Message-induced polarization and depolarization through the emotional, authoritarian, and rational justification of attitude change

Date of Completion

January 2005


Psychology, Social|Mass Communications




The literature on self-persuasion indicates that mere thought inductions should polarize attitudes whereas reason analysis inductions should depolarize attitudes, but that these effects may be moderated by trait variables (Need for Cognition, emotionalism, dogmatism) or state variables (subjective knowledge or attitude accessibility). The main purpose of this dissertation was to subsume the full scope of these findings under Belief Systems Theory. Participants from the University of Connecticut (N = 204) responded to an on-line experiment regarding attitudes toward alcohol or non-organic food. Participants were assigned to one of 8 cells in a 3 (mere thought, distraction, or reason analysis) by 2 (schema induction or not) by 2 (alcohol topic or non-organic food topic) design. Results indicated that the mere thought and reason analysis effects depend on whether single-item or multi-item scales are used. ^ On the single-item scales, results indicated the presence of a 3-way interaction on the magnitude of change with neutral point measure of polarization. Analysis of the means showed that both the mere thought and reason analysis manipulations polarized attitudes for women who did not receive a schema induction and depolarized attitudes for women who received a schema induction. Mere thought and reason analysis had little impact on polarization for men. It was unclear whether this finding is a sampling or measurement artifact or a complex interaction of gender and knowledge activation. There was no evidence that Need for Cognition, emotionalism, or dogmatism moderated mere thought effects. Nor was there evidence that subjective knowledge or attitude accessibility moderated reason analysis effects. ^ On the multi-item scale, causal modeling revealed that mere thought generated a small degree of attitude polarization whereas reason analysis generated a small degree of depolarization. Need for cognition decreased both emotionalism and dogmatism. Two types of effect were observed within attitude structures. First, top-down effects occurred for both the alcohol and food topics. Second, spreading activation effects occurred. A more comprehensive polarization was then conducted to examine the primary and secondary effects of the thought manipulations. ^