published_or_final_version / abstract / Politics and Public Administration / Doctoral / Doctor of Philosophy

This dissertation proposes a theoretical framework of attitude change under threatening conditions based on parallel updating. More specifically, I focus on public preferences for policies to address terrorist attacks, pandemics, climate change and natural disasters in periods when these threats are elevated. I test my argument with four original survey experiments, which include eleven interventions and draw on a nationally diverse sample of a total of 9,110 American citizens. These interventions identify the effects of factual information, partisan cues, incidental emotions, ideological and non-ideological framing, and memory priming. Evidence from these experiments provides consistent support that public opinion updating exhibits five characteristics. First, citizens change their views by a small amount. Second, citizens’ opinions move in the direction of information. Third, attitude change occurs regardless of political predispositions and individual attributes. Fourth, exposure to information about a specific policy area does not impact preferences for policies unrelated to this area. The only exception to this rule is when the treatment is emotionally strong. Finally, attitude- and identity-based cross pressures may introduce only minimal bias in the manner citizens update their opinions. These conclusions strongly challenge theories of public opinion which argue that individual differences in more-or-less enduring political and psychological characteristics can lead to political polarization. Although the persuasive techniques studied here are not equally potent in changing political views, the findings invite cautious optimism about the capacity of citizens to update opinions in a reasonable and accurate manner, even when the circumstances are unfavorable. Finally, the results suggest that the roots of polarization should be searched for more directly, notably in the increasingly fragmented political, social, and media environments.