Purpose: The objectives of the present study were to investigate 189 male athletes playing soccer, rugby and water polo 1) to demonstrate the presence of the team-serving bias and thereby examine the validity of the Causal Dimension Scale for Teams (CDS-T), a measurement tool of team attribution, and 2) to verify the effects of group cohesion on the team-serving bias. Results: A team-serving bias was found in the locus of causality, stability and external control factors of team attribution, thus validating the measurement tool used in the study. Further, the moderating effect between group cohesion and team-serving bias was verified via hierarchical regression analysis. The direction of the moderating effects, as shown in the interaction graphs, indicated that the team-serving bias was higher for Individual-centred Task-oriented cohesion and Group-centred Task-oriented cohesion. Conclusion: we have shown that the team-serving bias tends to be stronger when group cohesion is higher. Further, we found that among the different types of group cohesion, there were moderating effects of task cohesion but not of social cohesion.