Qualitative Change Process Research on Psychotherapy: Alternative Strategies.

Abstract

Originally, research on psychotherapy fell into two divisions, outcome research, which dealt with the extent to which clients over the course of therapy, and process research, which investigated what occurs within therapy sessions. Change process research (Greenberg, 1986) bridges these two fields: It is the study of the processes (sometimes referred to as effective ingredients) that bring about changes, including the temporal course of those changes. Thus process research concerns itself with explaining both how and why occurs in therapy. Traditionally the mode of understanding assumed to operate in process research has been causal in nature, an assumption which is implicit in physicalist metaphors such as change mechanisms and effective ingredients. Furthermore, most process research to date has been quantitative and hypothesistesting, reflecting not only the influence of positivism but also researchers' desires to test strong causal theories, such as Rogers' (1957) formulation of the necessary and sufficient conditions for in therapy. For most researchers, causal inference and quantitative assessment have been perceived as tightly linked elements of the standard modus operandi. Change process research has thus most commonly taken the form of process-outcome studies, in which in-session processes are measured quantitatively on indices of quantity (e.g., number of interpretations in a session) or intensity (e.g., degree to which the therapist was empathic); in some cases, process was manipulated experimentally (e.g., randomly assigning clients to leading or following therapist approaches). Measured or manipulated therapy process was then used to predict outcome, usually post-therapy clinical distress or other quantitative measures of personal functioning, often controlling for pre-therapy status. This genre of research has produced well in excess of 2,300 findings since 1950 (Orlinsky, Grawe & Parks, 1994). The process-outcome literature, however, is filled with disappointingly general or contradictory results (cf. Shapiro, Harper, Startup, Reynolds, Bird & Suokas, 1994), particularly when one attempts to move from global, evaluative, relational variables (e.g., therapeutic alliance) to more specific, clinically useful phenomena (e.g., type of therapist intervention; Orlinsky et al., 1994; Stiles, 1996). Furthermore, Stiles (1988, 1996; Stiles & Shapiro, 1989) has strongly criticized the quantitative process-outcome paradigm on various grounds, mostly having to do with the simplistic assumptions it makes about the nature of the therapy process (e.g., if something is good, then more of it must always be better). This emerging disillusionment with quantitative process-outcome research is by no means universally shared (Hayes, Castonguay & Goldfried, 1996), and the continuing accumulation of quantitative process research can be expected to continue into the foreseeable future. Nevertheless, even at their best, such research designs are blunt instruments for understanding anything.as complex and nuanced as the process of in psychotherapy. The vast accumulation of general or contradictory research findings conceals our fundamental ignorance about how individual clients actually grow and in the course of their therapies. Perhaps it is past time for researchers to adopt more open, discovery-oriented methods for understanding therapeutic change. It is true that the era of qualitative therapy research has begun, marked by the publication of special issues on the subject in the Journal of Counseling Psychology (1994) and Psychotherapy Research (1999). However, as Polkinghorne (1994) noted, the range of qualitative research strategies so far applied has been fairly limited. Thus, the potential of qualitative approaches for illuminating the process in therapy has not yet been fully realized. Our purpose in writing this chapter is to encourage the use of a broad range of options for qualitative data collection and analysis in psychotherapy process research. …