A Coherence Model of Cognitive Consistency: Dynamics of Attitude Change During the Persian Gulf War

We describe Co3 (Coherence Model of Cognitive Consistency), a computational model that we used to simulate attitudinal shifts toward various factors related to the Persian Gulf War. Co3 is based on “parallel constraint satisfaction,” a mechanism that revises a set of attitudes so as to maximize overall coherence, with each attitude simultaneously influencing every other related attitude. The Gulf War provided a naturalistic case study for examining the dynamics of attitude change. A survey of attitudes toward U.S. military involvement was administered to 129 students at the University of California, Los Angeles, first during the initial two days of the war, and again two weeks later. At each time, support for U.S. military action was highly correlated (either positively or negatively) with factors indicative of attitudes toward pacifism, the legitimacy of U.S. intervention, isolationism, and President Saddam Hussein of Iraq. A within-subject analysis revealed that shifts in support for the war were correlated with consistent shifts in all four of the major predictors, including those (e.g., pacifism) that would not seem to have been directly affected by events over the intervening time period. This pattern of attitude change demonstrates cognitive consistency. Co3 was used to model how a shift in one attitude due to external inputs (e.g., media reports) can trigger correlated shifts in related attitudes. Computational methods of the sort exemplified by Co3 may be useful in modeling various social psychological phenomena.