This study presents, summarizes, and analyzes the results of 178 Groups Support Systems (GSS) experiments based on categories of groups per treatment condition. The results suggest that for studies designed with less than 7 groups per condition there are significantly lower probabilities of finding significant differences. Task type and GSS type further moderate these results.
Fjermestad, Jerry, "In GSS Research How Many Groups per Treatment Condition Are Enough?" (1998). AMCIS 1998 Proceedings. 160.