Background: Research has suggested that therapists differ in their effectiveness (e.g., Baldwin & Imel, 2013); however, less is known about favorable therapist characteristics, especially for group therapy outcomes. The current study is part of a larger project investigating the personal and professional development of trainees that undergo group-analytic training including participation in experiential groups at the Institute of Group Analysis and Group Psychotherapy (IGA) in Oslo. A description of the larger project is given in a separate report (Leitemo & Vestbø, 2017). In this larger study we found a significant reduction in trainee self-reported interpersonal problems over a training period of 10 months. Objective: The aim of the current study was to explore whether measures of attachment dimensions (i.e., anxiety and avoidance) in the trainer (group analyst), the trainees and in the group, as well as measures of therapeutic relationships quality in the group, could explain the change in trainee interpersonal problems found earlier. Of interest were both direct effects of attachment dimensions as well as possible interactions between the participants’ attachment dimensions on change in interpersonal problems. Method: In total, 53 candidates and 8 trainers involved at the IGA training program participated in the study. The participants completed the Circumplex of Interpersonal Problems at the beginning and at the end of the study period, the Group Questionnaire at the beginning of the study period and the Experiences in Close Relationships just prior to the study period. Due to the hierarchical nature of the data, we used multilevel modeling in order to analyze the impact of attachment dimensions on change in trainee interpersonal problems. Results: We found that lower attachment anxiety in the trainer group analyst and in the group (modelled as the group members’ average level) predicted a reduction in trainee interpersonal problems. Furthermore, we found a significant interaction effect suggesting that trainees with higher levels of attachment anxiety benefitted from participating in a group with higher average level of attachment avoidance. Measures of the quality of group relationships were insignificant predictors of trainee outcome. Conclusion: Our findings support the use of attachment theory as a valuable framework to understand differences in group therapist and group effectiveness and as a promising avenue for future group psychotherapy research.