Date of Award


Document Type


Degree Name

Doctor of Philosophy (PhD)


The present study attempted to either extend, modify or challenge Rokeach's (1973) Self-confrontation theory of value change. Rokeach (1973) found that subjects who were confronted with specific self-confrontation messages, which were designed to make differences between their value systems and self-concepts salient, were more likely than control subjects to change the rank-ordering of the values Freedom and Equality in a ranking task containing 18 values. Rokeach explained this phenomenon by invoking the concept of self-dissatisfaction, which he said subjects came to feel as a result of being exposed to the self-confrontation messages. The present study compared this explanation with the Reference Group Influence theory of Kelman (1958) and others who felt that subjects would modify values merely by identifying with certain positively perceived reference groups. In a simple single-factor design, four experimental conditions were compared with a control condition. The experimental conditions consisted of separate self-confrontation manipulations derived from Rokeach's original manipulation. These were: (1) an Intrinsic Inconsistency condition which implied to subjects an inconsistency between their own value systems and self-concept, based on information not taken from an external source; (2) a Reference Group only condition, which exposed subjects to value rankings of a potential reference group; (3) an Extrinsic Inconsistency condition, which implied an inconsistency between values and self-concept, based on the value rankings of the reference group; (4) a Replication condition, which was a virtual replication of Rokeach (1973); and (5) a Control condition. Behavioral measures which were thought to reflect value change were also taken. Results indicated that the Intrinsic Inconsistency subjects were the only ones to elevate rankings of the value "Equality" significantly more than the controls. Although the Intrinsic Inconsistency condition's responses to behavioral measures were also greatest among the five conditions, this relationship was not significant. Thus only limited support was found for Rokeach's self-confrontation theory, and no evidence was found for a reference group influence effect. Possible explanations for these findings were offered as well as possibilities for future research.