EXTREMITIZATION OF ATTITUDES - DOES THOUGHT-INDUCED AND DISCUSSION-INDUCED POLARIZATION CUMULATE

A within-subjects experimental design with 240 college undergraduates revealed that the typical paradigms used to make attitudes more extrem e in the laboratory (individual thought and group discussion) failed t o produce cumulative change over time. Over the course of 4 sessions o f thought and/or discussion spread over 2 weeks, attitude polarization for both political and personal issues was observed only from pretest to following Session 1 of individualized thought-listing. Reading mes sages from 2 to 5 other participants in Session 2 did not extremify at titudes beyond the thought-induced polarization created in Session 1. Even so, attitude polarization was most consistent when participants d iscussed the issue with other participants over electronic mail-compar ed to thought listing (significant polarization following some session s) and thought only (a control condition of repeated expression; no po larization found). Although polarization did not cumulate, it also did not dissipate (or depolarize) after 2 weeks. Correlational analyses i ndicated that increases in attitude importance tended to accompany inc reases in attitude extremity. Results are discussed from the perspecti ve of the catastrophe theory of attitudes, dynamic social impact theor y, and group polarization literature.