A Dynamic Model of Political Trust: Information, Intention, and Vaccination

• In Event: Science, Trust, and Public Opinion

Abstract

A substream of research has emerged exploring the relationship between political trust, a term broadly characterized as the degree to which individual citizens trust political institutions and policymakers, and individual behavior during crises like the COVID-19 pandemic. Empirical evidence suggests political trust is one of strongest predictors of COVID-19 vaccine uptake as well as the adoption of various non-pharmaceutical interventions like mask wearing.

This research has provided important theoretical and real-world insights for the implications of waning levels of political trust for effective governance during a crisis, however studies to date leave two distinct shortcomings. First, they fail to fully account for the potential malleability of trust in the face of new and, in some cases, misleading information. This is particularly important for novel and emerging risks like COVID-19, where uncertainty, ambiguity, or the omission of facts can prove to be highly problematic given the increasingly politicized and polarized media environment that has emerged over the last several years. Second, individual behavior to reduce risk, including vaccination, is often treated as a static, binary outcome (yes/no). Conceiving of individual behavior in this way effectively ignores the range of possible mismatches between an individual’s intention to reduce risk and their actual realized risk mitigating behaviors in the face of new information and other contextual factors that are framed as socially desirable or undesirable, thereby potentially altering an individual’s trust.

We address these gaps in the existing literature, and provide some of the most comprehensive evidence to date on how political trust as well as other forms institutional trust (e.g. CDC, FDA and public health authorities, healthcare providers, scientists) influence individual COVID-19 vaccination intention and subsequent vaccination behavior. Unlike previous research on political trust, our longitudinal analysis accounts for not only the direct effects of trust on COVID-19 vaccine intentions and behavior, but also the extent to which these effects are mediated by sources of information (e.g., the news, podcasts, public service announcements, etc.) individuals consumed over the course of the pandemic. Capitalizing on the longitudinal information, we further model the dynamic evolution of trust, informing the extent to which political and institutional trust is persistent versus malleable in the face of information flows. Finally, we recognize that the effects of political and institutional trust, and the extent to which individuals’ trust remains sticky over time, are not likely to be uniformly distributed; hence, we also assess heterogeneity in the various responses across sub-populations and geographies.

To accomplish these objectives, we rely on detailed longitudinal information derived from a nationally-representative panel survey of U.S. adult residents at three time points during the first few months of the COVID-19 vaccine rollout:, January-February 2021 (n=3,900 ), March-April 2021 (n=2,268), and June 2021 (n=1,348). Analyses are based on multinomial logit regression models and discrete time hazard models, which carefully account for unobserved time-invariant heterogeneity across persons and across geographic areas (states), and for unobserved national trends, in addition to observed confounding factors.

Authors

• Rob A. DeLeo, Bentley University

• Dhaval Dave, Bentley University

• Kristin Taylor, Wayne State University

• Danielle Blanch-Hartigan

• Deserai Anderson Crow, University of Colorado Denver

• Thomas A. Birkland, North Carolina State University

• Elizabeth Ann Shanahan, Montana State University