high visibility within the inmate system, thereby generating and reinforcing the image of a culture marked in conflict with the values of the administration." (p. 291). In a similar vein, Klofas and Toch ( 1982) found that those guards and prisoners with the most hardline positions (subculture custodians, in their terms) were inclined to adopt self-appointed roles as spokesmen for their respective constituencies. Packard and Willower ( 1972) also acknowledged that the pluralistic ignorance they observed among school teachers may have been triggered by witnessing custodial behavior in places of high visibility within the school.
The tendency of those who embody the norms of a group to find themselves in positions of high influence has been observed in less formal situations as well. For example, Newcomb, ( 1943) in his classic study of social life at Bennington College, noted that those students who espoused most vocally Benningtons' tradition of liberalism tended to be popular and prominent in positions of leadership. Korte ( 1972) study of social life at Vassar tied the greater prominence of cultural conformists to pluralistic ignorance more directly. Specifically, he attributed the finding that Vassar students perceived themselves to be less socially and politically liberal than the majority of their peers to the fact that those students who embodied the liberal values of the institutional norm tended to be particularly prominent and conspicuous on campus.
The mistaken perception of personal deviance among the majority may arise initially because of the high visibility of those minority members who best embody the prototypical group norms. It may be perpetuated, however, by the widespread public conformity it induces. The illusion of personal deviance that begins with the misinterpretation of the representativeness of a vocal minority is thus perpetuated by a misinterpretation of the conformity of a formerly silent majority.
In recent years, students of the relation between attitudes and behaviors have contributed significantly to our understanding of the nature and conditions of attitude-behavior consistency. In the process, they have made a number of choices regarding the kinds of situations and behaviors they would study and the kind of consistency they would try to predict. They have chosen to focus on correlational consistency between attitudes and reasoned or planned actions, a strategy that seems to us quite sensible if the goal is to improve models of behavioral prediction. At the same time, this strategy has left a wide range of behaviors and many interesting questions about attitude-behavior consistency unexplored.