Session impact in counseling process research: Comment on Elliott and Wexler (1994) and Stiles et al. (1994).

Abstract

R. Elliott and M. M. Wexler (1994) have introduced the Session Impacts Scale, and W. B. Stiles, S. Reynolds, G. E. Hardy, A. Rees, M. Barkham, and D. A. Shapiro (1994) have undertaken a new factor analysis of the Session Evaluation Questionnaire with a British sample of clients. These 2 measures of session-level psychotherapy effects have great potential as research tools. In this comment on the 2 articles, issues of retest reliability, the connection of session-level measures to research strategies based on other units of analyses, and suggested uses of the SIS and the SEQ are discussed. The importance of context-sensi tive session impact research and of a comprehensive theory relating session impacts to therapy outcome is also stressed. Session impact has received sharply increased attention from counseling process researchers in the past decade. No doubt part of this interest stems from disappointment over the small yield, in terms of research results therapists can readily apply in practice, of two other research traditions, namely, studies of therapist technical activity in momentto-moment interactions and comparative outcome studies of the entire course of therapy. For example, despite the impressive body of knowledge gained from microanalytic studies connecting in-session therapist intentions and behavior with client reactions (cf. Hill, 1992), researchers despair of ever cataloging enough procedural knowledge, given the contextual complexity, to adequately inform a science of psychotherapy (Friedlander, 1992; Martin, 1992). At the macroanalytic level, studies comparing the outcome of various treatments have yielded disappointingly few specific results and have raised the question of whether there are any important practical differences between types of therapy (Stiles, Shapiro, & Elliott, 1986). Elliott, Stiles, and their associates in Elliott and Wexler (1994), Stiles et al. (1994), and elsewhere (Stiles, 1980) have convincingly argued that examination of session-level variables offers a bridge between the molecular and molar levels of analysis that promises to be more productive than either approach alone.