Republicans,Not Democrats,Are More Likely to Endorse Anti-Vaccine Misinformation

Abstract

Vaccine safety skeptics are often thought to be more likely to self-identify as Democrats (vs. Independents or Republicans). Recent studies, however, suggest that childhood vaccine misinformation is either more common among Republicans, or is uninfluenced by partisan identification (PID). Uncertainty about the partisan underpinnings of vaccine misinformation acceptance is important, as it could complicate efforts to pursue pro-vaccine health policies. I theorize that Republicans should be more likely to endorse anti-vaccine misinformation, as they tend to express more-negative views toward scientific experts. Across six demographically and nationally representative surveys, I find that—while few Americans think that “anti-vaxxers” are more likely to be Republicans than Democrats—Republican PID is significantly associated with the belief that childhood vaccines can cause autism. Consistent with theoretical expectations, effect is strongly mediated by anti-expert attitudes—an effect which supplemental panel analyses suggest is unlikely to be reverse causal.

Keywords

vaccine skepticism misinformation political psychology partisanship public opinion

The idea that self-identified Democrats in the US adult population are primarily to blame for the popularity of childhood vaccine misinformation (i.e., that childhood vaccines like MMR can cause autism) has become a mainstream argument in contemporary discourse about vaccine safety. Journalistic outlets, for example, have noted that vaccine refusal tends to be more common in traditionally liberal (“blue”) versus traditionally conservative (“red”) states; potentially indicative of higher levels of vaccine skepticism and misinformation endorsement in those areas (Berezow, 2014). This view is supported by academic research identifying liberal metropolitan areas as “hot spots” for vaccine exemptions and refusal (Olive et al., 2018).

Additionally, some journalists have portrayed those who view childhood vaccines as unsafe—and/or those who avoid vaccinating their children—as so-called “Whole Foods Moms” (e.g., Lubrano, 2019); that is, middle-income women who shop at higher-end grocery stores like Whole Foods, which tend to sell a variety of popular homeopathic and natural remedies to common medical ailments (Stone, 2019). Although these pieces do not always expressly typecast “Whole Foods Moms” as Democrats, the Whole Foods brand itself has come to take on a decisively liberal reputation in the US. Election analysts, for example, often note sharp differences in voting behavior between areas that have Whole Foods stores versus those that do not (e.g., Collins, 2018; Fearnow, 2019), which has been referred to be some election commentators as the “Whole Foods Bubble” (Wasserman, 2020).

Still others (e.g., Bricker & Justice, 2019) note anecdotally that prominent anti-vaccine activist Robert Kennedy Jr. (Mnookin, 2017), has longstanding dies to the Democratic party and has endorsed several Democratic presidential candidates (e.g., Welch, 2003). Some celebrities who question vaccine safety (e.g., Bricker & Justice, 2019) have also made public statements indicating that they hold politically liberal views (Furdyk, 2017).

Although the anecdotal link between identification with the Democratic party and vaccine misinformation has found a footing in popular discourse, academic research on the issue is considerably more mixed. In a synthesis of recent research on anti-vaccine opinion in the US, Bricker and Justice (2019) point to public opinion data suggesting that Democrats are more likely than Republicans to view childhood vaccines as unsafe. Although these differences are substantively small and statistically not differentiable from one another (Funk et al., 2015), they are consistent with research suggesting that parents who self-identify as liberal tend to be more willing to be vaccine hesitant; that is, pursue or consider pursuing vaccination schedules that do not reflect CDC guidelines (see: Callaghan et al., 2019 for more on the conceptualization and operationalization of vaccine hesitancy).

Others, however, have found that ideological conservatives—who tend to be much more likely to self-identify as Republicans compared to Democrats (Bafumi & Shapiro, 2009)—are more likely to endorse misinformation about childhood vaccine safety (e.g., Joslyn & Sylvester, 2019) and conspiracy theories about vaccines more generally (Featherstone et al, 2019; see also Hornsey et al., 2018 for a cross-national demonstration of this point). Still others find that political ideology is uncorrelated with vaccine skepticism (e.g., Rutjens et al., 2017). Conceptually, and more generally, elevated levels of pathogen sensitivity on the ideological right (e.g., O’Shea et al., 2021) may even lead Republicans expressing higher levels of confidence in vaccine safety; that is, as a means by which to avoid becoming sick with infectious diseases.

Past research therefore presents both empirical and conceptual ambiguity with respect to how partisan identification might influence anti-vaccine attitudes. However, because past studies differ with respect to how they measure, model, and conceptualize anti-vaccine opinion—as well as with survey sample composition and sampling strategy, it is difficult to know whether the inconsistent pattern of effects observed across studies is due to truly null relationship between partisan identification (PID) and anti-vaccine attitudes, or to differences in how each study was conducted. This implies that our understanding of how PID might shape anti-vaccine attitudes is at best conflicted, and, at worst, incomplete.

An incomplete understanding of how PID might impact childhood vaccine misinformation acceptance could have important health communication and policy consequences. Understanding the social, psychological, and political reasons people reject scientific consensus can help inform effective communication interventions aimed at reducing misinformation about childhood vaccine safety (e.g., Lunz Trujilo et al., 2020; see also: Lupia, 2015), and thereby increasing vaccine uptake in line with CDC recommended scheduling (Brewer et al., 2018). Moreover, because childhood vaccine skepticism is correlated with anti-vaccination behavior and opposition to policies that encourage universal vaccination (Benecke & DeYoung, 2019; Brewer et al., 2018; Motta et al., 2018), an accurate understanding of public opinion about vaccine skepticism may help policymakers better preempt potential political challenges to pro-vaccine policies and vaccine uptake. In this paper, I remedy this shortcoming by providing a novel and systematic investigation of the effect of PID on skepticism about the safety of childhood vaccines; specifically, misinformation regarding the possibility that they can cause children to develop autism. I theorize that self-identified Republicans should be more likely than Democrats or Independents to hold anti-vaccine views, due to their comparatively high levels of distrust in scientific research and the scientific community more broadly.

Across four demographically representative and two nationally representative cross-sectional surveys, I find that—while few Americans expect that Republicans are more likely than Democrats or Independents to hold anti-vaccine views—self-identified Republicans are consistently more likely to believe that childhood vaccines are unsafe ( $M_{e x p^{β}}$ = 1.84 [95% CI: 1.43, 2.25]). Consistent with theoretical expectations, these effects are strongly mediated (approximately 45%, on average) by the degree to which Republicans hold negative feelings toward the scientific community. This means that the effect of partisan identification may best be thought about—at least in part—as an expression of anti-expert attitudes. Supplemental panel analyses further suggest that the relationship between anti-expert attitudes and vaccine skepticism is unlikely to be reverse causal. I conclude by considering how these findings might help influence effective health policymaking and communication.

Conceptually Linking Political Partisanship and Anti-Vaccine Misinformation

Credibly connecting political partisanship to anti-vaccine views requires not only the standardization of measures and methods, but a critical investigation into why Republicans might be more likely than Democrats to believe that childhood vaccines are unsafe. I argue that the prevalence of anti-expert attitudes on the ideological right can help conceptually untangle past research on the link between partisan identification and vaccine safety.

A growing body of public opinion research suggests that Americans who embrace conservative ideological and Republican partisan labels are significantly more likely to hold negative views toward the scientific community and expertise more generally (Gauchat, 2012; Merkley, 2020; Motta, 2018; Oliver & Rahn, 2016). According to this line of work, increased reliance on scientific research when formulating public policy decisions has allowed scientific expertise itself to become grounds for political contestation (Gauchat, 2012, 2015). As Republican elites have staked out increasingly-clear opposition to federal government intervention in a wide range of policy areas—such as preferences for less regulation of energy production (see: McCright & Dunlap, 2010, 2011)—Republicans in the mass public have become increasing likely to hold positions inconsistent with scientific research (e.g., Brulle et al., 2012; Merkley & Stecula, 2020.), and to harbor negative feelings toward the scientific community (Gauchat, 2012, 2015; Motta, 2018).

Partisan polarization with respect to public attitudes toward the scientific community has also coincided with the rise of the Christian Right as a dominant force in the Republican Party, throughout the final quarter of the 20th century. The ascendency of the Christian Right has promoted deference to religious leaders in addition to and/or in place of scientific experts on culturally polarizing issues (Gauchat, 2012, 2015); such as the theory of evolution (Kahan, 2017), the risks posed by gene editing and nano-technology (Kahan et al., 2009), and concerns about whether or not immunizations transgress bodily sanctity and purity (Clifford & Wendell, 2016; Lunz Trujillo et al., 2020).

Relatedly, negative feelings toward the scientific community further reflect a more-recent rise in right-wing populist sentiment in the US (Oliver & Rahn, 2016). In addition to a more-general skepticism about the role that social and political elites play in American public life, the popularity of populist sentiments on the ideological right has included an “anti-intellectual” resentment of scientific experts, and a preference that policy decisions be informed by popular wisdom in place of scientific expertise (Merkley, 2020; Motta, 2018).

Because individuals who distrust scientific experts are more likely to reject claims informed by the best available scientific research (see: Motta, 2018), growing levels of distrust in experts on the right has facilitated partisan polarization with respect to the acceptance of anti-science conspiracy theories (e.g., Miller, 2020; Uscinski et al., 2020; see also Miller et al., 2016; van der Linden et al., 2021 for more on how the psychology of distrust enables conspiracy theory acceptance) and science misinformation more generally (e.g., Jost et al., 2018; Motta & Callaghan, 2020).

Correspondingly, people who hold negative views toward scientific experts have been shown to reject scientific consensus regarding the safety of childhood vaccines (Motta et al., 2018; Stecula, Kuru, Albaraccin, et al., 2020), and less likely to consider vaccinating themselves in adulthood (Callaghan et al., 2020; Motta, 2020; Stecula, Kuru, Jamieson, 2020; see also Brewer et al., 2018 for an extensive review on the link between negative views toward vaccination and decreased vaccine uptake). Consequently, the comparatively strong prevalence of anti-expert attitudes among self-identified Republicans offers a plausible theoretical mechanism connecting partisan identification to vaccine skepticism. Inasmuch as Republicans are less deferential to scientific expertise than Democrats or Independents, they may also be more likely to express skepticism about the safety of childhood vaccines.

Stated more formally, I expect that the effect of partisan identification on vaccine skepticism is mediated by views toward the scientific community. Specifically:

Hypothesis 1: Republicans are more likely than Democrats to believe that childhood vaccines are unsafe.

Hypothesis 2: The effect of Republican partisan identification on views that childhood vaccines are unsafe is at least partially mediated by negative attitudes toward the scientific community.

This means that, in addition to expecting that Republicans are more likely than Democrats to hold anti-vaccine views (H1), any potential link between Republican self-identification and vaccine skepticism might best be thought about as (at least in part) the result of distrust in scientific and medical experts (H2). In what follows, I devise an analytical strategy designed to put H1 and H2 to the empirical test.

Before doing so, however, it is important to recognize that the hypothesized link between partisanship, anti-expert attitudes, and vaccine misinformation represents just one potential pathway by which partisanship might influence attitudes toward vaccination. As suggested by the divergence between popular wisdom (i.e., that Democrats are more likely than Republicans to hold anti-vaccine views) academic literature on the partisan foundations of anti-vaccine views (i.e., that Republicans are either more likely, or neither more nor less likely, than Democrats to subscribe to anti-vaccine misinformation) it could alternatively be the case that Republicans are less likely than Democrats to hold anti-vaccine views.

Some have argued, for example, that that self-identified ideological conservatives (Hatemi & McDermott, 2012; Crawford, 2017) and individuals who hold politically conservative policy attitudes (Aarøe et al., 2020; O’Shea et al., 2021)—both of which are strongly correlated with Republican self-identification (Bafumi & Shapiro, 2009)—are more sensitive to pathogen-based threats. As a result, conservatives and Republicans may be comparatively more likely than Democrats to hold positive feelings toward vaccines, and to pursue vaccination; that is, in order to stave off the possibility of infection.

This is a plausible conceptual alternative. However, there at least two potential reasons why some scholars might raise doubts about this point of view. First, and most generally, while some scholars have noted political asymmetries in pathogen avoidance preferences, others argue that these differences are substantively small (e.g., Bakker et al., 2020; Kam & Estes, 2016; Tybur et al., 2010). Second, and more specifically, pathogen aversion has been shown to be correlated with moral purity values (Van Leeuwen et al., 2017), which include preferences for bodily sanctity (Haidt, 2012). As noted earlier, because vaccinations necessarily inject foreign substances into ones body, moral purity values have been shown to be correlated with aversion to vaccination (Clifford & Wendell, 2016; Lunz Trujillo et al., 2020). This may undermine a potential link between pathogen sensitivity on the ideological right and vaccine acceptance.

Still, if the analyses presented in this paper suggest were to suggest that Republicans are less likely than Democrats to hold anti-vaccine views and subscribe to anti-vaccine misinformation, this conceptual alternative (and others) would warrant additional consideration in future empirical testing. Presently, though, the relationship between partisan identification and vaccine attitudes is an open question; and one I put to the empirical test.

Analytical Strategy

In what follows, I offer a new and systematic assessment of the effect of partisan identification on anti-vaccine views. The analyses proceed in three steps. All data and code necessary to replicate this study are available at REPOSITORY BLINDED FOR REVIEW.

First, I assess whether or not popular narratives that downplay a potential link between Republican self-identification and vaccine skepticism might reflect public stereotypes about people who hold anti-vaccine views, via a survey question embedded in Study 4 (see: Data). Of course, self-reported stereotype acceptance does not necessarily imply receptivity to the (anecdotal) summary of journalistic coverage provided earlier. Still, it can at least help to document the prevalence of these narratives in the mass public.

Second, I calculate the average effect of partisan identification on anti-vaccine misinformation across studies. To do this, I first construct a series of logistic regression models that model acceptance of the view that childhood vaccines can cause autism on partisan identification and a series of demographic controls. Demographic controls include a variety of factors that past research suggests may influence vaccine misinformation acceptance, including respondents’ age, race, educational attainment, household income, and gender; that is, as men, more-educated individuals, people who self-identify as non-White, and older people have been shown to be more likely to hold skeptical views about childhood vaccine safety both in the US and cross-nationally (Callaghan et al., 2019; de Figueiredo et al., 2020). I also control for political ideology (i.e., self-placement on a standard left-right “symbolic ideology” scale; see Ellis & Stimson, 2012) which is highly correlated with partisan self-identification (Bafumi & Shapiro, 2009) in order to account for the possibility that observed partisan differences are the result of differences in ideological preferences.

Note that, because I theoretically expect a strong mediating relationship between anti-expert attitudes and PID, I estimate the meta-analytic results without anti-expert attitudes included in the models. This allows me to establish baseline effects of PID on vaccine skepticism.

Next, I extract parameter estimates and standard errors from each of these models, exponentiate them, and perform a formal meta-analysis of their effect size using a fixed effect modeling approach, via the metan command in Stata 15. This approach to summarizing the results is advantageous, because it allows me to weight the results from each study by statistical precision; meaning that more variable (and less certain) results are deemphasized in favor of less variable (more certain) results.

Third, I assess the degree to which anti-expert attitudes mediate the effect of PID on vaccine skepticism, via observational mediation analysis. In experimental contexts, mediation analyses assume that people are randomly assigned a “treatment” (in this case, PID); an assumption violated in observational mediation research (Green, Ha, & Bullock, 2010). However, observational analyses can provide valid mediation estimate so long as models are properly specified (Imai et al., 2011); that is, to account for other factors that could potentially cause people to (in this case) hold anti-vaccine opinions. Consequently, like the baseline models, all mediation models account for the demographic factors noted above (gender and racial self-identification, household income, political ideology, educational attainment, and age) that have been shown to be correlated with vaccine confidence attitudes.

Using Imai and colleagues’ mediate package in Stata 15, I specify models analogous to those in Step 2 (above)—with the addition of an anti-expert attitudes indicator—in order to estimate three quantities; (1) the indirect effect of PID explained through anti-expert attitudes, (2) the total effect of PID on vaccine skepticism, and (3) the percent of the total effect explained via anti-expert attitudes.

Before moving on, it is important to note that—while many surveys ask questions related to vaccine misinformation (and related topics, like support for pro-vaccine policies)—I selected these studies for two key reasons. First, all studies featured vaccine skepticism outcome variables that reference childhood vaccines misinformation specifically (i.e., as opposed to vaccines in general, or vaccines administered in adulthood), and in relation to the possibility that they might cause autism (perhaps the most well-known, well-studied, and socially/politically impactful aspect of vaccine misinformation; Benecke & DeYoung, 2019). Second, all studies included the same measure of anti-expert attitudes—used to conduct the mediation analyses—as well as PID measures and an analogous set of control variables.

Additionally, it is important to note that, while all models control for well-studied demographic correlates of vaccine confidence, they may not account for all factors that could potentially influence attitudes toward childhood vaccines. For example, parenting styles (see: Reich, 2014, 2018), and viewing participation in online anti-vaccine communities (Kata, 2010) as central to one’s self-concept—that is, as a form of social identification (Tajfel &Turner, 1978; see also: Huddy, 2001; Huddy et al., 2015) with the anti-vaccine community (see: Motta et al., 2021)—may conceptually play a role in shaping whether or not Americans believe that childhood vaccines are safe.

More generally, to achieve analytic consistency when comparing the meta-analytic and mediation models across studies, it is important to ensure that all models include the same sets of control variables, measured as similarly as possible. Unfortunately, this means that controlling for factors available in one set of studies (e.g., religious affiliation and religiosity measures in the ANES; see Data, below), but not available in another (i.e., the Lucid studies; again, see Data) would make it difficult to ascertain whether differences observed across studies are the result of true differences in vaccine attitudes, or merely the result of methodological artifact (i.e., modeling differences). Moreover, neither set of models can control for factors not measured in either of the two data sources; such as the aforementioned parenting style indicators, and/or social identification with the anti-vaccine community.

This means that, as is often the case in political and social science research (Clarke, 2009), both the meta-analytic and mediation models are unlikely to control for all potential correlates of anti-vaccine views (i.e., they may be subject to omitted variable bias). Consequently, while my analytic approach has the benefit of offering standardization across data sources, I urge caution when interpreting the results from these models. I encourage future work to expand on this approach by considering other factors, either conceptually related or unrelated to political partisanship, that might alternatively explain the patterns of effects observed in this study.

Additionally, some may be particularly concerned that the mediation models presented throughout this paper may be particularly sensitive to the influence of omitted variables that are correlated with both anti-expert attitude endorsement and vaccine misinformation acceptance. Consequently, I assess the robustness of the mediation analyses by providing series of sensitivity analyses, as recommended by Imai et al. (2011). Sensitivity analysis allows me to determine how much of an influence a potential unobserved confounder (or confounders) must have in order to render null the mediation effects presented later on in this piece. As I discuss later on, unobserved confounders must be fairly large in substantive size in order to explain away the mediating effect of anti-expert attitudes on vaccine beliefs.

Data

Data for this study come from two different sources. Studies 1 to 4 are cross-sectional surveys of approximately 1,000 people—conducted four times in April (N = 1,015), June (N = 1,015), August (N = 990), and October (N = 982), 2020. These studies employ quota sampling, via Lucid Theorem’s online opt-in sampling service, to achieve demographic representativeness on respondents’ age, gender, partisan identification, household income, educational attainment, race, and Census region. Lucid’s proprietary quota sampling procedure ensures marginal (but not necessarily joint) demographic representativeness on each of the above factors (Coppock & McClellan, 2019).

Although these data are not formally nationally representative of US adults (i.e., because they are not probability samples), data from Lucid have been shown to closely approximate known Census benchmarks, and tend to replicate both main and heterogeneous experimental effects from well-studied experimental paradigms (see: Coppock & McClellan, 2019). Lucid data have also been used extensively in the study of public health attitudes and misinformation (e.g., Callaghan et al., 2020; Miller, 2020), including skepticism about childhood vaccine safety (Lunz-Trujillo et al., 2020), and have been shown to produce political attitude estimates on factors like partisan identification and ideology that match nationally representative population benchmarks (Coppock & McClellan, 2019).

To further account for any remaining deviations between the sample and Census benchmarks, I calculate survey weights in each cross-sectional wave that adjust for respondents’ age, race, gender, income, and educational attainment (see: Motta & Callaghan, 2020 for an example of recent research published using the same weighting formula). Comparisons of these data to nationally representative benchmarks can be found in the Supplemental Materials. Full weighting syntax, including both the weighting formula and population totals used to calibrate the calculation of these weights, is available at REPOSITORY BLINDED FOR REVIEW.

Recognizing the limits of drawing inferences from data that are not formally nationally representative, Studies 5 to 6 feature data from the 2019 and 2016 ANES Pilot Studies, which interviewed N = 3,000 and N = 1,200 US adults (respectively) via YouGov. YouGov provides nationally representative data by first drawing a random sample of respondents in the nationally-representative American Community Survey, and then using propensity score matching techniques to find analogs in YouGov’s large online opt-in panel that match selected persons on a variety of demographic and political factors.

Measures

The primary outcome variable in this study is Vaccine Skepticism. In the Lucid studies (1–4), respondents were asked whether childhood vaccines “definitely” or “probably” can (vs. cannot) cause children to develop autism. In the 2019 ANES Pilot Study, respondents were asked to choose which of two statements “is most likely to be true” in a randomly assigned split ballot study. One ballot (ANES, 2019a, table 1) asked respondents to choose between statements suggesting that “childhood vaccines cause autism” and the other suggesting that they do not, while he second (ANES, 2019b, table 1) asked respondents whether “most scientific evidence” shows that childhood vaccines cause autism (vs. that they do not). Similarly, in the 2016 Pilot Study, respondents were asked “how likely or unlikely is it that vaccines cause autism?” and provided responses on a six point scale ranging from “extremely likely” to “extremely unlikely.”

Table 1.

Vaccine Misinformation Question Wording and Response Option Information Across Studies.

Study	Question wording	Coding rule	Misinfo. (%)
ANES (2016)	How likely or unlikely is it that vaccines cause autism? (1) Extremely likely; (2) Moderately likely; (3) Slightly likely; (4) Slightly unlikely; (5) Moderately unlikely (6) Extremely unlikely	1–3 = Misinformed	32
ANES (2016)		4–6 = Not misinformed	32
ANES (2019a)	Which of these two statements do you think is most likely to be true? (1) Childhood vaccines cause autism; (2) Childhood vaccines do not cause autism	1 = Misinformed	17
ANES (2019a)		2 = Not misinformed	17
ANES (2019b)	Which of these two statements do you think is most likely to be true? (1) Most scientific evidence shows childhood vaccines cause autism; (2) Most scientific evidence shows childhood vaccines do not cause autism	1 = Misinformed	14
ANES (2019b)		2 = Not misinformed	14
Lucid (2020)	Can vaccines administered to children at young ages cause them to become autistic? (1) They definitely can; (2) They probably can; (3) They probably cannot; (4) They definitely cannot	1–2 = Misinformed	4/20: 38
		3–4 = Not misinformed	6/20: 34
			8/20: 42
			10/20: 46

Note. Summary of differences in question wording and response options offered across studies. Note that information about how each measure was standardized into a dichotomous indicator of childhood vaccine misinformation acceptance is available in the column titled “Coding Rule.” The proportion of each sample holding anti-vaccine views, based on these coding rules, is summarized in the final column (with survey weights applied). Note also a high degree of consonance between the nationally representative 2016 ANES and the Lucid studies with respect to anti-vaccine misinformation endorsement (both of which were measured at the ordinal level), and somewhat lower levels of correspondence with the 2019 ANES data (which may be the result of the question offering a dichotomous choice response format).

As these questions vary in construction across surveys (see: Table 1), and because the 2019 ANES items are administered in a dichotomous format, I dichotomize responses in the Lucid and 2016 ANES surveys in order to facilitate comparison. Response option standardization across studies is important, as variables measured at different levels (e.g., dichotomous vs. ordinal), and with different numbers of ordered scale points, could imply that any differences I might observe across studies in the meta-analyses are the result of estimation (i.e., logistic vs. ordered logistic regression modeling) and/or differences in the amount of response-level variation on the outcome variables under investigation.

Consequently, in the Lucid surveys, I assigned respondents to receive a score of 1 if they think that vaccines definitely or probably can cause autism (0, otherwise). Similarly, participants in the 2019 ANES receive a score of 1 if they think that vaccines cause autism (and 0 if they do not). In the 2016 ANES Pilot, respondents receive a score of 1 if they indicated that it is extremely, moderately, or slightly likely (vs. unlikely) that vaccines cause autism. Table 1 summarizes differences in both question wording and response options across studies, as well as information about the coding rules used to produce estimates of anti-vaccine misinformation acceptance.

The primary independent variables in this analysis are Partisan Identification and Anti-Expert Attitudes. In all studies, I measure partisan identification using a standard branched seven-point indicator; rescaled to range from 0 (Strong Democrat) to 1 (Strong Republican). Additionally, I measure anti-expert attitudes—sometimes referred to as anti-intellectualism (e.g., Merkley, 2020; Motta, 2018)—using an item developed by Oliver and Rahn (2016) which asks respondents whether they trust “ordinary people or experts” more “when it comes to public policy decisions.” Responses were provided on a five point scale rescaled to range from 0 (trust experts much more) to 1 (trust ordinary people much more). Unfortunately, the expertise item was not available in the 2016 ANES Pilot. Consequently, I am unable to calculate mediation analyses in that study.

As noted earlier, models control for indicators of respondents’ Gender (whether or not respondents self-identify as women), Race (two indicators denoting whether respondents self-identify as Black [Non-Hispanic] or Hispanic), Age (indicators denoting whether respondents are aged 18–24, 25–44, 44–65, or 65+), Educational Attainment (whether respondents earned a college degree), Household Income (dichotomous income terciles denoting whether respondents live in homes making less than $25,000 per year, between $25 and 75,000 per year, or greater than $75,000 per year in the Lucid studies, and—due to minor differences in scale administration across studies—those making less than $30,000 per year, between $30 and 80,000 per year, and over $80,000 per year in the ANES studies); and Political Ideology (self-placement on a standard left-right symbolic ideological scale; recoded to range from 0 to 1 such that a score of 1 corresponds to identifying as extremely conservative).

Results

Before assessing the effects of PID on anti-vaccine misinformation acceptance, it is first important to demonstrate that many Americans accept conventional wisdom about who holds anti-vaccine beliefs. In Study 4 (Lucid, October 2020) I asked respondents whether “people who think that childhood vaccines can cause autism” are “more likely to be Democrats than Republicans”; “more likely to be Republicans than Democrats”; or “about equally likely to be Democrats or Republicans.”

I find that public opinion about who holds anti-vaccine views is consistent with the anecdotal stereotypes featured in popular press narratives. Most respondents (72%) think that people who hold negative views toward vaccines are more likely to be Democrats than Republicans (23%), or that there no partisan differences in childhood vaccine misinformation acceptance (48%). Just 28% believe that Republicans are more likely than Democrats to hold anti-vaccine views.

The meta-analytic results, in contrast, tell a very different story. The results are displayed in Figure 1. Point estimates from each multivariate logistic regression model are displayed as circles, with 95% confidence intervals extending from each one. Shaded squares correspond to the weight assigned to each result (listed on the right-hand side of the figure).

Figure 1.

Meta-analysis of the effect of PID on vaccine misinformation acceptance.

Results that are statistically different from 1.00 (i.e., a null effect; when expressed as an odds ratio) have 95% confidence intervals that do not intersect with the dashed red line. Likewise, results that are significantly different from the weighted average across studies produce confidence intervals that do not intersect with the solid black line. If my expectations are correct, most points should not intersect with the dashed red line—indicating a significant effect of Republican identification on vaccine skepticism—and should intersect with both the solid black line and with one another; indicating similar effect sizes and variability across studies.

Consistent with H1, the results suggest that Republican identification is positively and significantly associated with vaccine misinformation acceptance. Across studies, Republican identification is associated with a $e^{β}$ = 1.84 [95% CI: 1.43, 2.25] increase in the odds of believing that childhood vaccines can cause autism. In fact, Republican identification failed to produce a statistically significant increase in just one out of seven tests (ANES, 2019a; see the Supplemental Materials for full model output)—and, even in that case, was correctly signed.

The results also suggest both visually and statistically ( $χ_{6}^{2}$ = 7.64, p = n.s.) a high degree of similarity in effect sizes across studies. Heterogeneous study-level effects account for less than a quarter (22%) of the variation in pooled effect size. This is an important finding, as it suggests that both the nationally representative samples (ANES) and demographically representative quota samples (Lucid) tend to produce similar estimates of the effect of partisanship on anti-vaccine views.

Furthermore, Figure 1 summarizes the results of the observational mediation analyses described earlier. Consistent with H2, I find strong support for the idea that anti-expert attitudes mediate the effect of PID on vaccine skepticism. The indirect effects of PID (I) range from 0.03 to 0.08, with total effects of PID (T) ranging from 0.12 to 0.16; indicating that between 18% and 56% of the total effect of PID on vaccine skepticism is explained via anti-expert attitudes. On average, across all studies producing statistically significant parameter estimates, the mediation effect is approximately 45%.

Robustness Checks and Sensitivity Analyses

At this point, however, some might raise concerns about a potentially-endogenous relationship between anti-expert attitudes and vaccine skepticism. While I conceptualize anti-expert attitudes as a mediating factor (with respect to the effect of PID on vaccine skepticism), it could be the case that people hold negative views toward scientific experts because they view vaccines as unsafe.

In analyses available in the Supplementary Materials, I construct cross-lagged models of the effect of anti-expert attitudes on vaccine skepticism (and vice versa), using nationally representative panel data from Pew’s American Trends Panel. While these data did not meet criteria for inclusion in the meta-analyses (i.e., it measures only general vaccine skepticism; not misinformation about a link between childhood vaccines and the onset of autism), they nevertheless present an opportunity to offer a preliminary investigation into this question.

In the Supplemental Materials, I find that whereas lagged vaccine skepticism is not associated with change in negative feelings toward the scientific community ( $β = 0.00$ , p > .90; $N$ = 303), negative feelings toward the scientific community are associated with an increased likelihood to view vaccines as unsafe ( $β = 0.22$ , p < .05, $N = 313$ ). These preliminary findings suggest that the relationship between the hypothesized mediator in Figure 1 (anti-expert attitudes) does not appear to be subject to endogeneity concerns.

Additionally, and as alluded to in the paper’s Analytical Strategy, I provide a series of sensitivity analyses in the Supplemental Materials (Figure S2) that assess the robustness of each observed mediation effect across studies to an unobserved potential confounder. In the case of the four Lucid studies, I find that the effect of omitted variables (collectively represented as $ρ$ ; the correlation between error terms from the outcome and mediator stages of the mediation model) must be both negatively signed (i.e., having a negative influence on childhood vaccine misinformation acceptance) and moderately-large in substantive magnitude (i.e., producing a moderately-strong correlation between error terms; ranging from $ρ$ = −0.30 to −0.40) in order to explain away the mediation effects documented in Figure 1. In the 2019 ANES study, I find that $ρ$ must be positively signed (i.e., like anti-expert attitudes, having a positive influence on childhood vaccine misinformation acceptance) and again moderately-large in substantive size ( $ρ$ = 0.30, in both cases) in order to explain away the mediating influence of anti-expert attitudes on vaccine misinformation acceptance.

A more tractable way to assess sensitivity is to use the information presented above to determine the point at which unobserved confounders account for enough amount of the shared variance between the hypothesized mediator (in this case, anti-expert attitudes) and outcome variable (vaccine misinformation) to explain away a potential mediation effect (i.e., ${R_{M}^{2}}^{*}$ ${R_{Y}^{2}}^{*}$ ). Across all six studies, that quantity ranges from 0.09 to 0.16, meaning that an unobserved confound must account for 30% to 40% of the shared variance between anti-expert attitudes and anti-vaccine misinformation endorsement (calculated by taking the square root of ${R_{M}^{2}}^{*}$ ${R_{Y}^{2}}^{*}$ ; e.g., $\sqrt{0.09}$ = 0.30. See Rathbun et al. (2016) for a demonstration of this procedure). In other words, an unobserved confounder must have a substantial influence on the relationship between anti-expert attitudes and childhood vaccine misinformation endorsement in order to explain away the effects observed in Figure 1.

Taken together, these results suggest that, while my analytical approach can account for many (but certainly not all) factors that could alternatively explain the mediation effects observed in Figure 1, unobserved confounders would have to be fairly large in substantive magnitude in order to render mediation effects null. I am therefore confident that the mediating effects of anti-expert attitudes are robust to any potential model misspecification errors.

Conclusion

The results presented in this paper suggest that self-identified Republicans, not Democrats, are more likely to endorse misinformation about childhood vaccine safety. This effect may be, at least in part, the result of Republicans’ elevated distrust toward scientists and medical experts.

These findings have important implications for efforts to communicate vaccine safety, as well as for pro-vaccine public health policies. Concerning the former, the meta-analytic findings underscore the importance of targeting Republicans with pro-vaccine communications—that is, identifying the “correct” audience—and suggest potential pathways for crafting effective messages capable of doing so—for example, by enlisting prominent GOP figures to endorse pro-vaccine messages. Relatedly, the mediation analyses results further caution that, while most Americans defer to medical experts as a source of expertise about vaccine-related issues (Villa, 2019), messages designed for Republicans may be more effective when delivered by non-expert sources. Additionally, while this study investigates the anti-vaccination attitudes in the US, cross-national research has documented a relationship between right-leaning political ideology (which is strongly related, both conceptually and empirically, and Republican self-identification in the US) and anti-vaccination attitudes (see: Hornsey et al., 2018). Results from this work can help facilitate research on anti-vaccine attitudes outside the US; particularly in countries where right-wing political ideology, the parties that espouse those views, and feelings toward scientific experts have become (or are becoming) increasingly entangled.

Given the link between vaccine misinformation and support for pro-vaccine policies, these results also have important implications for health policy. Conceivably, policymakers who subscribe to conventional wisdom about public views on vaccine safety may be reluctant to introduce pro-vaccine policies in more-liberal parts of the country; such as densely populated urban areas. However, this research cautions that these reservations may be unfounded, or at least more likely to apply to more conservative geographic areas.

To get leverage on this question, future research might consider investigating how public opinion might shift in response localized changes in vaccine policy (and vice versa). More generally, researchers should investigate how partisan asymmetries in misinformation acceptance and other factors (e.g., individualism, support for limited government) shape perceptions about the role local and federal government actors ought to play in crafting vaccine-related policies.

It is also important to note that the research presented here is fairly general in scope, as it assesses the effects of partisan identification on vaccine misinformation acceptance across the entire US adult population. Efforts to study the complex interplay of other socio-political influences on childhood vaccine skepticism—and whether or not those influences are stronger for some population sub-groups relative to others—are worth exploring in future research. Experimental designs that manipulate either the presence/absence of partisan cues in the vaccine communication environment, for example, may be particularly helpful in assessing the intervening conditions under which partisanship may play more (vs. less) of a role in influencing vaccine misinformation endorsement. Future mediational studies—whether focusing on anti-expert attitudes or other potential mediators—ought to consider experimentally manipulating hypothesized mediators in order to further guard against the possibility of unobserved confounding effects.

Of course, this research is not without limitations. For example, while I am able to offer preliminary and supplemental analyses suggesting no reverse causation between the outcome variable (anti-vaccine misinformation) and the hypothesized mediator (anti-expert attitudes), the surveys analyzed in this the study’s meta-analysis and mediational tests are nevertheless limited by the fact that they are correlational in nature. Consequently, I cannot disentangle the causal ordering between partisan identification and the development of anti-expert attitudes. Future efforts to study the partisan dynamics of vaccine misinformation endorsement ought to do so in a longitudinal framework. Scholars should consider measuring anti-expert attitudes, partisan identification, and vaccine misinformation among the same survey panelists over time, in order to assess how change in each factor might correspond to change in the others.

Moreover, as noted throughout the manuscript, standardizing the comparison of the effects of partisanship across studies necessarily limits the extent to which I can control for potential confounds in my analyses. Additionally, data limitation restrictions prevent me from identifying, measuring, and accounting for all potential omitted variables that might alternatively explain partisan differences in anti-vaccine misinformation acceptance. While sensitivity analyses are helpful at addressing this concern, researchers in the future should make an effort to update and expand upon the results presented in this piece by accounting for factors not studied here; such as respondents’ social identification with anti-vaccine communities online, parenting styles, and media diets (e.g., the consumption of alternative medical news). Further probing potential confounds can help increase confidence in the effects documented here, and provide fertile grounds for future research.

Supplemental Material

sj-pdf-1-apr-10.1177_1532673X211022639 – Supplemental material for Republicans, Not Democrats, Are More Likely to Endorse Anti-Vaccine Misinformation

Supplemental material, sj-pdf-1-apr-10.1177_1532673X211022639 for Republicans, Not Democrats, Are More Likely to Endorse Anti-Vaccine Misinformation by Matthew Motta in American Politics Research

Footnotes

Declaration of Conflicting Interests

The author declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author received no financial support for the research, authorship, and/or publication of this article.

ORCID iD

Matthew Motta

Supplemental Material

Supplemental material for this article is available online.

Author Biography

Matthew Motta is an assistant professor of Political Science at Oklahoma State University. His research assesses the social-psychological origins and policy consequences of misinformation about science and politics. He is also interested in using insights from that research to devise strategic communication strategies that reduce misinformation endorsement.

References

Aarøe

Petersen

M. B.

Arceneaux

(2020). The behavioral immune system shapes partisan preferences in modern democracies: Disgust sensitivity predicts voting for socially conservative parties. Political Psychology, 41(6), 1073–1091.

Bafumi

Shapiro

R. Y.

(2009). A new partisan voter. The Journal of Politics, 71(1), 1–24.

Bakker

B. N.

Schumacher

Gothreau

Arceneaux

(2020). Conservatives and liberals have similar physiological responses to threats. Nature Human Behaviour, 4(6), 613–621.

Benecke

DeYoung

S. E.

(2019). Anti-vaccine decision-making and measles resurgence in the United States. Global Pediatric Health, 6, 2333794X19862949.

Berezow

A. B.

(2014). Are liberals or conservatives more anti-vaccine? Real Clear Science. https://www.realclearscience.com/journal_club/2014/10/20/are_liberals_or_conservatives_more_anti-vaccine_108905.html

Brewer

N. T.

Chapman

G. B.

Rothman

A. J.

Leask

Kempe

(2017). Increasing vaccination: Putting psychological science into action. Psychological Science in the Public Interest, 18(3), 149–207.

Bricker

Justice

(2019). The postmodern medical paradigm: A case study of anti-MMR vaccine arguments. Western Journal of Communication, 83(2), 172–189.

Brulle

R. J.

Carmichael

Jenkins

J. C.

(2012). Shifting public opinion on climate change: An empirical assessment of factors influencing concern over climate change in the US, 20022010. Climatic Change, 114(2), 169–188.

Callaghan

Motta

Sylvester

Trujillo

K. L.

Blackburn

C. C.

(2019). Parent psychology and the decision to delay childhood vaccination. Social Science & Medicine, 238, 112407.

10.

Clarke

K. A.

(2009). Return of the phantom menace: Omitted variable bias in political research. Conflict Management and Peace Science, 26(1), 46–66.

11.

Clifford

Wendell

D. G.

(2016). How disgust influences health purity attitudes. Political Behavior, 38(1), 155–178.

12.

Collins

(2018). Whole foods election: Majority of seats the democrats flipped had the trendy market nearby. USA Today. https://www.usatoday.com/story/news/politics/elections/2018/11/08/whole-foods-election-majority-seats-democrats-flipped-had-whole-foods-nearby/1930109002/

13.

Coppock

McClellan

O. A.

(2019). Validating the demographic, political, psychological, and experimental results obtained from a new source of online survey respondents. Research & Politics, 6(1), 2053168018822174.

14.

Crawford

J. T.

(2017). Are conservatives more sensitive to threat than liberals? It depends on how we define threat and conservatism. Social Cognition, 35(4), 354–373.

15.

de Figueiredo

Simas

Karafillakis

Paterson

Larson

H. J

. (2020). Mapping global trends in vaccine confidence and investigating barriers to vaccine uptake: A large-scale retrospective temporal modelling study. The Lancet, 396(10255), 898–908.

16.

Ellis

Stimson

J. A.

(2012). Ideology in America. Cambridge University Press.

17.

Fearnow

(2019). Democrats ‘becoming the whole foods party,’ says Iowa candidate who lost to Republican Steve King. Newsweek. https://www.newsweek.com/democrats-party-whole-foods-voters-1419755

18.

Featherstone

J. D.

Bell

R. A.

Ruiz

J. B.

(2019). Relationship of peoples sources of health information and political ideology with acceptance of conspiratorial beliefs about vaccines. Vaccine, 37(23), 2993–2997.

19.

Funk

Rainine

Page

(2015). Public and scientists views on ccience and society. Pew Research Center. https://www.pewresearch.org/science/2015/01/29/public-and-scientists-views-on-science-and-society/

20.

Furdyk

(2017). Big bang theory star Mayim Bialik, Madeleine Albright promise to join Trumps Muslim registry. Global News. https://globalnews.ca/news/3209639/big-bang-theory-star-mayim-bialik-madeleine-albright-promise-to-join-trumps-muslim-registry/

21.

Gauchat

(2012). Politicization of science in the public sphere: A study of public trust in the United States, 1974 to 2010. American Sociological Review, 77(2), 167–187.

22.

Gauchat

(2015). The political context of science in the United States: Public acceptance of evidence-based policy and science funding. Social Forces, 94(2), 723–746.

23.

Green

D. P.

S. E.

Bullock

J. G.

(2010). Enough already about “black box” experiments: Studying mediation is more difficult than most scholars suppose. The Annals of the American Academy of Political and Social Science, 628(1), 200–208.

24.

Haidt

(2012). The righteous mind: Why good people are divided by politics and religion. Vintage.

25.

Hatemi

McDermott

(2012). Policing the perimeter: Disgust and purity in democratic debate. PS: Political Science & Politics, 45(4), 675–687.

26.

Hornsey

M. J.

Harris

E. A.

Fielding

K. S.

(2018). The psychological roots of anti-vaccination attitudes: A 24-nation investigation. Health Psychology, 37(4), 307.

27.

Huddy

(2001). From social to political identity: A critical examination of social identity theory. Political Psychology, 22(1), 127–156.

28.

Huddy

Mason

Aare

(2015). Expressive partisanship: Campaign involvement, political emotion, and partisan identity. American Political Science Review, 109(1), 1–17.

29.

Imai

Keele

Tingley

Yamamoto

(2011). Unpacking the black box of causality: Learning about causal mechanisms from experimental and observational studies. American Political Science Review, 105(4), 765–789.

30.

Joslyn

M. R.

Sylvester

S. M.

(2019). The determinants and consequences of accurate beliefs about childhood vaccinations. American Politics Research, 47(3), 628–649.

31.

Jost

J. T.

van der Linden

Panagopoulos

Hardin

C. D.

(2018). Ideological asymmetries in conformity, desire for shared reality, and the spread of misinformation. Current Opinion in Psychology, 23, 77–83.

32.

Kahan

D. M.

(2017). Ordinary science intelligence: A science-comprehension measure for study of risk and science communication, with notes on evolution and climate change. Journal of Risk Research, 20(8), 995–1016.

33.

Kahan

D. M.

Braman

Slovic

Gastil

Cohen

(2009). Cultural cognition of the risks and benefits of nanotechnology. Nature Nanotechnology, 4(2), 87–90.

34.

Kam

C. D.

Estes

B. A.

(2016). Disgust sensitivity and public demand for protection. The Journal of Politics, 78(2), 481–496.

35.

Kata

(2010). A postmodern Pandora’s box: Anti-vaccination misinformation on the Internet. Vaccine, 28(7), 1709–1716.

36.

Lubrano

(2019). Anti-vaccine parents are often white, college-educated, Whole Foods moms. The Philadelphia Inquirer. https://www.inquirer.com/news/middle-class-working-class-vaccine-anti-vaxxers-measles-cdc-20190410.html

37.

Lunz Trujillo

Motta

Callaghan

Sylvester

. (2020). Correcting misperceptions about the MMR vaccine: Using psychological risk factors to inform targeted communication strategies. Political Research Quarterly, 74(2), 1065912920907695.

38.

Lupia

(2015). Uninformed: Why people know so little about politics and what we can do about it. Oxford University Press.

39.

McCright

A. M.

Dunlap

R. E.

(2010). Anti-reflexivity. Theory, Culture & Society, 27(2–3), 100–133.

40.

McCright

A. M.

Dunlap

R. E.

(2011). The politicization of climate change and polarization in the American public’s views of global warming, 2001–2010. The Sociological Quarterly, 52(2), 155–194.

41.

Merkley

Stecula

D. A.

(2020). Party cues in the news: Democratic elites, Republican backlash, and the dynamics of climate skepticism. British Journal of Political Science, 1–18.

42.

Merkley

(2020). Anti-intellectualism, populism, and motivated resistance to expert consensus. Public Opinion Quarterly, 84(1), 24–48.

43.

Miller

J. M.

(2020). Psychological, political, and situational factors combine to boost COVID-19 conspiracy theory beliefs. Canadian Journal of Political Science, 53(2), 327–334.

44.

Miller

J. M.

Saunders

K. L.

Farhart

C. E.

(2016). Conspiracy endorsement as motivated reasoning: The moderating roles of political knowledge and trust. American Journal of Political Science, 60(4), 824–844.

45.

Mnookin

(2017). How Robert F. Kennedy, Jr., distorted vaccine science. Scientific American. https://www.scientificamerican.com/article/how-robert-f-kennedy-jr-distorted-vaccine-science1/

46.

Motta

(2018). The dynamics and political implications of anti-intellectualism in the United States. American Politics Research, 46(3), 465–498.

47.

Motta

(2020). Could concern about climate change increase demand for a Lyme disease vaccine in the US?” Vaccine, 38(40), 6191–6193.

48.

Motta

Callaghan

(2020). The pervasiveness and policy consequences of medical folk wisdom in the US. Scientific Reports, 10(1), 1–10.

49.

Motta

Callaghan

Sylvester

(2018). Knowing less but presuming more: Dunning–Kruger effects and the endorsement of anti-vaccine policy attitudes. Social Science & Medicine, 211, 274–281.

50.

Motta

Callaghan

Sylvester

Lunz-Trujillo

(2021). Identifying the prevalence, correlates, and policy consequences of anti-vaccine social identity, Politics, Groups, and Identities.https://doi.org/10.1080/21565503.2021.1932528

51.

Olive

J. K.

Hotez

P. J.

Damania

Nolan

M. S.

(2018). The state of the antivaccine movement in the United States: A focused examination of nonmedical exemptions in states and counties. PLoS Medicine, 15(6), e1002578.

52.

Oliver

J. E.

Rahn

W. M.

(2016). Rise of the Trumpenvolk: Populism in the 2016 election. The ANNALS of the American Academy of Political and Social Science, 667(1), 189–206.

53.

O’Shea

Vitriol

J. A.

Federico

Appleby

Williams

(2021). Exposure and aversion to human transmissible diseases predict conservative ideological and partisan preferences. Forthcoming at Political Psychology, 1–24.

54.

Rathbun

B. C.

Kertzer

J. D.

Reifler

Goren

Scotto

T. J.

(2016). Taking foreign policy personally: Personal values and foreign policy attitudes. International Studies Quarterly, 60(1), 124–137.

55.

Reich

J. A.

(2014). Neoliberal mothering and vaccine refusal: Imagined gated communities and the privilege of choice. Gender & Society, 28(5), 679–704.

56.

Reich

J. A.

(2018). Calling the shots: Why parents reject vaccines. NYU Press.

57.

Rutjens

B. T.

Sutton

R. M.

van der Lee

(2017). Not all skepticism is equal: Exploring the ideological antecedents of science acceptance and rejection. Personality and Social Psychology Bulletin, 44(3), 384–405.

58.

Stecula

D. A.

Kuru

Albarracin

Jamieson

K. H.

(2020). Policy views and negative beliefs about vaccines in the United States, 2019. American Journal of Public Health, 110(10), 1561–1563.

59.

Stecula

D. A.

Kuru

Jamieson

K. H.

(2020). How trust in experts and media use affect acceptance of common anti-vaccination claims. Harvard Kennedy School Misinformation Review,.

60.

Stone

(2019). Whole Foods is selling dangerous anti-vaccine propaganda in its checkout aisles. Insider. https://www.insider.com/whole-foods-amazon-anti-vaxxer-vaccines-2019-12

61.

Tajfel

Turner

J. C.

(1978). Intergroup behavior. In Introducing Social Psychology, 401–466.

62.

Tybur

J. M.

Merriman

L. A.

Hooper

A. E. C.

McDonald

M. M.

Navarrete

C. D.

(2010). Extending the behavioral immune system to political psychology: Are political conservatism and disgust sensitivity really related? Evolutionary Psychology, 8(4), 147470491000800406.

63.

Uscinski

J. E.

Enders

A. M.

Klofstad

Seelig

Funchion

Everett

Wuchty

Premaratne

Murthi

(2020). Why do people believe COVID-19 conspiracy theories? Harvard Kennedy School Misinformation Review, 1(3).

64.

van der Linden

Panagopoulos

Azevedo

Jost

J. T

. (2021). The paranoid style in American politics revisited: An ideological asymmetry in conspiratorial thinking. Political Psychology, 42(1), 23–51.

65.

Van Leeuwen

Dukes

Tybur

J. M.

Park

J. H

. (2017). Disgust sensitivity relates to moral foundations independent of political ideology. Evolutionary Behavioral Sciences, 11(1), 92.

66.

Villa

(2019). Five facts about vaccines in the US. Pew Research Center. https://www.pewresearch.org/fact-tank/2019/03/19/5-facts-about-vaccines-in-the-u-s/

67.

Wasserman

(2020). To beat Trump, democrats may need to break out of the whole foods bubble. https://www.nytimes.com/interactive/2020/02/27/upshot/democrats-may-need-to-break-out-of-the-whole-foods-bubble.html

68.

Welch

(2003). RFK Jr. blasts Bush, champions Kerry. The Seattle Times. https://archive.seattletimes.com/archive/?date=20031029&slug=rfkjr29m

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.19 MB