Impact of Narrative Messages on COVID-19 Vaccine Uptake Intentions

• In Event: Advances in Theories of the Policy Process

Abstract

In the face of increasing vaccine hesitancy, public health officials are fervently working to identify effective communication strategies to increase COVID-19 vaccine uptake. This NSF-funded study uses an experimental survey design to test (i) the direct effect of narrative and non-narrative messages on intention to get the COVID-19 vaccine (others and self), (ii) the indirect mediating effects of affective response to the treatment conditions on intent to vaccinate, and (iii) the indirect moderating effects of a priori beliefs of individualism and collectivism (Triandis & Gelfland, 1998) on affective response and intent to vaccinate. The sample is representative of U.S. citizens (n = 3400). The questionnaire was launched January 15, 2021 and is in the field as of January 28, 2021.

The message treatments are graphic images pairing narrative words with a call to action. Three treatments test the narrative mechanism of character selection (Shanahan et al., 2019) by manipulating the victim character: yourself, your circle (e.g., family and close friends), and your community. The hero character (target audience) and call to action (vaccinate) are held constant across all conditions, with the latter being presented separately as the control condition. Assignment to treatment and control conditions is randomized with approximately equal numbers of participants in each condition. The survey also includes a suite of questions gauge the participants’ experience with COVID-19, vaccine history and attitudes, risk perceptions, risk mitigation and exacerbation behaviors, worldviews on individualism and collectivism, and demographics. Outcome variables include questions about affective response and intention to vaccinate.

Unlike previous studies of the impacts of narrative risk communication on individual behavior, ours uses graphic images with few words (as opposed to paragraphs of text) as treatment conditions and the control condition. In addition, it measures the effects of different victim characters that represent a range from micro (individual) to macro (broader community): yourself, your circle, your community. The motivation for this research is to gain further precision regarding the power of narratives in communicating COVID-19 risk reduction by testing the effects of character selection on affective response and intent to get the COVID-19 vaccination. The context of the COVID-19 vaccines is important because vaccine refusal and hesitancy are enduring issues with a long history of risk communication challenges for public health agencies.
Research Questions
1. Do narrative messages have greater impact than non-narrative messages on intention to get the COVID-19 vaccination (for self and others)?
2. Does affective response (valance toward risk message) mediate the relationship between messages (narrative and non-narrative) and intention to get the COVID-19 vaccination (for self and others)?
3. Do a priori beliefs on individualism/collectivism moderate the relationship between messages (narrative and non-narrative) and affect?
4. Do a priori beliefs on individualism/collectivism moderate the relationship between messages (narrative and non-narrative) and intention to get the COVID-19 vaccination (for self and others)?
Hypotheses
H1: Respondents receiving the narrative treatment of yourself, your circle, and your community as victim will report a greater affective response on average than in the control condition.
H2: Respondents reporting a strong negative affective response will report lower intentions to vaccine (self and others), while respondents reporting a strong positive affective response will report higher intentions to vaccine (self and others).
H3: Respondents receiving the narrative treatment of yourself, your circle, and your community as victim will report a greater intention to vaccine (self and others) on average than in the control condition.
H4: Respondents with strong prior beliefs of individualism/collectivism will condition the strength of the relationships, based on congruency of beliefs and treatment received.
Estimation Strategy
We will estimate the effects of treatments using two- and one-tailed tests in the hypothesized direction of the effects. We will use Hayes’ PROCESS model in SPSS (Hayes, 2017) to conduct a conditional process analysis (Fig 1). Our objective is to test the mediating effects of affective response to the treatment and control conditions and the moderating effects of a priori beliefs of individualism/collectivism in shaping the impact of narrative risk messages on intention to vaccine (self and others). Therefore, we control for indirect effects in our model. This is a regression-based moderated mediation model, the regression coefficients of which can be estimated using two OLS regression analyses.

Authors

• Elizabeth Ann Shanahan, Montana State University

• Elizabeth Ann Albright, Duke University

• Deserai Anderson Crow, University of Colorado Denver

• Elizabeth Koebele, University of Nevada, Reno

• Kristin Taylor, Wayne State University