Date of Award


Degree Type


Degree Name

Doctor of Philosophy (PhD)

Graduate Group


First Advisor

Robert C. Hornik


Public health messaging campaigns can be useful tools for addressing vaccine hesitancy. Behavior change theory-driven formative research, which guides campaign planners in message topic selection, is necessary for campaigns’ success. However, such research cannot advise on how to communicate campaign messages. Information processing theories propose avenues through which messages can be made more persuasive.

Focusing on Black Americans, this dissertation aims to test an approach to improve future public health messaging campaigns targeting COVID-19 vaccine hesitancy. Specifically, it investigates whether source credibility (an information processing theory variable) can enhance the persuasiveness of differentially promising campaign message topics. These message topics refer to beliefs associated with intention to receive a COVID-19 vaccine.

Study 1 was a two-part belief elicitation. In the first part (Study 1A) participants (N=117) reported their most salient beliefs related to the COVD-19 vaccine in an open-ended survey. The second part involved topic modeling on 5945 tweets discussing the vaccine. Beliefs identified from these studies were transformed into cross-sectional survey items for Study 2. Results from this study highlighted the utility in combining traditional survey data with publicly available social media data to learn about the health concerns of a target population.

Study 2 (N=500) identified beliefs varying in promise, following the Hornik & Woolf approach, to be transformed into message topics. Study 3 was a message experiment where participants (N=245) were exposed to messages, addressing either a “more” or “less” promising topic, designed to increase vaccination intention. The messages were also either communicated by a health professional, or layperson source.

Results of Study 3 indicated no main effects of neither message promise, nor source credibility on vaccination intention. Messages targeting the “less” promising topic shifted the linked belief, but the belief was not associated with intention. Messages on the “more” promising topic did not shift the linked belief, but the belief was strongly associated with intention. Evidence from these studies underline the benefit of choosing the most promising message topics when planning a public health messaging campaign. These results also suggest that additional research is needed to identify credible sources for vaccine non-intenders, and other differentiating characteristics of this unvaccinated population to develop effective public messaging campaigns in the future.