Group processes in short- and long-term psychodynamic group psychotherapy

This thesis examined the processes of alliance, cohesion, and group climate in a sample of 145 patients attending 9 short-term (20 sessions) and 9 long-term (80 sessions) psychodynamic psychotherapy groups. Concepts were operationalized through the Working Alliance Inventory – Short Form (WAI-S), the Therapeutic Factors Inventory, subscale Cohesiveness (COH), and the Group Climate Questionnaire – Short Form (GCQ-S). Three waves of data collection were applied for the measurement of alliance and cohesion (sessions 3, 10, and 17), whereas five time-points were used in the measurement of group climate development (sessions 3, 10, 17, 39, and 77). Study I examined the interrelatedness of alliance, cohesion, and group climate (GCQ-S). Five hypothesized models of group processes were tested early in therapy using multilevel confirmatory analyses. The two three-factor models that approached conventional standards of model fit were merged, and a three-factor model consisting of member-leader alliance, positive bonding relationship, and negative relationship fit the data well. Later in therapy, the bonding between member and leader was no longer important for the member-group bonding, and the model was then better described as member-leader alliance, member-group cohesion, and negative relationship. Results indicated that the processes of alliance and cohesion, and the member-leader versus the member-group relationship structure, evolve as different processes of psychodynamic group psychotherapy. There were no differences in factor structure relative to group format (short-term, long-term). Study II examined the sources of influence on alliance and cohesion. Within the framework of generalizability theory the 14 variance components identifiable by the research design were estimated. Results indicated that patient variability was the strongest clinically relevant contribution to both alliance and cohesion. Therapists were important for alliance through all the measured stages, but for cohesion only in the middle stage. The therapist x group interaction accounted for a substantial proportion of alliance variability early in therapy and for cohesion variability within the first two stages, but this contribution then decreased. Group length did not account for any of the variance in alliance or cohesion measures. Study III examined the development of group climate (engagement, avoiding, and conflict) in short- and long-term groups. Linear mixed models were used to compare changes in group climate over time. The development of engagement was similar in the two psychotherapy formats. During the first 18 sessions, conflict and avoidance decreased toward the termination of the short-term groups, in contrast to an increase in this still-early stage of the long-term groups. When compared according to the stages of therapy (early, middle, and late), a low-high-low pattern for conflict and avoidance emerged in both psychotherapy formats, with a stronger decrease toward termination in long-term groups. Results suggest an accelerated progress of group climate development within short-term groups, compared to a delayed but strengthened process development in long-term groups