One member of each pair of 52 undergraduates, referred to as a learner, was trained BC conditional discriminations, with B stimuli as the samples and C stimuli as the correct comparisons. Responses of the learner were either reinforced or punished by another member of each pair, an instructor, who had previously mastered AC conditional discriminations, with A stimuli as the samples. The B sample stimuli were presented simultaneously with the A stimuli so that only the instructor could see the latter. Each learner mastered the BC conditional discriminations from his or her instructor, regardless of whether the learner saw the instructor in advance or not. Performance was less accurate when the B stimuli corresponded 83.3% with the A stimuli than when they corresponded 100%. These results extend previous findings on private events obtained from only two of eight pairs when the stimuli always corresponded. (Contains 1 table and 7 figures.)