In the 1st 2 of 3 experiments with 99 female undergraduates, the movement of group members toward or away from an S's preference was manipulated while holding constant the content of group discussion. The only significant effect of such shifts was an antimomentum effect. Data demonstrate that the inverse relationship between overall level of support and likelihood of changing one's position is not solely attributable to large factions' ability to generate more arguments. In a final experiment, 30 male and 28 female undergraduates were given an opportunity to defend their preference. There was no momentum effect. Analyses of the content of Ss' speech paralleled current data on opinion change: Ss were more sensitive to current levels of support than to changes in the level of support.