Objective: Group-based physical training interventions have been shown to be effective in increasing quality of life in cancer survivors. Until now, however, the impact of cohesion within the group on intervention outcome has not been investigated. Methods: We examined self-reported individual group cohesion ratings collected in the first half of a 12-week rehabilitation programme for cancer survivors (N = 132). Four dimensions of group cohesion were measured, i.e. the bond with the group as whole, the bond with other members, cooperation within the group and the instrumental value. Quality of life, physical functioning and fatigue were assessed before and after the intervention using the European Organization for Research and Treatment of Cancer Quality of Life Questionnaire-C30. Linear multiple multivariate regression analysis was conducted to explore the relationship between group cohesion and intervention outcome. Results: The relationship between group cohesion and outcome was significantly modified by gender. Higher ratings of cooperation within the group predicted better post-intervention quality of life and physical functioning and less fatigue in men, and better quality of life and physical functioning in women. Additionally, women who reported a stronger bond with other members showed a lower quality of life after the intervention. No relationship was found between the instrumental value and the outcome variables. Conclusion: Some dimensions of group cohesion seem to be associated with intervention outcome. The underlying mechanisms need to be unravelled. Copyright (C) 2007 John Wiley & Sons, Ltd.