There are many ways in which a group support system (GSS) can be used to support group brainstorming. This paper reports the results of an experiment that manipulated task structure and time structure. Groups electronically brainstormed on intact tasks (where all parts of the task were presented simultaneously) or on partitioned tasks (where three subcategories of the task were presented to the groups). The time periods in which groups worked were either one 30-minute time period or three 10-minute periods separated by two-minute breaks. Groups in the partitioned task treatment generated 40% more ideas, but there were no time effects. These differences are attributed to the ability of the partitioned task to refocus members' attention more evenly across the entire solution space.