FIGURE 12.4 Factors That Decrease Destructive Obedience By systematically varying his basic experimental design, Milgram identified several factors that diminish the likelihood of destructive obedience. In this graph, you can see the percentage of subjects who administered the maximum shock in different experimental variations. For example, when Milgram’s subjects observed what they thought were two other subjects disobeying the experimenter, the real subjects followed their lead 90 percent of the time and refused to continue.

Source: Data from Milgram (1974a).