Twelve four-person female groups of subjects displaying the typical underestimation of their peers' (relative to their own) risk acceptance were compared with twelve groups of subjects who (slightly) overestimated their peers' risk acceptance. Risk level was measured by responses to a set of hypothetical decision situations known to elicit risky shift on the basis of previous research. Risky shift following group discussion was not found to be different for the two types of groups, casting doubt on the widely suggested role of peer underestimation in risky shift. Nor was risky shift affected by whether or not group members stated their individual decisions publicly at the close of discussion. Larger group risky shifts were accompanied by higher self-ratings given by group members on a number of polarity scales. In discussing the findings, we outline an explanation of group-induced shifts in risk taking, emphasizing the motivational and informational inducements provided by group discussion whereby group members come to discard their prior positions in favor of more aspired ones.