Modeling group process constructs at three stages in group psychotherapy.

This study examined the relationships among group therapy processes measured by the Working Alliance Inventory-Short Form, the Therapeutic Factors Inventory Cohesiveness subscale, and the Group Climate Questionnaire-Short Form in a sample of 145 patients attending 18 psychodynamic groups. Five hypothesized models were tested early in therapy (Sessions 3 and 4) using multilevel confirmatory factor analysis. Two three-factor models approached conventional standards of model fit. By merging these two models, a three-factor model consisting of member-leader alliance, positive bonding relationship, and negative relationship fit the data well. Later in therapy, member-leader bonding was no longer important to member-group cohesion, indicating that cohesion and alliance and the member-leader versus member-group bonding represent different processes.