The ability to analyze the therapeutic components of an activity is an important skill for occupational therapists. This study examined two potentially significant factors in activity analysis: the use of humor and the effect of group structure. Four groups (two with a parallel structure and two with a project structure) participated in a hat-making activity designed to elicit humor. Four groups (two with a parallel structure and two with a project structure) participated in a bookmark-making activity. The 28 female subjects’ affective responses were measured by Osgood’s short-form semantic differential, and the cohesion among group members was assessed by the Group Environment Scale. Results indicated that subjects who participated in groups which included humor rated their activity significantly higher on two factors of affective meaning (evaluation and action) and significantly higher in terms of cohesion. There was a significant interaction between the two activities and group structure in terms of the action factor and cohesion. In both cases the parallel groups making bookmarks received particularly low scores. The findings have implications for conceptualizing occupational therapy group activities.