Affective-cognitive consistency and thought-induced attitude polarization.

Subjects whose preexperimental attitudes toward either capital punishment or censorship were high or low in affective-cognitive consistency were identified. These four groups thought about their attitudes by writing two essays, one on the topic for which consistency had been assessed (relevant essay) and one on the unassessed topic (distractor essay). In accord with the hypothesis that thought-induced attitude polarization requires the presence of a well-developed knowledge structure, high-consistency subjects evidenced greater polarization than low-consistency subjects only on the relevant topic after writing the relevant essay. Content analyses of subjects' relevant essays yielded additional data confirming Tesser's ideas regarding mediation: High (vs. low) consistency subjects expressed a greater proportion of cognitions that were evaluatively consistent with their prior affect toward the attitude object and a smaller proportion of evaluatively inconsistent and neutral cognitions. Moreover, although high-and low-consistency subjects did not differ in the amount of attitudinally relevant information they possessed or their awareness of inconsistent cognitions, their method of dealing with discrepant information diverged: High-consistency subjects evidenced a greater tendency to assimilate discrepant information by generating refutational thoughts that discredited or minimized the importance of inconsistent information.