Despite the fact that leading psychotherapy researchers have persistently endorsed the use of intensive analysis procedures over the years, the field has been slow to act upon these recommendations. This article identifies a number of factors responsible for this slowness: the uncritical utilization of conventional research methodologies, the failure to distinguish between different types of intensive analysis designs, the assumption that intensive analysis is necessarily unscientific, and the fact that the behavioral single-case experimental design are inappropriate for studying other forms of therapy. The task-analysis approach to psychotherapy research is presented as a methodologically rigorous alternative to the behavioral single-case experimental design.

A classification system, designed to focus on 2 aspects of client behavior, voice quality and expressive stance, was applied to interviews of S3 clients. The behavior thus coded was vector analyzed by columns, yielding factor loadings for interviews for each client for 1st, 2nd, and llth interviews. Estimated factor loadings were obtained for 12 attrition clients for 1st and 2nd interviews. Attrition could be predicted from 1st interview process. For the 2nd and llth interviews, 2 of the 4 factors extracted were found to differentiate among outcome groups formed by combining client's and therapist's vantage points. The authors discuss the value of these client-process variables not only for prognosis, but also as a background against which to evaluate therapist participation or experimental intervention.

In this article, we describe a task-focused approach that uses sequential analyses as deductive techniques for studying therapist-client interactions in the context of clinical microtheories of change events. The methodology is demonstrated in a study of a class of change events in client-centered therapy referred to as the resolution of problematic reactions. Logit-loglinear analysis and binomial z scores were used to test the effects of therapist behavior, assessed by therapist vocal quality and by the therapist task-relevant system, on client process assessed by client vocal quality and by the Experiencing Scale. The results indicated (a) that therapist productive voice facilitated a shift from poor to productive client voice and a shift from low to intermediate experiencing and (b) that therapist task-specific interventions directed toward resolution facilitated shifts to high experiencing. This task-focused approach to sequential analyses has the potential to yield clinically and theoretically relevant findings.

This study reports three cases in which the effects on client depth of experiencing and voice quality of the Gestalt "two chair" operation were investigated. In a design using the clients as their own

The purpose of this study was to develop and test a system for classifying some aspects of the therapist's behavior during the interview. The focus was on process rather than on content and involved both vocal and lexical aspects of the therapist's style of participation. 30 therapist responses were sampled from the 2nd and from the next-to-last interviews of each of 20 cases and were classified on each of 3 aspects. The behavior thus coded was vector analyzed by columns, thereby yielding factor loadings for interviews. Loadings on 2 of the resultant 3 factors were found to be related to case outcome as viewed by client and therapist. The findings for late interviews were more clear-cut than those for early interviews. Loadings on the same 2 factors were found to be related to the therapist's level of experience.