The present experiment was designed to explore the effects of sets toward the characteristics of a partner upon an actor's subsequent behavior when the actor is involved in a group problem-solving situation. Two sets were created; one in which the subject believed he or she was interacting with a "human" as partner and the second, with a "machine" as partner. The hypothesis tested was that the individual will approach the task in the same manner, regardless of differences in these sets toward the origin of the response to one's acts. The data showed no difference for the subjects under the two conditions. The attitudinal measure showed that those subjects working under the "human" condition regarded their partners as more confused than did those working under the "machine" condition. Although the results of this experiment cannot be generalized to all kinds of group problem-solving tasks, they do indicate the existence of components in such tasks that are not dependent on social characteristics that have been commonly studied. Further research on these basic components may lead to better understanding of mutual problem solving.