Public attitudes toward COVID-19 vaccination: The role of vaccine attributes, incentives, and misinformation

While efficacious vaccines have been developed to inoculate against severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2; also known as COVID-19), public vaccine hesitancy could still undermine efforts to combat the pandemic. Employing a survey of 1096 adult Americans recruited via the Lucid platform, we examined the relationships between vaccine attributes, proposed policy interventions such as financial incentives, and misinformation on public vaccination preferences. Higher degrees of vaccine efficacy significantly increased individuals’ willingness to receive a COVID-19 vaccine, while a high incidence of minor side effects, a co-pay, and Emergency Use Authorization to fast-track the vaccine decreased willingness. The vaccine manufacturer had no influence on public willingness to vaccinate. We also found no evidence that belief in misinformation about COVID-19 treatments was positively associated with vaccine hesitancy. The findings have implications for public health strategies intending to increase levels of community vaccination.


The Correlates of Belief in Misinformation Concerning Covid-19 Treatments
Aside from assessing the influence of the vaccine attributes randomly manipulated in the conjoint experiment on US adults' willingness to vaccinate, a secondary objective of the study was to assess the association between individual respondents' levels of belief in misinformation about COVID-19 treatments and willingness to vaccinate.
As described in the text, we asked subjects to evaluate the veracity of eight statements about COVID-19 treatments, five of which were false and three of which were true. Correct answers were scored a -1; incorrect answers were scored 1; and responses of unsure/don't know were scored 0. The resulting index of belief in misinformation ranged from -8 to 8. The distribution of this index across our sample is presented in Supplementary Figure 2.
Our treatment knowledge question battery allowed us to construct two alternate measures of misinformation about COVID-19 treatments. The first is the simple additive count of false headlines about COVID-19 treatments that each subject believed. This distribution is shown in Supplementary Figure 3.
A second additive index measure again examines accuracy perceptions of the five false headlines, but assigns 1 point for each false headline believed, 0 points for an unsure response, and -1 points for a response that the headline is incorrect. The distribution of this measure is shown in Supplementary  Figure 4.
Supplementary Table 2 presents a series of regression analyses modeling the factors associated with beliefs in COVID-19 misinformation. While there is some variation across models depending on the operationalization of belief in misinformation, these analyses broadly find that political partisanship, educational attainment, gender, and age are significantly associated with belief in misinformation.

Robustness Check: Alternate Operationalization of Willingness to Vaccinate
As described in the text, after seeing each hypothetical vaccine profile in the conjoint experiment, subjects were asked whether they would take the vaccine or choose not be vaccinated. This binary operationalization of willingness to vaccinate is the dependent variable for the analyses in the text.
After answering this binary question, we also asked subjects how likely they would be to take each vaccine. This variable is measured on a seven-point scale from extremely unlikely to extremely likely. As a robustness check, we re-estimated the analyses in Table 3 using this alternate operationalization of willingness to vaccinate. As shown in Supplementary Table 3, results are substantively similar.

Robustness Check: Modeling the Association between Willingness to Vaccinate and Alternate Operationalizations of Belief in Misinformation
As shown in Supplementary Table 4, re-estimating the regression analyses in Table 3 using the alternate measures of belief in misinformation described above yields substantively similar results. Most important, subjects who were more susceptible to believing COVID-19 misinformation were more likely to report willingness to vaccinate, all else being equal.

Moderating Role of Attitudes toward General Vaccine Safety on EUA Treatment
The EUA treatment significantly decreased willingness to vaccinate, on average, by 7%. However, additional analyses interacting the EUA treatment with the measure of general vaccine safety beliefs show that this effect is concentrated among those who believe vaccines are generally safe. As shown in Supplementary Figure 5, among those who believe vaccines generally are extremely safe, the estimated negative effect of the EUA treatment is 11%.

Moderating Role of Misinformation on EUA Treatment
Belief in misinformation also significantly moderated the effect of the EUA treatment. As shown in Supplementary Figure 6, the negative effect of the EUA is concentrated among subjects who scored lower on the COVID-19 treatment misinformation index. This is consistent with speculation that subjects who scored high on this index may be more concerned about COVID-19 in general, which in turn makes them more willing to vaccinate and less concerned by the EUA process vs. full FDA approval.  Note: The first model is a negative binomial event count regression; models 2 and 3 are ordinary least squares regressions. Standard errors in parentheses. All significance tests are two-tailed. * p < .10 ** p < .05 *** p < .01 Supplementary