https://academic.oup.com/poq/article/84/1/24/5758079 We use cookies to enhance your experience on our website.By continuing to use our website, you are agreeing to our use of cookies. You can change your cookie settings at any time. Find out more Skip to Main Content Advertisement Oxford University Press * Search Menu * Account Menu * Menu * Sign In * Register Navbar Search Filter [This issue ] Mobile Microsite Search Term [ ] Search * Sign In + * Register * Issues * Advance articles * Submit + Author Guidelines + Submission Site + Open Access * Purchase * Alerts * About + About Public Opinion Quarterly + About the American Association for Public Opinion Research + Editorial Board + Advertising and Corporate Services + Journals Career Network + Self-Archiving Policy + Dispatch Dates + Contact Us Public Opinion Quarterly American Association for Public Opinion Research * Issues * Advance articles * Submit + Author Guidelines + Submission Site + Open Access * Purchase * Alerts * About + About Public Opinion Quarterly + About the American Association for Public Opinion Research + Editorial Board + Advertising and Corporate Services + Journals Career Network + Self-Archiving Policy + Dispatch Dates + Contact Us Close search filter [All Public Opinion Quarterly] search input [ ] Search Advanced Search Search Menu Skip Nav Destination Article Navigation Close mobile search navigation Article Navigation Issue Cover Volume 84 Issue 1 Spring 2020 Article Contents * Abstract * Anti-Intellectualism and the Rejection of Expert Consensus * Populist Rhetoric and the Priming of Anti-Intellectualism * Data and Methods for Observational Analysis * Observational Results * Experimental Design * Experimental Results * Discussion * Appendix * References * Footnotes * < Previous * Next > Article Navigation Anti-Intellectualism, Populism, and Motivated Resistance to Expert Consensus Eric Merkley Eric Merkley Address correspondence to Eric Merkley, University of Toronto, Munk School of Global Affairs and Public Policy, 1 Devonshire Place, Toronto, ON M5S 3K7, Canada; email: eric.merkley@utoronto.ca. OrcId icon https://orcid.org/0000-0001-7647-9650 Search for other works by this author on: Oxford Academic Google Scholar Public Opinion Quarterly, Volume 84, Issue 1, Spring 2020, Pages 24-48, https://doi.org/10.1093/poq/nfz053 Published: 26 February 2020 * pdfPDF * Split View * Views + Article contents + Figures & tables + Video + Audio + Supplementary Data * Cite Cite Eric Merkley, Anti-Intellectualism, Populism, and Motivated Resistance to Expert Consensus, Public Opinion Quarterly, Volume 84, Issue 1, Spring 2020, Pages 24-48, https://doi.org/10.1093/ poq/nfz053 Select Format [Select format ] Download citation Close * Permissions Icon Permissions * Share + Email + Twitter + Facebook + More Navbar Search Filter [This issue ] Mobile Microsite Search Term [ ] Search * Sign In + * Register Close search filter [All Public Opinion Quarterly] search input [ ] Search Advanced Search Search Menu Abstract Scholars have maintained that public attitudes often diverge from expert consensus due to ideology-driven motivated reasoning. However, this is not a sufficient explanation for less salient and politically charged questions. More attention needs to be given to anti-intellectualism--the generalized mistrust of intellectuals and experts. Using data from the General Social Survey and a survey of 3,600 Americans on Amazon Mechanical Turk, I provide evidence of a strong association between anti-intellectualism and opposition to scientific positions on climate change, nuclear power, GMOs, and water fluoridation, particularly for respondents with higher levels of political interest. Second, a survey experiment shows that anti-intellectualism moderates the acceptance of expert consensus cues such that respondents with high levels of anti-intellectualism actually increase their opposition to these positions in response. Third, evidence shows anti-intellectualism is connected to populism, a worldview that sees political conflict as primarily between ordinary citizens and a privileged societal elite. Exposure to randomly assigned populist rhetoric, even that which does not pertain to experts directly, primes anti-intellectual predispositions among respondents in the processing of expert consensus cues. These findings suggest that rising anti-elite rhetoric may make anti-intellectual sentiment more salient in information processing. Citizens often disagree with scientific opinion on a wide range of issues that have important implications for policymaking. The bulk of scholarly attention has been dedicated to climate change. At some level, this is understandable: Climate change is one of the most pressing issues of our time, and one where we have struggled to find and implement long-term policy solutions. However, a focus on climate change potentially distorts our understanding of citizens' acceptance of expert advice on other issues. A large majority of citizens in the United States agree with the climate change consensus, but opinion is heavily structured by ideology and partisanship. Thus, reasons for the failure of citizens to accept expert advice tend to focus on ideology-driven motivated reasoning (Kahan, Jenkins-Smith, and Braman 2011; Lewandowsky and Oberauer 2016) anchored in psychological theories of information processing (Kunda 1990; Ditto and Lopez 1992 ). But this is not obviously the case with other, less politicized issues (e.g., the safety of genetically modified organisms [GMOs], nuclear power, and water fluoridation), where experts and the public differ substantially.^1 One important predisposition that governs citizens' acceptance of expert knowledge is anti-intellectualism, a generalized mistrust of experts and intellectuals. Little work has explored the nature of this predisposition and how it may shape attitudes toward areas of expert consensus. This paper contributes to this nascent literature in three ways: (1) by establishing anti-intellectualism as a strong predictor of disagreement with positions of expert consensus on both highly salient and polarizing issues, like climate change, and other issues that score lower on these characteristics, such as water fluoridation; (2) by demonstrating that anti-intellectualism moderates the persuasiveness of messages of expert consensus on these same issues using a survey experiment; and (3) by connecting anti-intellectualism to the broader predisposition of populism, and showing that anti-elite rhetoric--even rhetoric that does not directly pertain to experts and intellectuals--can prime anti-intellectualism as a predisposition. Anti-Intellectualism and the Rejection of Expert Consensus Anti-intellectualism has a long history in American politics. In his foundational work, Hofstadter (1962, p. 19) argues this worldview is anchored in a belief that "intellectuals . . . are pretentious, conceited . . . and snobbish; and very likely immoral, dangerous, and subversive." Experts are seen as dangerous because they occupy the halls of power and profess to know how citizens should better run their lives. The rising importance of experts with the growth of government after World War II may have helped spark a rise in anti-intellectualism (Hofstadter 1962). Not all scholars agree on how to conceptualize anti-intellectualism. Rigney (1991) identifies three distinct components to anti-intellectualism: (1) anti-rationalism, or the dismissal of critical thinking as a desirable trait; (2) unreflexive instrumentalism, or the devaluing of long term payoffs for short-term material gain; and (3) anti-elitism, or the disparagement of intellectuals and experts. Some have viewed anti-intellectualism as a rhetorical style that emphasizes plainspokenness (Shogan 2007; Lim 2008), while others see it as an important component of populist rhetoric (Kazin 1995; Harris 2010; Brewer 2016). For this study, anti-intellectualism is defined as a generalized suspicion and mistrust of intellectuals and experts of whatever kind. Such contempt can have a number of sources. Some citizens might perceive expert authority as fundamentally at odds with religious authority that they may privilege. Or, they might not see the value of education and critical thought, particularly if they see it as coming at the expense of practical knowledge and common sense (Rigney 1991). Or, they may be skeptical of acquired knowledge because they see it as a tool of an exploitative societal elite (Brewer 2016)--a point which will be returned to below. Whatever the source, the result is a generalized mistrust of intellectuals and expert authority. Anti-intellectualism has important implications for citizens' acceptance of expert consensus where it exists. Perceived speaker knowledge is important for messages to be persuasive to the lay citizen, but they are not sufficient. Citizens need to trust speakers as well, which largely depends on the perception of common interests (Lupia and McCubbins 1998). By definition, those that hold anti-intellectual predispositions lack this trust in expert sources. Thus, they should exhibit lower levels of agreement with important positions of scientific consensus. Motta (2018) found this to be the case for climate change and the safety of nuclear power, but it should also apply to issues of lesser salience. Hypothesis 1 (H1): Anti-intellectualism is correlated with opposition to positions of expert consensus holding other factors constant. Science communication scholars have turned to consensus cues--clear statements that signal there is a scientific or expert consensus on a question--as a mechanism to persuade citizens on issues like climate change (van der Linden et al. 2015; van der Linden, Leiserowitz, and Maibach 2019). These works are part of a broader literature showing that corrections or "fact-checks" are often effective at rectifying misperceptions (Nyhan and Reifler 2010, 2015; Berinsky 2017; Wood and Porter 2019). The willingness to accept and be persuaded by consensus cues, however, should be highly conditional on one's level of anti-intellectual sentiment. Hypothesis 2 (H2): The effect of consensus cues on support for positions of expert consensus will be weaker among those with higher levels of anti-intellectual sentiment. Populist Rhetoric and the Priming of Anti-Intellectualism Anti-intellectual sentiment should shape how willing one is to be persuaded by expert consensus where it exists. Factors in the real world may intensify this reaction, and one possibility is populist rhetoric. The study of populism has been extensive, but there has been considerable disagreement on how to define it. Kazin (1995) argues that populism is a worldview that pits average citizens, imagined as a collective, against elites in political and economic conflict. Other scholars have treated populism as a rhetorical strategy that links populist movements globally (Rooduijn 2014). Following these scholars and others (Taggart 2000; Mudde 2004), I treat populism as both a worldview and a rhetorical strategy employed by politicians that emphasizes conflict between the people, imagined as a collective, and political elites or the establishment. In short, populism is minimally defined by its anti-elitism--a hostility toward elites, of whatever kind and for whatever reason. The roots of this anti-elitism can vary. On the political left, hostility toward elites is anchored in concern about the wealth and privilege of economic elites and the resulting effects on marginalized communities. On the political right, this suspicion may be fueled by a concern about excessive government power over individuals. In either case, ire is directed toward societal elites for reasons that are not linked to the level of intellect or education of those elites. Recently, populist sentiment has seen a resurgence in the aftermath of the Financial Crisis, manifesting in the rise of the Tea Party and Donald Trump on the right (Skocpol and Williamson 2013; Motta 2018) and the growing clout of liberal populists in the Democratic Party (Oliver and Rahn 2016). There is likely to be a strong connection between populism and anti-intellectualism. Suspicion of experts can be rooted in a perception that their knowledge will be used to control ordinary citizens, which shades into populist discourse. As Brewer notes, "American populism tends to be highly resentful of being told by experts 'we know best'" (2016, p. 253). Some populists may see experts as part of the ruling elite because of their status and importance in policy debates. In short, populist sentiment is likely a source of anti-intellectualism, though it is not likely to be the only one. Mistrust of intellectuals and experts can be rooted in, for example, religious fundamentalism, where intellectuals and experts are seen as a threat to religious authority. It can be fueled by ideology and partisanship, particularly in periods where intellectual and expert opinions crystalize on one side of the political divide. We can also imagine that anti-intellectualism has some connection to other personality characteristics, like a propensity for intuitionist over rationalist thinking (Oliver and Wood 2018). The important point is that these other possible sources of anti-intellectualism are conceptually distinct from populism. Populist sentiment is not necessary for anti-intellectualism. Similarly, anti-intellectualism is not an inevitable by-product of populism. The degree to which intellectuals and experts are identified as part of the ruling elite likely varies at the individual level and over time (Rigney 1991). For example, populist progressives in the early twentieth century saw expertise and professionalism as a solution to the machine politics they abhorred. Marxist leaders often make considerable use of anti-elite rhetoric, but their movement historically has often been led by intellectuals and fueled by important philosophical texts. Populism and anti-intellectualism have a complex relationship. They are connected to one another, but the latter should not be seen as a component of the former. The conceptual connection between populism and anti-intellectualism suggests that anti-elite rhetoric may have important implications for the public's support for positions with expert consensus. Rhetoric has the power to shape political attitudes (Petty and Cacioppo 1979; Chaiken 1980). It can also prime citizens to evaluate candidates and policy based on certain issues or underlying predispositions (Iyengar and Kinder 1987). For example, presidential rhetoric related to specific issues can shape the criteria by which he or she is evaluated by the public (Druckman and Holmes 2004), and racial rhetoric--at least that which is implicit--activates racial attitudes when citizens evaluate candidates and policies like social welfare ( Mendelberg 2001). Rhetoric can also link people's worldviews to issues of science. Scholars have found that moral rhetoric can activate citizens' moral intuitions in shaping attitudes toward environmental protection and science-related questions (Haider-Markel and Joslyn 2001; Barker 2005 ; Shen and Edwards 2005; Winterich, Zhang, and Mittal 2012; Feinberg and Willer 2013; Kidwell, Farmer, and Hardesty 2013; Clifford et al. 2015). For example, rhetoric on stem cell research that tapped into the care foundation--according to Haidt's moral foundations theory ( 2001)--was more persuasive among those that score high on that foundation (Clifford et al. 2015). An association between populism and anti-intellectualism would suggest that for many people, experts are seen as elites. If this is true, anti-elite rhetoric should prime anti-intellectualism as a predisposition that shapes people's propensity to accept expert consensus cues. More specifically, anti-elite rhetoric should further diminish the persuasive effect of consensus cues among those with higher levels of anti-intellectualism, even when that rhetoric does not directly pertain to experts and related issues. Hypothesis 3 (H3): The effect of consensus cues on support for positions of expert consensus will be weaker among those with higher levels of reported anti-intellectualism, particularly when exposed to anti-elite rhetoric. Data and Methods for Observational Analysis To test these hypotheses, I use the General Social Survey (GSS) and a non-probability sample of 3,614 voting-age American citizens collected July 15-17, 2018 through Amazon Mechanical Turk (MTurk). The latter sample cannot make claims to representativeness, but some of its broad characteristics are similar to the public as a whole. Table S1 of the online supplementary material compares the 2016 GSS and the MTurk sample used in this article. MTurk respondents are reasonably representative of the American population in terms of gender, race, partisanship, and ideology, but they are substantially younger, more educated, less religious, and more affluent. I use the GSS to check whether the observational findings in my MTurk sample hold for a more representative sample. MEASURING ANTI-INTELLECTUALISM Research on theorizing and measuring anti-intellectualism is scarce. One recent exception (Motta 2018) used a question in the GSS that asked respondents their degree of confidence in the scientific community (a great deal/only some/hardly any). I use this question for my GSS analyses as well, rescaled from 0 to 1, where 1 is having hardly any confidence in the scientific community. (Descriptive statistics of the variables analyzed appear in table S2 of the online supplementary material.) This single item is obviously insufficient on its own. Confidence is not the same concept as trust, while the scientific community only represents one set of actors in a broader constellation of experts and intellectuals in society. Oliver and Rahn (2016) measured anti-intellectualism with a three-question battery that they found correlates strongly with conservative ideology and religious fundamentalism. These three statements are: (1) I'd rather put my trust in the wisdom of ordinary people than the opinion of experts and intellectuals; (2) When it comes to really important questions, scientific facts don't help that much; and (3) Ordinary people can really use the help of experts to understand complicated things like science and health (reverse-coded). But their questions tap strongly into populist themes, while emphasizing attitudes toward science rather than intellectuals and experts more broadly. Anti-intellectualism is certainly related to these issues, but as discussed in the previous section, it is likely a more complex concept that is not fully captured by these questions. Absent theoretical work that teases out the dimensions of anti-intellectualism for measurement purposes, I lean on the conceptualization advanced here. Whatever the particular source of anti-intellectualism--religious fundamentalism, ideology, populism, or intuitionism--citizens with such sentiment will have a generalized mistrust of experts. So, respondents in this study were given a randomized battery of questions where they rated their trust in a number of different groups in society with the following lead ("distrust a lot" to "trust a lot," seven-point scale): "Below is a list of groups in society. Please tell us the degree to which you trust or distrust members of these groups." Among these groups were experts, scientists, economists, university professors, doctors and medical professionals, legal professionals, and financial experts. The intent was to provide respondents with a number of different professions where members engage in intellectual or academic pursuits to acquire established forms of expertise with which respondents would likely be familiar. These groups also provide a breadth of professions that are not exclusively related to the natural sciences. The distributions of these variables are displayed below in the left panel of figure 1 as box plots. Americans are generally trusting of experts across the board, but scientists and doctors have an edge over most groups with a median of 5 on the 0-6 scale, while legal professionals are trusted the least with a median of 4. Legal professionals aside, only one-quarter of respondents or less are distrusting--at any level--of any given expert community. Figure 1. Distributions of trust in expert communities, box plots (left). Distribution of anti-intellectualism, histogram (right). Open in new tabDownload slide Distributions of trust in expert communities, box plots (left). Distribution of anti-intellectualism, histogram (right). Figure 1. Distributions of trust in expert communities, box plots (left). Distribution of anti-intellectualism, histogram (right). Open in new tabDownload slide Distributions of trust in expert communities, box plots (left). Distribution of anti-intellectualism, histogram (right). I constructed an index of all of these groups, and rescaled it from 0 to 1, where 1 is the most anti-intellectual as indicated by a consistent and complete mistrust of various expert communities (Cronbach's alpha = 0.86; factor loadings and item drop scores provided in table S3 of the online supplementary material). The distribution of this measure is displayed in the right panel of figure 1. Anti-intellectualism as measured here is not all that common among respondents. The average score is approximately 0.34 on the 0-1 index. Further, approximately two-thirds of Americans find themselves between 0.17 and 0.51, indicating a reasonably narrow distribution. Only about 20 percent of respondents find themselves at the midpoint of the scale or higher. MODEL The GSS includes three questions related to expert consensus on climate change, nuclear power, and GMOs. The first item, asked in 1993, 1994, and 2000, used a four-point scale ranging from "definitely true" to "definitely not true," and asked respondents their opinion about the statement that "using coal or gas contributes to the greenhouse effect." The second question, posed in 1993, 1994, and 2010, asked respondents, "How dangerous is nuclear power for the environment?" Responses fell on a five-point scale, anchored by "extremely dangerous" and "not dangerous at all." Using this same scale, the third question (in 2000 and 2010) asked, "How dangerous is modifying genes in crops for the environment?" Nearly three-quarters (72 percent) of Americans believed coal and gas definitely or probably contributed to the greenhouse effect. Roughly four in five Americans (83 percent) viewed nuclear power as somewhat to extremely dangerous for the environment, while 72 percent thought the same for GMOs. These three issues represent important areas of expert consensus, but there are a few critical limitations with these questions. First, they have only been asked sporadically and fairly long ago (e.g., 1993, 1994, and 2000 for climate change). Second, the questions related to nuclear power and GMOs do not address beliefs about the safety of these technologies for humans, which is the central area of scientific consensus on these issues. Finally, while opposition to GMOs is historically less ideologically charged than climate change or nuclear power, it would be useful to have a question related to an issue that completely lacks salience. As a result, I had my MTurk respondents report their level of agreement with four positions of expert consensus related to climate change, nuclear power, GMOs, and water fluoridation. The first three statements featured different question wordings from the GSS using seven-point scales ranging from "strongly agree" to "strongly disagree": (1) Earth's climate is warming and this is due to the human production of greenhouse gases like carbon dioxide; (2) Nuclear power is a safe and environmentally friendly form of energy production compared to conventional sources of energy like fossil fuels; and (3) Genetically modified foods are safe, and pose no greater risk to human health than non-GM foods. The fourth statement, a seven-point scale anchored by "strongly agree" and "strongly disagree," read: Water fluoridation improves oral and dental health with no safety risk. Eighty percent of MTurk respondents agreed at some level with the expert position on climate change, compared to 48 percent on nuclear power, 46 percent on GMO safety, and 53 percent on water fluoridation. I rescaled these measures from 0 to 1, where 1 indicates full support for the expert position. The following OLS regression model examines the association between anti-intellectualism and support for each expert position, where X represents a vector of additional control variables:^2 support for expert position = a + b1anti-intellectualism+ X + e (1) I controlled for ideology (a seven-point scale ranging from "extremely liberal" to "extremely conservative") and partisanship (another seven-point scale anchored by "strong Democrat" and "strong Republican"). Both variables were rescaled from 0 to 1. Controls for education and political interest were also included. After all, it is possible that anti-intellectuals are simply not as informed about positions of expert consensus, and it is this lack of information, rather than the motivated rejection of expert consensus, that is driving any observed relationship. b [1] should be negative and significant to support H1.^3 Observational Results The results for the observational analyses testing H1 are displayed in figure 2. The top panel plots the coefficients for anti-intellectualism (operationalized as confidence in the scientific community) and ideology in the GSS. The full estimation results can be found in tables A1 and A2 of the Appendix. The results display a consistent link between anti-intellectualism and opposition to positions of expert consensus. Figure 2. Determinants of support for expert consensus in GSS (top) and MTurk sample (bottom). Controls include gender, employment status, race, age, income, education, church attendance, partisanship, generalized trust, and political interest (MTurk only); 95 and 90 percent confidence intervals. Open in new tabDownload slide Determinants of support for expert consensus in GSS (top) and MTurk sample (bottom). Controls include gender, employment status, race, age, income, education, church attendance, partisanship, generalized trust, and political interest (MTurk only); 95 and 90 percent confidence intervals. Figure 2. Determinants of support for expert consensus in GSS (top) and MTurk sample (bottom). Controls include gender, employment status, race, age, income, education, church attendance, partisanship, generalized trust, and political interest (MTurk only); 95 and 90 percent confidence intervals. Open in new tabDownload slide Determinants of support for expert consensus in GSS (top) and MTurk sample (bottom). Controls include gender, employment status, race, age, income, education, church attendance, partisanship, generalized trust, and political interest (MTurk only); 95 and 90 percent confidence intervals. In the GSS sample, moving from having a great deal of confidence in the scientific community to having no confidence is associated with a 0.05 drop in support for the scientific consensus on the greenhouse effect (p ~ 0.03, two-tailed here and throughout). It is also associated with a 0.05 (p ~ 0.02) and 0.08 reduction (p ~ 0.008) in the perceived safety of nuclear power and GMOs, respectively, on 0-1 scales. Ideology, in contrast, has inconsistent effects. Conservative ideology is negatively associated with support for the scientific consensus on the greenhouse effect (p ~ 0.03), while it is positively associated with the expert position on nuclear power (p ~ 0.07). Anti-intellectualism is a more consistent predictor of resistance to expert consensus than ideology in the GSS sample. Even stronger findings are found in the MTurk sample, the coefficients for which are displayed in the bottom panel of figure 2. The consistent strength of the association between anti-intellectualism and opposition to scientific consensus is striking. Moving across the anti-intellectualism index is associated with a reduction of 0.39 points in support for the scientific consensus for climate change (p ~ 0.003), 0.23 points for nuclear power (p < 0.001), 0.29 points for GMOs (p < 0.001), and 0.28 points for water fluoridation on 0-1 scales (p < 0.001). These are sizable effects. When averaging across all issues, movement across the range of the anti-intellectualism index is associated with a reduction in support for positions of expert consensus of 0.30 points (p < 0.001). ^4^,^5 Conservative ideology is associated with opposition to the climate consensus (p < 0.001), GMOs (p ~ 0.04), and fluoride (p ~ 0.02). Averaging across all issues, moving from extreme liberals to extreme conservatives is associated with a reduction in support for areas of expert consensus of a more modest 0.20 points (p < 0.001). Taken together, there is strong support in our observational analyses for H1. Anti-intellectualism is a strong predictor of support for expert consensus above and beyond the effects of left-right ideology. If, as this paper argues, anti-intellectualism motivates people to resist expert consensus where it exists, what else might we observe besides the above correlation? We know that elite discourse can have a polarizing effect on politically aware citizens (Zaller 1992). In this context, politically interested citizens are more likely to pick up on signals of expert consensus in discourse, but only those who are trusting of experts and intellectuals should accept them when forming their opinions. I ran the same estimations as above, but interacting the anti-intellectualism index with political interest. The linear predictions from these models are shown below in figure 3, and the estimates are found in table A3 of the Appendix. Political interest polarizes citizens in their support of positions of expert consensus across levels of anti-intellectualism. Of course, politically interested citizens may not be reacting to consensus cues per se. The experiment in the next section provides clearer causal evidence of this process.^6 Figure 3. Predicted support for positions of expert consensus across levels of political interest for respondents with low (10pctl) and high levels of anti-intellectualism (90pctl). A) Climate change; B) Nuclear power; C) GMOs; D) Fluoride; E) Average across issues. Confidence intervals: 90 percent. AI = Anti-intellectualism. Open in new tabDownload slide Predicted support for positions of expert consensus across levels of political interest for respondents with low (10pctl) and high levels of anti-intellectualism (90pctl). A) Climate change; B) Nuclear power; C) GMOs; D) Fluoride; E) Average across issues. Confidence intervals: 90 percent. AI = Anti-intellectualism. Figure 3. Predicted support for positions of expert consensus across levels of political interest for respondents with low (10pctl) and high levels of anti-intellectualism (90pctl). A) Climate change; B) Nuclear power; C) GMOs; D) Fluoride; E) Average across issues. Confidence intervals: 90 percent. AI = Anti-intellectualism. Open in new tabDownload slide Predicted support for positions of expert consensus across levels of political interest for respondents with low (10pctl) and high levels of anti-intellectualism (90pctl). A) Climate change; B) Nuclear power; C) GMOs; D) Fluoride; E) Average across issues. Confidence intervals: 90 percent. AI = Anti-intellectualism. Experimental Design There appears to be an observational connection between anti-intellectualism and opposition to areas of expert consensus. An experiment can more convincingly establish a causal link between these factors. I embedded a 3X2 factorial design in the MTurk survey for the purpose of testing H2 and H3. First, I randomly assigned subjects into three groups. The first treatment group was asked to read a mock news article from Reuters describing a political rally during which Senator Angus King (I-ME) used anti-elite rhetoric. The mock article was based on coverage of a real rally held by Donald Trump as a Republican presidential candidate. The language was altered so that it could plausibly come from either a Democratic or a Republican politician. Importantly, none of the rhetoric taps into anti-intellectual themes or addresses the scientific issues used in this paper. The discussion is focused on corruption on Wall Street and in Washington. This allows me to test whether generalized anti-elite rhetoric activates anti-intellectualism. Senator King's status as an independent senator from a small state allows me to examine the effect of rhetoric independent of the partisanship of the source. The articles can be found in figures S1-S4 in the online supplementary material. The second group received an identical article with one exception--the source of the rhetoric was Donald Trump, for respondents who identify or lean toward the Republican Party, or Bernie Sanders, for everyone else. Sometimes the effectiveness of a message is dependent on speaker characteristics (Kuklinski and Hurley 1994). Bernie Sanders and Donald Trump have reputations as populist, anti-establishment politicians and are both very popular within their respective political constituencies.^7 The third group, the control condition, instead read a short article related to a NASA discovery.^8 Second, I randomly assigned subjects into two groups independent of the rhetoric treatment. One group received the battery of questions asking their opinions on climate change, nuclear power, GMOs, and water fluoridation with the following lead: Surveys indicate that most scientists and policy experts agree with the following statements. To what extent do you agree or disagree with these positions? Subjects in the control group were simply asked to state their agreement or disagreement with each of those positions without such a lead. Respondents in the treatment condition should be more supportive of the expert consensus position, but in support of H2, this effect should be weaker among those with higher levels of anti-intellectual sentiment. Further, this moderation effect should be stronger when respondents are exposed to anti-elite rhetoric (H3). The combined experimental conditions are shown in table 1. Table 1. Experimental conditions and sample sizes . No consensus cue Consensus cue . Total . . No rhetoric 619 573 1,192 Rhetoric--Non-partisan 590 615 1,205 Rhetoric--Partisan 580 637 1,217 Total N = 1,789 N = 1,825 N = 3,614 . No consensus cue Consensus cue . Total . . No rhetoric 619 573 1,192 Rhetoric--Non-partisan 590 615 1,205 Rhetoric--Partisan 580 637 1,217 Total N = 1,789 N = 1,825 N = 3,614 Open in new tab Table 1. Experimental conditions and sample sizes . No consensus cue Consensus cue . Total . . No rhetoric 619 573 1,192 Rhetoric--Non-partisan 590 615 1,205 Rhetoric--Partisan 580 637 1,217 Total N = 1,789 N = 1,825 N = 3,614 . No consensus cue Consensus cue . Total . . No rhetoric 619 573 1,192 Rhetoric--Non-partisan 590 615 1,205 Rhetoric--Partisan 580 637 1,217 Total N = 1,789 N = 1,825 N = 3,614 Open in new tab The experimental protocol was as follows. Respondents consented to the survey and completed a number of pretreatment questions gauging their political attitudes and demographics. They were then asked to read the mock news article, which they believed to be real news content, and answered a battery of questions gauging their support for positions of expert consensus. Finally, they were debriefed on the nature of the deception in the experiment, given the opportunity to withdraw their consent, and provided a code to receive payment through Amazon.^9^,^10 MODELS H2 was tested using OLS regression with a model that includes an interaction of the scientific consensus cue and anti-intellectualism. However, confounders are a concern because moderating variables are observational (Kam and Trussler 2017). Thus, the treatment is also interacted with a series of controls (X), including ideology and partisanship, generalized trust, and political interest. Ideology, partisanship, and generalized trust may affect how respondents process consensus cues from experts and are all correlated with anti-intellectualism as conceptualized and measured here. We also have some expectation that more politically sophisticated respondents will be less responsive to experimental manipulations, so it is controlled for as well. b [3] should be negative and significant to provide support for H2: support for expert position = a + b1consensus cue + b2anti-intellectualism + b3 consensus cue * anti-intellectualism + X + consensus cue * X + e (2) Finally, the anti-elite rhetoric treatment should prime anti-intellectuals to resist messages of expert consensus. This requires a three-way interaction between both treatments and anti-intellectualism, as shown in equation 3: support for expert position = a + b1 consensus cue + b2 rhetoric + b3 anti-intellectualism + b4 consensus cue * rhetoric + b5 consensus cue * anti-intellectualism + b6 rhetoric * anti-intellectualism + b7 consensus cue * rhetoric * anti-intellectualism + X + consensus cue * X + e (3) Experimental Results The treatment was effective for nuclear power (p ~ 0.05), and for water fluoridation (p ~ 0.001), but the effects are substantively small--a 0.02 point increase in support of the expert position for the former and a 0.03 point increase for the latter. This is to be expected, as treatment effects for consensus cues are likely to be highly heterogeneous. It appears that anti-intellectualism consistently moderates the effectiveness of the consensus cue treatment, in line with H2. The regression estimates are provided in table A4 of the Appendix.^11 The interaction is significant on three of the four issues below the 0.05 level. The marginal effects of these estimates are shown in figure 4. Respondents that are the most trusting of experts are expected to increase their support for the climate consensus by a slight 0.02 points in response to the consensus cue, which is non-significant (p ~ 0.28). This decreases approximately 0.10 points for those with the highest levels of anti-intellectual sentiment such that the treatment will reduce the support of these respondents for the expert position by a sizable 0.08 points (p ~ 0.02). That is, anti-intellectuals double down on their rejection of expert positions in response to a consensus cue. This finding is similar to the "backfire effect" sometimes (but not often) found in fact-checking experiments where directionally motivated experimental subjects become more entrenched in their misperceptions in response to an intervention (Nyhan and Reifler 2010). Figure 4. Marginal effects of expert consensus cue conditioned by anti-intellectualism. A) Climate change; B) Nuclear power; C) GMOs; D) Fluoride; E) Average across issues. Controls for ideology, partisanship, generalized trust, and political interest. Confidence intervals: 90 percent. Open in new tabDownload slide Marginal effects of expert consensus cue conditioned by anti-intellectualism. A) Climate change; B) Nuclear power; C) GMOs; D) Fluoride; E) Average across issues. Controls for ideology, partisanship, generalized trust, and political interest. Confidence intervals: 90 percent. Figure 4. Marginal effects of expert consensus cue conditioned by anti-intellectualism. A) Climate change; B) Nuclear power; C) GMOs; D) Fluoride; E) Average across issues. Controls for ideology, partisanship, generalized trust, and political interest. Confidence intervals: 90 percent. Open in new tabDownload slide Marginal effects of expert consensus cue conditioned by anti-intellectualism. A) Climate change; B) Nuclear power; C) GMOs; D) Fluoride; E) Average across issues. Controls for ideology, partisanship, generalized trust, and political interest. Confidence intervals: 90 percent. The strong moderating effect of anti-intellectualism is common for all of the issues used here. Respondents with the lowest levels of anti-intellectualism are expected to increase their support for the expert position on nuclear power by 0.08 points (p ~ 0.001). This effect decreases 0.19 points for those with the highest levels of anti-intellectualism such that they also reduce their agreement with the expert position by 0.11 points (p ~ 0.02). The interaction term is highly significant (p ~ 0.004). The consensus cue also increases support for the expert position on fluoride by a substantial 0.10 points among those who are most trusting of experts (p < 0.001), but with a similar backfire effect of 0.10 points for those with the highest levels of anti-intellectual sentiment (p ~ 0.02). Again, the interaction term is highly significant (p ~ 0.002). The interaction for GMOs is not quite significant after including controls, but the results are pointing in the same direction (p ~ 0.11). Averaging across all issues, those most trusting of experts increase their support for positions of expert consensus by 0.06 points (p < 0.001), while we expect a backfire effect of 0.09 points among those who have the highest levels of anti-intellectual sentiment (p ~ 0.001).^12 The interaction term is highly significant (p < 0.001). All told, there is strong support for H2. Anti-intellectualism appears to have a consistent moderating effect on the acceptance of consensus cues from experts. Echoing the observational findings, those trusting of experts accept consensus information, while those who do not fail to do so. These effects are modest in size, but are precisely estimated because of the large sample used here. CAN ANTI-INTELLECTUALISM BE PRIMED BY POPULIST RHETORIC? We have strong theoretical grounds to expect an association between populism and anti-intellectualism. I constructed an index of populist sentiment in the GSS, based on institutional confidence questions, and in the MTurk survey with questions developed by Oliver and Rahn (2016).^13 There is a strong association between these two variables in both the 2016 GSS and the MTurk sample. These results are provided in figure S7 and table S10 in the online supplementary material. This finding is not particularly surprising given the anti-intellectual themes scholars have found in populist discourse. But it also means that anti-elite rhetoric may prime those with strong anti-intellectual sentiment to resist signals of consensus from expert communities. It may do this even when the rhetoric is not about experts per se. The results for this test are presented below. The three-way interaction is difficult to interpret, so marginal effects plots are shown below in figure 5. The estimates themselves are provided in table A5 of the Appendix. Figure 5. Marginal effects of expert consensus cue conditioned by anti-intellectualism and anti-elite rhetoric. A) Climate change; B) Nuclear power; C) GMOs; D) Fluoride; E) Average across issues. Confidence intervals: 90 percent. Open in new tabDownload slide Marginal effects of expert consensus cue conditioned by anti-intellectualism and anti-elite rhetoric. A) Climate change; B) Nuclear power; C) GMOs; D) Fluoride; E) Average across issues. Confidence intervals: 90 percent. Figure 5. Marginal effects of expert consensus cue conditioned by anti-intellectualism and anti-elite rhetoric. A) Climate change; B) Nuclear power; C) GMOs; D) Fluoride; E) Average across issues. Confidence intervals: 90 percent. Open in new tabDownload slide Marginal effects of expert consensus cue conditioned by anti-intellectualism and anti-elite rhetoric. A) Climate change; B) Nuclear power; C) GMOs; D) Fluoride; E) Average across issues. Confidence intervals: 90 percent. The results provide some evidence that the anti-elite rhetoric primed those that mistrusted experts to resist the cue on three of four issues. Anti-intellectualism moderated the effectiveness of the consensus cue for climate change, nuclear power, and GMOs, but only when exposed to anti-elite rhetoric.^14 These results provide compelling support for H3 and some troubling evidence that rising anti-elite rhetoric may undermine the persuasiveness of experts among the people that are in most need of persuading. Discussion Anti-intellectualism has been well documented in American political life. However, we do not have a strong understanding of what this concept is, which citizens are more attracted to it, and the implications of this predisposition for political behavior. This study offers three primary contributions. First, anti-intellectualism, the generalized mistrust and suspicion of intellectuals and experts, has relevance for political behavior in its own right, and not simply as a component of conservative ideology. Anti-intellectualism is a strong predictor of opposition to positions with widespread expert agreement, particularly for those with higher levels of political interest. This is true both on historically salient and highly polarized issues like climate change and nuclear power, as well as on issues of lesser salience and ideological charge, such as GMOs and water fluoridation (H1). Their consistency and strength as a predictor exceed that of left-right ideology. Second, experimental evidence shows that anti-intellectualism limits the persuasiveness of cues signaling expert consensus (H2). Scholars seeking to understand what influences citizens' acceptance of expert messages should place more attention on anti-intellectualism as a structuring predisposition. The finding that consensus cues have the potential to backfire on those who have the strongest levels of anti-intellectual sentiment is a troubling unintended consequence of these interventions. More work should be done to extend this finding to other science-based issues, and perhaps issues of consensus for communities of experts other than scientists and medical professionals. Third, populist rhetoric may play a role in priming anti-intellectualism as a predisposition when citizens process information in their environment, even when that rhetoric doesn't directly pertain to experts (H3). Anti-intellectualism limited the acceptance of cues signaling expert consensus much more strongly when respondents were exposed to generalized anti-elite rhetoric. The implication is that rising populist rhetoric may make anti-intellectualism a more salient determinant of support for expert positions in the future. One important limitation of this study is its use of a nonrepresentative MTurk sample. However, there is some evidence that my sample compares favorably to others used in similar research. Motta (2018), for example, used an online panel, weighted to match American population characteristics, to identify an association between anti-intellectualism and skepticism of climate science. He found that crossing the whole range of his measure of anti-intellectualism, drawn from the work of Oliver and Rahn (2016), increases the likelihood of being a climate skeptic by 17-21 percent depending on the survey wave. Using the same measure of anti-intellectualism, I found an increase of 16 percent in that same likelihood in my MTurk sample. More details on this analysis can be found in table S11 of the online supplementary material. The important role that anti-intellectualism plays in structuring attitudes toward expert consensus poses a challenge to science communicators who have focused on communication strategies to overcome left-right ideological conflict. The prospect that consensus cues may backfire on a sizable, ideologically heterogeneous segment of citizens is troubling. It is unlikely that appeals to authority in the form of expert consensus cues can ever be persuasive to citizens that by definition are hostile to these information sources. Scholars should explore alternative messaging strategies with the aim of persuading these citizens of the merits of mainstream expert positions. This problem is perhaps even more acute now amid populist resurgence in the United States and Europe (Moffitt 2016; Oliver and Rahn 2016). As populist rhetoric increasingly saturates political discourse, we can expect anti-intellectualism to be a stronger force in shaping public attitudes, including those related to areas of scientific and expert consensus. The politics of climate change may make it seem like anti-intellectualism is more prevalent in the United States, but misinformation on the safety of nuclear power and GMOs is strong in Europe as well. Future work should extend the study of anti-intellectualism beyond the United States. But before that happens, we need stronger theorization of anti-intellectualism as a complex concept with a number of distinct sources, like ideology, populism, religious fundamentalism, and intuitionism. We could then map out how each of these sources of anti-intellectualism affect opposition to positions of expert consensus directly and indirectly through their influence on anti-intellectualism. Much more work needs to be done in theorizing anti-intellectualism to situate this concept among related constructs and guide efforts at measurement so we can fully understand how it relates to mass behavior. ERIC MERKLEY is a postdoctoral fellow in the Munk School of Global Affairs and Public Policy at the University of Toronto, Toronto, ON, Canada. The author is grateful for the helpful feedback from his committee, Paul Quirk, Richard Johnston, and Fred Cutler, as well as his external examiner John Bullock. Thanks also go to Dominik Stecula, April Clark, Shane Singh, and participants at the Comparative-Canadian Workshop at UBC for useful comments. This work was supported by the Social Sciences and Humanities Research Council of Canada [752-2015-2504 to E.M.]. Appendix Table A1. Determinants of support for expert positions, GSS . Climate . Nuclear . GMO . . 1 . 2 . 3 . Confidence in science (Reverse-coded) -0.05* -0.05* -0.08** (0.02) (0.02) (0.03) Ideology -0.07* 0.06# 0.05 (0.03) (0.03) (0.04) PID -0.08** 0.09** 0.03 (0.02) (0.02) (0.03) Male 0.01 0.09** 0.11** (0.01) (0.01) (0.02) Employed -0.03# 0.01 0.01 (0.01) (0.01) (0.02) White 0.02 0.07** 0.04 (0.02) (0.02) (0.02) Age -0.00* 0.00** 0.00* (0.00) (0.00) (0.00) Income 0.00 0.01* 0.00 (0.00) (0.00) (0.00) Education -0.00 0.02** 0.01 (0.01) (0.01) (0.01) Church attendance -0.00 -0.00 0.00 (0.00) (0.00) (0.00) Trust 0.00 -0.03** -0.01 (0.01) (0.01) (0.01) Constant 0.75** -0.04 0.22** R^2 0.03 0.17 0.10 N 1,868 2,024 1,047 . Climate . Nuclear . GMO . . 1 . 2 . 3 . Confidence in science (Reverse-coded) -0.05* -0.05* -0.08** (0.02) (0.02) (0.03) Ideology -0.07* 0.06# 0.05 (0.03) (0.03) (0.04) PID -0.08** 0.09** 0.03 (0.02) (0.02) (0.03) Male 0.01 0.09** 0.11** (0.01) (0.01) (0.02) Employed -0.03# 0.01 0.01 (0.01) (0.01) (0.02) White 0.02 0.07** 0.04 (0.02) (0.02) (0.02) Age -0.00* 0.00** 0.00* (0.00) (0.00) (0.00) Income 0.00 0.01* 0.00 (0.00) (0.00) (0.00) Education -0.00 0.02** 0.01 (0.01) (0.01) (0.01) Church attendance -0.00 -0.00 0.00 (0.00) (0.00) (0.00) Trust 0.00 -0.03** -0.01 (0.01) (0.01) (0.01) Constant 0.75** -0.04 0.22** R^2 0.03 0.17 0.10 N 1,868 2,024 1,047 NOTE.--Sampling weight WTSSALL applied. Standard errors in parentheses. #p < 0.1, *p < 0.05, **p < 0.01 Open in new tab Table A1. Determinants of support for expert positions, GSS . Climate . Nuclear . GMO . . 1 . 2 . 3 . Confidence in science (Reverse-coded) -0.05* -0.05* -0.08** (0.02) (0.02) (0.03) Ideology -0.07* 0.06# 0.05 (0.03) (0.03) (0.04) PID -0.08** 0.09** 0.03 (0.02) (0.02) (0.03) Male 0.01 0.09** 0.11** (0.01) (0.01) (0.02) Employed -0.03# 0.01 0.01 (0.01) (0.01) (0.02) White 0.02 0.07** 0.04 (0.02) (0.02) (0.02) Age -0.00* 0.00** 0.00* (0.00) (0.00) (0.00) Income 0.00 0.01* 0.00 (0.00) (0.00) (0.00) Education -0.00 0.02** 0.01 (0.01) (0.01) (0.01) Church attendance -0.00 -0.00 0.00 (0.00) (0.00) (0.00) Trust 0.00 -0.03** -0.01 (0.01) (0.01) (0.01) Constant 0.75** -0.04 0.22** R^2 0.03 0.17 0.10 N 1,868 2,024 1,047 . Climate . Nuclear . GMO . . 1 . 2 . 3 . Confidence in science (Reverse-coded) -0.05* -0.05* -0.08** (0.02) (0.02) (0.03) Ideology -0.07* 0.06# 0.05 (0.03) (0.03) (0.04) PID -0.08** 0.09** 0.03 (0.02) (0.02) (0.03) Male 0.01 0.09** 0.11** (0.01) (0.01) (0.02) Employed -0.03# 0.01 0.01 (0.01) (0.01) (0.02) White 0.02 0.07** 0.04 (0.02) (0.02) (0.02) Age -0.00* 0.00** 0.00* (0.00) (0.00) (0.00) Income 0.00 0.01* 0.00 (0.00) (0.00) (0.00) Education -0.00 0.02** 0.01 (0.01) (0.01) (0.01) Church attendance -0.00 -0.00 0.00 (0.00) (0.00) (0.00) Trust 0.00 -0.03** -0.01 (0.01) (0.01) (0.01) Constant 0.75** -0.04 0.22** R^2 0.03 0.17 0.10 N 1,868 2,024 1,047 NOTE.--Sampling weight WTSSALL applied. Standard errors in parentheses. #p < 0.1, *p < 0.05, **p < 0.01 Open in new tab Table A2. Determinants of support for expert positions, MTurk . Climate Nuclear GMO . Fluoride Combined . . . . . 1 . 2 . 3 . 4 . 5 . Anti-intellectualism -0.39** -0.23** -0.29** -0.28** -0.30** (0.06) (0.08) (0.08) (0.08) (0.05) Ideology -0.43** -0.03 -0.17* -0.18* -0.20** (0.06) (0.08) (0.08) (0.08) (0.05) PID -0.02 0.06 0.10 0.05 0.05 (0.05) (0.06) (0.07) (0.06) (0.04) Male 0.01 0.12** 0.04 0.05* 0.06** (0.02) (0.03) (0.03) (0.03) (0.02) Employed 0.02 -0.01 0.00 0.01 0.01 (0.02) (0.03) (0.03) (0.03) (0.02) White 0.02 0.06* 0.02 0.02 0.03 (0.02) (0.03) (0.03) (0.03) (0.02) Age 0.00 -0.00# -0.01** -0.00 -0.00** (0.00) (0.00) (0.00) (0.00) (0.00) Income -0.00 0.01# 0.00 0.01 0.00 (0.01) (0.01) (0.01) (0.01) (0.01) Education -0.00 0.05** 0.06** 0.04** 0.04** (0.01) (0.02) (0.02) (0.01) (0.01) Church attendance 0.00 -0.00 -0.00 -0.00 -0.00 (0.00) (0.00) (0.01) (0.00) (0.00) Trust -0.01 0.04* 0.06** 0.03* 0.03** (0.01) (0.02) (0.02) (0.02) (0.01) Political Interest 0.02 0.03# 0.03 0.04* 0.03** (0.01) (0.02) (0.02) (0.02) (0.01) Constant 1.03** 0.23** 0.44** 0.39** 0.52** R^2 0.34 0.14 0.15 0.13 0.26 N 552 552 552 552 552 . Climate Nuclear GMO . Fluoride Combined . . . . . 1 . 2 . 3 . 4 . 5 . Anti-intellectualism -0.39** -0.23** -0.29** -0.28** -0.30** (0.06) (0.08) (0.08) (0.08) (0.05) Ideology -0.43** -0.03 -0.17* -0.18* -0.20** (0.06) (0.08) (0.08) (0.08) (0.05) PID -0.02 0.06 0.10 0.05 0.05 (0.05) (0.06) (0.07) (0.06) (0.04) Male 0.01 0.12** 0.04 0.05* 0.06** (0.02) (0.03) (0.03) (0.03) (0.02) Employed 0.02 -0.01 0.00 0.01 0.01 (0.02) (0.03) (0.03) (0.03) (0.02) White 0.02 0.06* 0.02 0.02 0.03 (0.02) (0.03) (0.03) (0.03) (0.02) Age 0.00 -0.00# -0.01** -0.00 -0.00** (0.00) (0.00) (0.00) (0.00) (0.00) Income -0.00 0.01# 0.00 0.01 0.00 (0.01) (0.01) (0.01) (0.01) (0.01) Education -0.00 0.05** 0.06** 0.04** 0.04** (0.01) (0.02) (0.02) (0.01) (0.01) Church attendance 0.00 -0.00 -0.00 -0.00 -0.00 (0.00) (0.00) (0.01) (0.00) (0.00) Trust -0.01 0.04* 0.06** 0.03* 0.03** (0.01) (0.02) (0.02) (0.02) (0.01) Political Interest 0.02 0.03# 0.03 0.04* 0.03** (0.01) (0.02) (0.02) (0.02) (0.01) Constant 1.03** 0.23** 0.44** 0.39** 0.52** R^2 0.34 0.14 0.15 0.13 0.26 N 552 552 552 552 552 Note.--Standard errors in parentheses. #p < 0.1, *p< 0.05, **p< 0.01 Open in new tab Table A2. Determinants of support for expert positions, MTurk . Climate Nuclear GMO . Fluoride Combined . . . . . 1 . 2 . 3 . 4 . 5 . Anti-intellectualism -0.39** -0.23** -0.29** -0.28** -0.30** (0.06) (0.08) (0.08) (0.08) (0.05) Ideology -0.43** -0.03 -0.17* -0.18* -0.20** (0.06) (0.08) (0.08) (0.08) (0.05) PID -0.02 0.06 0.10 0.05 0.05 (0.05) (0.06) (0.07) (0.06) (0.04) Male 0.01 0.12** 0.04 0.05* 0.06** (0.02) (0.03) (0.03) (0.03) (0.02) Employed 0.02 -0.01 0.00 0.01 0.01 (0.02) (0.03) (0.03) (0.03) (0.02) White 0.02 0.06* 0.02 0.02 0.03 (0.02) (0.03) (0.03) (0.03) (0.02) Age 0.00 -0.00# -0.01** -0.00 -0.00** (0.00) (0.00) (0.00) (0.00) (0.00) Income -0.00 0.01# 0.00 0.01 0.00 (0.01) (0.01) (0.01) (0.01) (0.01) Education -0.00 0.05** 0.06** 0.04** 0.04** (0.01) (0.02) (0.02) (0.01) (0.01) Church attendance 0.00 -0.00 -0.00 -0.00 -0.00 (0.00) (0.00) (0.01) (0.00) (0.00) Trust -0.01 0.04* 0.06** 0.03* 0.03** (0.01) (0.02) (0.02) (0.02) (0.01) Political Interest 0.02 0.03# 0.03 0.04* 0.03** (0.01) (0.02) (0.02) (0.02) (0.01) Constant 1.03** 0.23** 0.44** 0.39** 0.52** R^2 0.34 0.14 0.15 0.13 0.26 N 552 552 552 552 552 . Climate Nuclear GMO . Fluoride Combined . . . . . 1 . 2 . 3 . 4 . 5 . Anti-intellectualism -0.39** -0.23** -0.29** -0.28** -0.30** (0.06) (0.08) (0.08) (0.08) (0.05) Ideology -0.43** -0.03 -0.17* -0.18* -0.20** (0.06) (0.08) (0.08) (0.08) (0.05) PID -0.02 0.06 0.10 0.05 0.05 (0.05) (0.06) (0.07) (0.06) (0.04) Male 0.01 0.12** 0.04 0.05* 0.06** (0.02) (0.03) (0.03) (0.03) (0.02) Employed 0.02 -0.01 0.00 0.01 0.01 (0.02) (0.03) (0.03) (0.03) (0.02) White 0.02 0.06* 0.02 0.02 0.03 (0.02) (0.03) (0.03) (0.03) (0.02) Age 0.00 -0.00# -0.01** -0.00 -0.00** (0.00) (0.00) (0.00) (0.00) (0.00) Income -0.00 0.01# 0.00 0.01 0.00 (0.01) (0.01) (0.01) (0.01) (0.01) Education -0.00 0.05** 0.06** 0.04** 0.04** (0.01) (0.02) (0.02) (0.01) (0.01) Church attendance 0.00 -0.00 -0.00 -0.00 -0.00 (0.00) (0.00) (0.01) (0.00) (0.00) Trust -0.01 0.04* 0.06** 0.03* 0.03** (0.01) (0.02) (0.02) (0.02) (0.01) Political Interest 0.02 0.03# 0.03 0.04* 0.03** (0.01) (0.02) (0.02) (0.02) (0.01) Constant 1.03** 0.23** 0.44** 0.39** 0.52** R^2 0.34 0.14 0.15 0.13 0.26 N 552 552 552 552 552 Note.--Standard errors in parentheses. #p < 0.1, *p< 0.05, **p< 0.01 Open in new tab Table A3. Political interest, anti-intellectualism interaction, MTurk, all respondents . Climate Nuclear GMOs . Fluoride Combined . . . . Anti-intellectualism -0.15# 0.02 -0.09 -0.07 -0.08 (0.08) (0.10) (0.10) (0.09) (0.06) Political interest 0.04** 0.07** 0.06** 0.08** 0.06** (0.01) (0.02) (0.02) (0.02) (0.01) -0.11** -0.09* -0.08# -0.14** -0.11** Interest x Anti-intellectualism (0.03) (0.04) (0.04) (0.04) (0.03) Constant 1.03# 0.32** 0.49** 0.43** 0.57 Controls Yes Yes Yes Yes Yes R^2 0.37 0.09 0.13 0.12 0.23 N 3,187 3,187 3,187 3,187 3,187 . Climate Nuclear GMOs . Fluoride Combined . . . . Anti-intellectualism -0.15# 0.02 -0.09 -0.07 -0.08 (0.08) (0.10) (0.10) (0.09) (0.06) Political interest 0.04** 0.07** 0.06** 0.08** 0.06** (0.01) (0.02) (0.02) (0.02) (0.01) -0.11** -0.09* -0.08# -0.14** -0.11** Interest x Anti-intellectualism (0.03) (0.04) (0.04) (0.04) (0.03) Constant 1.03# 0.32** 0.49** 0.43** 0.57 Controls Yes Yes Yes Yes Yes R^2 0.37 0.09 0.13 0.12 0.23 N 3,187 3,187 3,187 3,187 3,187 Note.--Controls for ideology, partisanship, gender, employment, race, age, income, religiosity, and generalized trust. Standard errors in parentheses. #p < 0.1, *p < 0.05, **p < 0.01 Open in new tab Table A3. Political interest, anti-intellectualism interaction, MTurk, all respondents . Climate Nuclear GMOs . Fluoride Combined . . . . Anti-intellectualism -0.15# 0.02 -0.09 -0.07 -0.08 (0.08) (0.10) (0.10) (0.09) (0.06) Political interest 0.04** 0.07** 0.06** 0.08** 0.06** (0.01) (0.02) (0.02) (0.02) (0.01) -0.11** -0.09* -0.08# -0.14** -0.11** Interest x Anti-intellectualism (0.03) (0.04) (0.04) (0.04) (0.03) Constant 1.03# 0.32** 0.49** 0.43** 0.57 Controls Yes Yes Yes Yes Yes R^2 0.37 0.09 0.13 0.12 0.23 N 3,187 3,187 3,187 3,187 3,187 . Climate Nuclear GMOs . Fluoride Combined . . . . Anti-intellectualism -0.15# 0.02 -0.09 -0.07 -0.08 (0.08) (0.10) (0.10) (0.09) (0.06) Political interest 0.04** 0.07** 0.06** 0.08** 0.06** (0.01) (0.02) (0.02) (0.02) (0.01) -0.11** -0.09* -0.08# -0.14** -0.11** Interest x Anti-intellectualism (0.03) (0.04) (0.04) (0.04) (0.03) Constant 1.03# 0.32** 0.49** 0.43** 0.57 Controls Yes Yes Yes Yes Yes R^2 0.37 0.09 0.13 0.12 0.23 N 3,187 3,187 3,187 3,187 3,187 Note.--Controls for ideology, partisanship, gender, employment, race, age, income, religiosity, and generalized trust. Standard errors in parentheses. #p < 0.1, *p < 0.05, **p < 0.01 Open in new tab Table A4. Consensus cue treatment moderated by anti-intellectualism . Climate Nuclear GMO . Fluoride Combined . . . . . 1 . 2 . 3 . 4 . 5 . Anti-intellectualism -0.36** -0.09# -0.23** -0.29** -0.24** (0.04) (0.05) (0.05) (0.04) (0.03) Cue x -0.10* -0.19** -0.11 -0.20** -0.15** Anti-intellectualism (0.05) (0.07) (0.07) (0.06) (0.04) Ideology -0.32** -0.00 -0.15** -0.12** -0.15** (0.03) (0.04) (0.05) (0.04) (0.03) Cue x Ideology -0.03 0.01 -0.03 0.11* 0.02 (0.05) (0.06) (0.06) (0.06) (0.04) PID -0.13** 0.06# 0.05 0.02 0.00 (0.03) (0.04) (0.04) (0.03) (0.02) Cue x PID -0.04 -0.02 -0.02 -0.07 -0.04 (0.04) (0.05) (0.05) (0.05) (0.03) Trust 0.01 0.05** 0.05** 0.04** 0.04** (0.01) (0.01) (0.01) (0.01) (0.01) Cue x Trust -0.02# -0.03* -0.02 -0.01 -0.02* (0.01) (0.01) (0.01) (0.01) (0.01) Political Interest 0.01 0.05** 0.05** 0.05** 0.04** (0.01) (0.01) (0.01) (0.01) (0.01) Cue x Political -0.03* -0.00 -0.03 -0.03# -0.02* interest (0.01) (0.02) (0.02) (0.01) (0.01) Consensus cue 0.13** 0.12** 0.12* 0.15** 0.13** (0.04) (0.05) (0.05) (0.04) (0.03) Constant 1.02** 0.34** 0.44** 0.53** 0.59** R^2 0.37 0.04 0.07 0.10 0.19 N 3,213 3,213 3,213 3,213 3,213 . Climate Nuclear GMO . Fluoride Combined . . . . . 1 . 2 . 3 . 4 . 5 . Anti-intellectualism -0.36** -0.09# -0.23** -0.29** -0.24** (0.04) (0.05) (0.05) (0.04) (0.03) Cue x -0.10* -0.19** -0.11 -0.20** -0.15** Anti-intellectualism (0.05) (0.07) (0.07) (0.06) (0.04) Ideology -0.32** -0.00 -0.15** -0.12** -0.15** (0.03) (0.04) (0.05) (0.04) (0.03) Cue x Ideology -0.03 0.01 -0.03 0.11* 0.02 (0.05) (0.06) (0.06) (0.06) (0.04) PID -0.13** 0.06# 0.05 0.02 0.00 (0.03) (0.04) (0.04) (0.03) (0.02) Cue x PID -0.04 -0.02 -0.02 -0.07 -0.04 (0.04) (0.05) (0.05) (0.05) (0.03) Trust 0.01 0.05** 0.05** 0.04** 0.04** (0.01) (0.01) (0.01) (0.01) (0.01) Cue x Trust -0.02# -0.03* -0.02 -0.01 -0.02* (0.01) (0.01) (0.01) (0.01) (0.01) Political Interest 0.01 0.05** 0.05** 0.05** 0.04** (0.01) (0.01) (0.01) (0.01) (0.01) Cue x Political -0.03* -0.00 -0.03 -0.03# -0.02* interest (0.01) (0.02) (0.02) (0.01) (0.01) Consensus cue 0.13** 0.12** 0.12* 0.15** 0.13** (0.04) (0.05) (0.05) (0.04) (0.03) Constant 1.02** 0.34** 0.44** 0.53** 0.59** R^2 0.37 0.04 0.07 0.10 0.19 N 3,213 3,213 3,213 3,213 3,213 Note.--Standard errors in parentheses. #p < 0.1, *p < 0.05, **p < 0.01 Open in new tab Table A4. Consensus cue treatment moderated by anti-intellectualism . Climate Nuclear GMO . Fluoride Combined . . . . . 1 . 2 . 3 . 4 . 5 . Anti-intellectualism -0.36** -0.09# -0.23** -0.29** -0.24** (0.04) (0.05) (0.05) (0.04) (0.03) Cue x -0.10* -0.19** -0.11 -0.20** -0.15** Anti-intellectualism (0.05) (0.07) (0.07) (0.06) (0.04) Ideology -0.32** -0.00 -0.15** -0.12** -0.15** (0.03) (0.04) (0.05) (0.04) (0.03) Cue x Ideology -0.03 0.01 -0.03 0.11* 0.02 (0.05) (0.06) (0.06) (0.06) (0.04) PID -0.13** 0.06# 0.05 0.02 0.00 (0.03) (0.04) (0.04) (0.03) (0.02) Cue x PID -0.04 -0.02 -0.02 -0.07 -0.04 (0.04) (0.05) (0.05) (0.05) (0.03) Trust 0.01 0.05** 0.05** 0.04** 0.04** (0.01) (0.01) (0.01) (0.01) (0.01) Cue x Trust -0.02# -0.03* -0.02 -0.01 -0.02* (0.01) (0.01) (0.01) (0.01) (0.01) Political Interest 0.01 0.05** 0.05** 0.05** 0.04** (0.01) (0.01) (0.01) (0.01) (0.01) Cue x Political -0.03* -0.00 -0.03 -0.03# -0.02* interest (0.01) (0.02) (0.02) (0.01) (0.01) Consensus cue 0.13** 0.12** 0.12* 0.15** 0.13** (0.04) (0.05) (0.05) (0.04) (0.03) Constant 1.02** 0.34** 0.44** 0.53** 0.59** R^2 0.37 0.04 0.07 0.10 0.19 N 3,213 3,213 3,213 3,213 3,213 . Climate Nuclear GMO . Fluoride Combined . . . . . 1 . 2 . 3 . 4 . 5 . Anti-intellectualism -0.36** -0.09# -0.23** -0.29** -0.24** (0.04) (0.05) (0.05) (0.04) (0.03) Cue x -0.10* -0.19** -0.11 -0.20** -0.15** Anti-intellectualism (0.05) (0.07) (0.07) (0.06) (0.04) Ideology -0.32** -0.00 -0.15** -0.12** -0.15** (0.03) (0.04) (0.05) (0.04) (0.03) Cue x Ideology -0.03 0.01 -0.03 0.11* 0.02 (0.05) (0.06) (0.06) (0.06) (0.04) PID -0.13** 0.06# 0.05 0.02 0.00 (0.03) (0.04) (0.04) (0.03) (0.02) Cue x PID -0.04 -0.02 -0.02 -0.07 -0.04 (0.04) (0.05) (0.05) (0.05) (0.03) Trust 0.01 0.05** 0.05** 0.04** 0.04** (0.01) (0.01) (0.01) (0.01) (0.01) Cue x Trust -0.02# -0.03* -0.02 -0.01 -0.02* (0.01) (0.01) (0.01) (0.01) (0.01) Political Interest 0.01 0.05** 0.05** 0.05** 0.04** (0.01) (0.01) (0.01) (0.01) (0.01) Cue x Political -0.03* -0.00 -0.03 -0.03# -0.02* interest (0.01) (0.02) (0.02) (0.01) (0.01) Consensus cue 0.13** 0.12** 0.12* 0.15** 0.13** (0.04) (0.05) (0.05) (0.04) (0.03) Constant 1.02** 0.34** 0.44** 0.53** 0.59** R^2 0.37 0.04 0.07 0.10 0.19 N 3,213 3,213 3,213 3,213 3,213 Note.--Standard errors in parentheses. #p < 0.1, *p < 0.05, **p < 0.01 Open in new tab Table A5. Consensus cue treatment moderated by anti-intellectualism and rhetoric . Climate Nuclear GMO . Fluoride Combined . . . . . 1 . 2 . 3 . 4 . 5 . Consensus cue 0.13** 0.05 0.11# 0.16** 0.11** (0.04) (0.06) (0.06) (0.05) (0.04) Anti-intellectualism -0.36** -0.27** -0.36** -0.32** -0.33** (0.06) (0.08) (0.08) (0.07) (0.05) Cue x -0.04 -0.00 -0.02 -0.18# -0.06 Anti-intellectualism (0.09) (0.11) (0.12) (0.10) (0.07) Rhetoric -0.00 -0.11** -0.04 0.00 -0.04# (0.03) (0.03) (0.04) (0.03) (0.02) Cue x Rhetoric 0.01 0.11* 0.01 -0.01 0.03 (0.04) (0.05) (0.05) (0.05) (0.03) Rhetoric x 0.00 0.28** 0.20* 0.04 0.13* Anti-intellectualism (0.07) (0.09) (0.10) (0.09) (0.06) -0.10 -0.29* -0.15 -0.02 -0.14# Cue x Rhetoric x Anti-intellectualism (0.10) (0.13) (0.14) (0.13) (0.09) Constant 1.02** 0.41** 0.47** 0.53** 0.61** R^2 0.37 0.04 0.08 0.10 0.19 N 3,213 3,213 3,213 3,213 3,213 . Climate Nuclear GMO . Fluoride Combined . . . . . 1 . 2 . 3 . 4 . 5 . Consensus cue 0.13** 0.05 0.11# 0.16** 0.11** (0.04) (0.06) (0.06) (0.05) (0.04) Anti-intellectualism -0.36** -0.27** -0.36** -0.32** -0.33** (0.06) (0.08) (0.08) (0.07) (0.05) Cue x -0.04 -0.00 -0.02 -0.18# -0.06 Anti-intellectualism (0.09) (0.11) (0.12) (0.10) (0.07) Rhetoric -0.00 -0.11** -0.04 0.00 -0.04# (0.03) (0.03) (0.04) (0.03) (0.02) Cue x Rhetoric 0.01 0.11* 0.01 -0.01 0.03 (0.04) (0.05) (0.05) (0.05) (0.03) Rhetoric x 0.00 0.28** 0.20* 0.04 0.13* Anti-intellectualism (0.07) (0.09) (0.10) (0.09) (0.06) -0.10 -0.29* -0.15 -0.02 -0.14# Cue x Rhetoric x Anti-intellectualism (0.10) (0.13) (0.14) (0.13) (0.09) Constant 1.02** 0.41** 0.47** 0.53** 0.61** R^2 0.37 0.04 0.08 0.10 0.19 N 3,213 3,213 3,213 3,213 3,213 Note.--Controls for ideology, partisanship, generalized trust, and political interest interacted with the treatment. Standard errors in parentheses. #p < 0.1, *p < 0.05, **p < 0.01 Open in new tab Table A5. Consensus cue treatment moderated by anti-intellectualism and rhetoric . Climate Nuclear GMO . Fluoride Combined . . . . . 1 . 2 . 3 . 4 . 5 . Consensus cue 0.13** 0.05 0.11# 0.16** 0.11** (0.04) (0.06) (0.06) (0.05) (0.04) Anti-intellectualism -0.36** -0.27** -0.36** -0.32** -0.33** (0.06) (0.08) (0.08) (0.07) (0.05) Cue x -0.04 -0.00 -0.02 -0.18# -0.06 Anti-intellectualism (0.09) (0.11) (0.12) (0.10) (0.07) Rhetoric -0.00 -0.11** -0.04 0.00 -0.04# (0.03) (0.03) (0.04) (0.03) (0.02) Cue x Rhetoric 0.01 0.11* 0.01 -0.01 0.03 (0.04) (0.05) (0.05) (0.05) (0.03) Rhetoric x 0.00 0.28** 0.20* 0.04 0.13* Anti-intellectualism (0.07) (0.09) (0.10) (0.09) (0.06) -0.10 -0.29* -0.15 -0.02 -0.14# Cue x Rhetoric x Anti-intellectualism (0.10) (0.13) (0.14) (0.13) (0.09) Constant 1.02** 0.41** 0.47** 0.53** 0.61** R^2 0.37 0.04 0.08 0.10 0.19 N 3,213 3,213 3,213 3,213 3,213 . Climate Nuclear GMO . Fluoride Combined . . . . . 1 . 2 . 3 . 4 . 5 . Consensus cue 0.13** 0.05 0.11# 0.16** 0.11** (0.04) (0.06) (0.06) (0.05) (0.04) Anti-intellectualism -0.36** -0.27** -0.36** -0.32** -0.33** (0.06) (0.08) (0.08) (0.07) (0.05) Cue x -0.04 -0.00 -0.02 -0.18# -0.06 Anti-intellectualism (0.09) (0.11) (0.12) (0.10) (0.07) Rhetoric -0.00 -0.11** -0.04 0.00 -0.04# (0.03) (0.03) (0.04) (0.03) (0.02) Cue x Rhetoric 0.01 0.11* 0.01 -0.01 0.03 (0.04) (0.05) (0.05) (0.05) (0.03) Rhetoric x 0.00 0.28** 0.20* 0.04 0.13* Anti-intellectualism (0.07) (0.09) (0.10) (0.09) (0.06) -0.10 -0.29* -0.15 -0.02 -0.14# Cue x Rhetoric x Anti-intellectualism (0.10) (0.13) (0.14) (0.13) (0.09) Constant 1.02** 0.41** 0.47** 0.53** 0.61** R^2 0.37 0.04 0.08 0.10 0.19 N 3,213 3,213 3,213 3,213 3,213 Note.--Controls for ideology, partisanship, generalized trust, and political interest interacted with the treatment. Standard errors in parentheses. #p < 0.1, *p < 0.05, **p < 0.01 Open in new tab References Barker , David C . 2005 . "Values, Frames, and Persuasion in Presidential Nomination Campaigns." Political Behavior 27 : 375 - 94 . Google Scholar Crossref Search ADS WorldCat Berinsky , Adam J . 2017 . "Rumors and Health Care Reform: Experiments in Political Misinformation." British Journal of Political Science 47 : 241 - 62 . Google Scholar Crossref Search ADS WorldCat Brewer , Mark D . 2016 . "Populism in American Politics." The Forum 14 : 249 - 64 . Google Scholar Crossref Search ADS WorldCat Chaiken , Shelly . 1980 . "Heuristic versus Systematic Information Processing and the Use of Source versus Message Cues in Persuasion." Journal of Personality and Social Psychology 37 : 1387 - 97 . Google Scholar Crossref Search ADS WorldCat Clifford , Scott , Jennifer Jerit , Carlisle Rainey , and Matt Motyl . 2015 . "Moral Concerns and Policy Attitudes: Investigating the Influence of Elite Rhetoric." Political Communication 32 : 229 - 48 . Google Scholar Crossref Search ADS WorldCat Ditto , Peter H. , and David F. Lopez . 1992 . "Motivated Skepticism: Use of Differential Decision Criteria for Preferred and Nonpreferred Conclusions." Journal of Personality and Social Psychology 63 : 568 - 84 . Google Scholar Crossref Search ADS WorldCat Druckman , James N. , and Justin W. Holmes . 2004 . "Does Presidential Rhetoric Matter? Priming and Presidential Approval." Presidential Studies Quarterly 34 : 755 - 78 . Google Scholar Crossref Search ADS WorldCat Feinberg , Matthew , and Robb Willer . 2013 . "The Moral Roots of Environmental Attitudes." Psychological Science 24 : 56 - 62 . Google Scholar Crossref Search ADS PubMed WorldCat Green , Donald P. , Shang E. Ha , and John G. Bullock . 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 : 200 - 208 . Google Scholar Crossref Search ADS WorldCat Haider-Markel , Donald P. , and Mark R. Joslyn . 2001 . "Gun Policy, Opinion, Tragedy, and Blame Attribution: The Conditional Influence of Issue Frames." Journal of Politics 63 : 520 - 43 . Google Scholar Crossref Search ADS WorldCat Haidt , Jonathan . 2001 . "The Emotional Dog and Its Rational Tail: A Social Intuitionist Approach to Moral Judgment." Psychological Review 108 : 814 - 34 . Google Scholar Crossref Search ADS PubMed WorldCat Harris , Lee . 2010 . The Next American Civil War: The Populist Revolt Against the Liberal Elite. New York : Palgrave-Macmillan . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Hofstadter , Richard . 1962 . Anti-Intellectualism in American life . New York : Random House . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Iyengar , Shanto , and Donald R. Kinder . 1987 . News That Matters: Television and American Opinion. Chicago : University of Chicago Press . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Kahan , Dan M. , Hank Jenkins-Smith , and Donald Braman . 2011 . "Cultural Cognition of Scientific Consensus." Journal of Risk Research 14 : 147 - 74 . Google Scholar Crossref Search ADS WorldCat Kam , Cindy D. , and Marc J. Trussler . 2017 . "At the Nexus of Observational and Experimental Research: Theory, Specification, and Analysis of Experiments with Heterogeneous Treatment Effects." Political Behavior 39 : 789 - 815 . Google Scholar Crossref Search ADS WorldCat Kazin , Michael . 1995 . The Populist Persuasion . New York : Basic Books . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Kidwell , Blair , Adam Farmer , and David M. Hardesty . 2013 . "Getting Liberals and Conservatives to Go Green: Political Ideology and Congruent Appeals." Journal of Consumer Research 40 : 350 - 67 . Google Scholar Crossref Search ADS WorldCat Kuklinski , James H. , and Norman L. Hurley . 1994 . "On Hearing and Interpreting Political Messages: A Cautionary Tale of Citizen Cue-Taking." Journal of Politics 56 : 729 - 51 . Google Scholar Crossref Search ADS WorldCat Kunda , Ziva . 1990 . "The Case for Motivated Reasoning." Psychological Bulletin 108 : 480 - 98 . Google Scholar Crossref Search ADS PubMed WorldCat Lewandowsky , Stephan , and Klaus Oberauer . 2016 . "Motivated Rejection of Science." Current Directions in Psychological Science 25 : 217 - 22 . Google Scholar Crossref Search ADS WorldCat Lim , Elvin T . 2008 . The Anti-Intellectual Presidency: The Decline of Presidential Rhetoric from George Washington to George W. Bush. New York : Oxford University Press . Google Scholar Crossref Search ADS Google Preview WorldCat COPAC Lupia , Arthur , and Matthew D. McCubbins . 1998 . The Democratic Dilemma: Can Citizens Learn What They Need to Know? New York : Cambridge University Press . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Mendelberg , Tali . 2001 . The Race Card: Campaign Strategy, Implicit Messages, and the Norm of Equality. Princeton, NJ : Princeton University Press . Google Scholar Crossref Search ADS Google Preview WorldCat COPAC Moffitt , Benjamin . 2016 . The Global Rise of Populism: Performance, Political Style, and Representation. Stanford, CA : Stanford University Press . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Montgomery , Jacob M. , Brendan Nyhan , and Michelle Torres . 2018 . "How Conditioning on Posttreatment Variables Can Ruin Your Experiment and What to Do about It." American Journal of Political Science 62 : 760 - 75 . Google Scholar Crossref Search ADS WorldCat Motta , Matthew . 2018 . "The Dynamics and Political Implications of Anti-Intellectualism in the United States." American Politics Research 46 : 465 - 98 . Google Scholar Crossref Search ADS WorldCat Mudde , Cas . 2004 . "The Populist Zeitgeist." Government and Opposition 39 : 541 - 63 . Google Scholar Crossref Search ADS WorldCat Nyhan , Brendan , and Jason Reifler . 2010 . "When Corrections Fail: The Persistence of Political Misperceptions." Political Behavior 32 : 303 - 30 . Google Scholar Crossref Search ADS WorldCat ------. 2015 . "Displacing Misinformation about Events: An Experimental Test of Causal Corrections." Journal of Experimental Political Science 2 : 81 - 93 . Crossref Search ADS WorldCat Oliver , J. Eric , and Wendy M. Rahn . 2016 . "Rise of the Trumpenvolk: Populism in the 2016 Election." The ANNALS of the American Academy of Political and Social Science 667 : 189 - 206 . Google Scholar Crossref Search ADS WorldCat Oliver , J. Eric , and Thomas Wood . 2018 . Enchanted America: How Intuition and Reason Divide Our Politics. Chicago : University of Chicago Press . Google Scholar Crossref Search ADS Google Preview WorldCat COPAC Petty , Richard E. , and John T. Cacioppo . 1979 . "Issue Involvement Can Increase or Decrease Persuasion by Enhancing Message-Relevant Cognitive Responses." Journal of Personality and Social Psychology 37 : 1915 - 26 . Google Scholar Crossref Search ADS WorldCat Pew Research Center . 2015 . "Public and Scientists' Views on Science and Society." http://assets.pewresearch.org/wp-content/uploads/sites/14/2015/01/ PI_ScienceandSociety_Report_012915.pdf. Rigney , Daniel . 1991 . "Three Kinds of Anti-Intellectualism: Rethinking Hofstadter." Sociological Inquiry 6 : 434 - 51 . Google Scholar Crossref Search ADS WorldCat Rooduijn , Matthijs . 2014 . "The Nucleus of Populism: In Search of the Lowest Common Denominator." Government and Opposition 49 : 573 - 99 . Google Scholar Crossref Search ADS WorldCat Shen , Fuyuan , and Heidi Hatfield Edwards . 2005 . "Economic Individualism, Humanitarianism, and Welfare Reform: A Value-Based Account of Framing Effects." Journal of Communication 55 : 795 - 809 . Google Scholar Crossref Search ADS WorldCat Shogan , Colleen J . 2007 . "Anti-Intellectualism in the Modern Presidency: A Republican Populism." Perspectives on Politics 5 : 295 - 303 . Google Scholar Crossref Search ADS WorldCat Skocpol , Theda , and Vanessa Williamson . 2013 . The Tea Party and the Remaking of American Conservatism. New York : Oxford University Press . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Taggart , Paul . 2000 . Populism . Buckingham : Open University Press . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC van der Linden , Sander L. , Anthony A. Leiserowitz , Geoffrey D. Feinberg , and Edward W. Maibach . 2015 . "The Scientific Consensus on Climate Change as a Gateway Belief: Experimental Evidence." PLoS One 10 . doi:10.1371/journal.pone.0118489. Google Scholar OpenURL Placeholder Text WorldCat van der Linden , Sander L. , Anthony A. Leiserowitz , and Edward W. Maibach . 2019 . "The Gateway Belief Model: A Large-Scale Replication." Journal of Environmental Psychology 62 : 49 - 58 . Google Scholar Crossref Search ADS WorldCat Winterich , Karen Page , Yinlong Zhang , and Vikas Mittal . 2012 . "How Political Identity and Charity Positioning Increase Donations: Insights from Moral Foundations Theory." International Journal of Research in Marketing 29 : 346 - 54 . Google Scholar Crossref Search ADS WorldCat Wood , Thomas , and Ethan Porter . 2019 . "The Elusive Backfire Effect: Mass Attitudes' Steadfast Factual Adherence." Political Behavior 41 : 135 - 63 . Google Scholar Crossref Search ADS WorldCat Zaller , John . 1992 . The Nature and Origins of Mass Opinion . Cambridge : Cambridge University Press . Google Scholar Crossref Search ADS Google Preview WorldCat COPAC Footnotes 1. A Pew survey (2015) found that 88 percent of members of the American Association for the Advancement of Science (AAAS) think GMOs are safe to eat, versus only 37 percent of Americans. Pew found an average gap of 24 percentage points between AAAS members and the public across 13 different items. 2. The estimation based on the MTurk data was only done on respondents in the control condition (N = 619) to ensure these estimates are not contaminated with experimental effects. 3. GSS analyses are weighted with WTSSALL, as recommended by the GSS user guide. MTurk results are unweighted. 4. It is possible that ideology exerts an influence on support for positions of expert consensus through anti-intellectualism. Engaging in observational mediation analysis is a daunting task (Green, Ha, and Bullock 2010), but nonetheless this is an unlikely explanation. Adding anti-intellectualism into the observational models does not dampen the coefficients on ideology at all, on any issue. Details of these analyses are available upon request. 5. The effect sizes are stronger in the MTurk sample compared to the GSS, but nonetheless they are comparable to results found in more representative samples. More details on this analysis can be found in table S11 of the online supplementary material. Smaller effect sizes for the GSS are likely a function of limited variation in the independent variable and limited measurement validity. 6. In contrast, there were only significant interactions of interest and ideology on climate change and fluoride. Estimates are provided in table S4 of the online supplementary material. There is only limited variation in the measure of political interest, so analyses were run on all respondents (N = 3,614), controlling for treatment assignment, for more precise estimates. 7. There was little evidence that the source of the rhetoric made a difference in the results that follow, as shown in table S5 and figure S5 of the online supplementary material, so these conditions are collapsed for the following analysis. 8. One might be concerned that this article might increase support for positions of expert consensus. It does not seem like this is the case. In an MTurk sample collected a year later, the share of subjects agreeing at some level with the climate consensus was an identical 80 percent. 9. A total of 83 percent passed a pair of attention checks embedded in the survey. Sixty-eight respondents withdrew their consent and were dropped from all analyses in the paper. 10. The expert trust battery was asked before the consensus-cue treatment was administered, but after the rhetoric treatment. Though this design could potentially bias the results (Montgomery, Nyhan, and Torres 2018), there is no indication that my rhetoric manipulation influenced respondents' reported trust in experts. Diagnostic tests on this can be found in table S6 of the online supplementary material. 11. Table S7 of the online supplementary material presents the results of an OLS estimation without controls. 12. There is little evidence that the observed backfire effect is the result of a linear specification. Table S8 and figure S6 show the results of estimations with a quadratic anti-intellectualism term. 13. (1) People like me don't have much say in what government does; (2) Politics usually boils down to a struggle between the people and the powerful; (3) The system is stacked against people like me; (4) It doesn't really matter who you vote for because the rich control both political parties; (5) People at the top usually get there from some unfair advantage (seven-point, "strongly agree" to "strongly disagree"). 14. This activation effect appears to be stronger for those who are less politically interested, as expected, and these respondents are generally more responsive to rhetoric from more clearly partisan sources. The marginal effects from this analysis can be found in figure S8 of the online supplementary material. Results should be treated with caution. A larger sample size is needed for more precise, confident estimates. (c) The Author(s) 2020. Published by Oxford University Press on behalf of the American Association for Public Opinion Research. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https:/ /academic.oup.com/journals/pages/open_access/funder_policies/chorus/ standard_publication_model) Issue Section: Articles Download all slides Supplementary data Supplementary data nfz053_suppl_Supplementary_Material - pdf file Advertisement 3,500 Views 12 Citations View Metrics x Email alerts Article activity alert Advance article alerts New issue alert Receive exclusive offers and updates from Oxford Academic Related articles in * Web of Science * Google Scholar Citing articles via Web of Science (12) Google Scholar Crossref * Latest * Most Read * Most Cited Using Administrative Records and Survey Data to Construct Samples of Tweeters and Tweets Lost in a Gallup: Polling Failure in U.S. Presidential Elections Family Matters: Education and the (Conditional) Effect of State Indoctrination in China Review: Presidential Job Approval Partisan Bias in Bipartisan Places?: A Field Experiment Measuring Attitudes Toward the Presidential Alert in Real Time Advertisement Advertisement * About Public Opinion Quarterly * Editorial Board * Author Guidelines * Contact Us * Facebook * Twitter * Purchase * Recommend to your Library * Advertising and Corporate Services * Journals Career Network Public Opinion Quarterly * Online ISSN 1537-5331 * Print ISSN 0033-362X * Copyright (c) 2021 American Association for Public Opinion Research * About Us * Contact Us * Careers * Help * Access & Purchase * Rights & Permissions * Open Access * Potentially Offensive Content Connect * Join Our Mailing List * OUPblog * Twitter * Facebook * YouTube * Tumblr Resources * Authors * Librarians * Societies * Sponsors & Advertisers * Press & Media * Agents Explore * Shop OUP Academic * Oxford Dictionaries * Epigeum * OUP Worldwide * University of Oxford Oxford University Press is a department of the University of Oxford. It furthers the University's objective of excellence in research, scholarship, and education by publishing worldwide Oxford University Press * Copyright (c) 2021 Oxford University Press * Cookie Policy * Privacy Policy * Legal Notice * Site Map * Accessibility Close This Feature Is Available To Subscribers Only Sign In or Create an Account Close This PDF is available to Subscribers Only View Article Abstract & Purchase Options For full access to this pdf, sign in to an existing account, or purchase an annual subscription. Close Scholarly IQ