To determine the relative contribution of group interaction and information exchange to the risky-shift effect, female undergraduates were assigned either to interacting or to listening groups. Tape recordings of discussions of risk dilemmas were derived from the interacting groups. These taped discussions comprised the stimulus material for the listening groups. Thus, information was held fairly constant for interacting-listening pairs. Individual levels of risk taking measured prior to and after group discussion were used to assess the magnitude of the risky-shift effect. Although both group types manifested significant risky shifts, the interacting groups significantly exceeded the listening groups in extent of shift. We concluded that informational processes alone could not fully account for the risky- shift phenomenon.