Using activity analysis, occupational therapists structure group activities to meet the needs and goals of group members. This study compared the effects of project group structure (subjects participate together on a common task) with parallel group structure (subjects each have a task) in a sample of healthy seniors (32 women and 9 men). Subjects randomly assigned to either experimental condition made creative and imitative collages in a counterbalanced order. Time-sampled observations revealed that project group subjects talked and looked at others significantly more than parallel group subjects. On-task and laughing behaviors were also measured, but results were inconclusive. Project group subjects rated their activities higher on the action factor of the Osgood semantic differential (OSD) than did parallel group subjects, but there were no other differences on the OSD or on MacKenzie’s Group Climate Questionnaire. Nor were there significant main effects between the creative collage and the imitative collage. Results are discussed both in terms of group task structure as an occupational therapy tool and in terms of methodologies for future activity analysis research.