to computer text measures of anxious dysfluency and hedging (Stinson et al, 1993); studies using multiple variables to locate and describe important moments in psychotherapy (Horowitz, Stinson, Fridhandler et al, 1993; Stinson & Horowitz, 1993); a study comparing and contrasting single subject and small group patterns of verbal defensiveness and computer measures of dysfluency and speaker stance (Stinson, Milbrath & Horowitz, 1993); and a study comparing multidimensional scalings of self-report measures with computer content measures of therapy session discourse about clinically relevant objects of discourse (Hart et al, 1993). The units have also been used in the development of multidimensional ratings of topic and in examining the relationship of topic sequence to avoidance and to change over course of therapy (Milbrath et al, 1993). Bucci (1988) has used TUs for coding referential activity in convergent research, and the approach is being applied by a consortium of researchers within the American Psychoanalytic Association who are using many measures on the same therapy session transcripts (Waldron, et al, 1991).…”

Clients' and Therapists' Stories about Psychotherapy

“…She found that treatment outcome was not associated with congruence between the perspectives, but instead, outcome was associated with the extent to which each participant, therapist or client, felt the key ingredient (either the development of insight or problem solving) had been present. Disconnections like these have been further identified in subsequent research (e.g., Levander & Werbart, ; Lietaer, ; Milbrath et al., ; Tyron, Blackwell, & Hammel, ; Wark, ; Werbart & Levander, , , ).…”

Expressive and Defensive Behavior during Discourse on Unresolved Topics: A Single Case Study of Pathological Grief

“…In an iterative process of reading the transcript and reflecting, judges identified, refined, and translated content in psychological terms to concrete descriptions that best fit the phenomenon. Wherever possible the subject's original language was retained in this reduction procedure until a final set of category labels and brief descriptions was derived (see Milbrath, Bauknight, Horowitz, Amaro, & Sugahara, 1993).…”