First graders, pretested on conservation tasks, were assigned to groups consisting of one conserver and one nonconserver. Groups then had to give a joint response. Nonconservers yielded more often to conservers. Subsequent tests indicated nonconservers improved on the tasks. (ST)