Deliberative Opinion Change and Consistency: Citizens in the 1994 Illinois Gubernatorial Election

Does deliberation lead to improvements in public opinion quality? The answer depends upon what standards are used. I argue that two widely used criteria ­ stability and consistency ­ should be modified for use in deliberation research. When citizens discuss politics we often desire instability after deliberation (through evidence of opinion change) and inconsistency during deliberation (when participants consider multiple viewpoints). In contrast to past work comparing pre- and post-deliberative opinions, here the actual deliberation captured in two focus group discussions during the 1994 Illinois gubernatorial election (N=2,564 statements) is used to predict the probabilities of vote change and candidate choice. I find that consistency operates selectively. Candidate consistency ­ when deliberation matches vote choice and when the deliberative statements themselves tend to favor one candidate ­ decreases the likelihood that citizens change their vote. Issue consistency ­ when deliberation matches voting explanations on a salient issue like crime ­ increases the probability of vote change. These same relationships hold for predicting the probability of voting for the group¹s most preferred candidate, but they are inverted for the least preferred candidate. That is, the people who eventually favored the winning candidate deliberated one way. Those in the minority who remained with an unpopular candidate tended to deliberate another way. This means that deliberation influences political opinions differently for different citizens. Statewide public opinion before and after the focus groups strongly suggests that the deliberative processes at work in the focus groups closely mirrored events during the rest of the election.