Racial Resentment

The first psychological predisposition we examine acknowledges the importance of racial attitudes in America. The issue of race predates the republic itself and is emblazoned into the nation’s founding documents and history. The struggle with slavery and its tragic consequences, such as the Civil War, Reconstruction, and Jim Crow, have plagued the nation for more than two centuries. American citizens across generations have formed, altered, shared, and harbored views about race and many of its consequences (e.g., Du Bois, Reference Du Bois and Huggins1903; Kinder and Sanders, Reference Kinder and Sanders1996; Myrdal, Reference Myrdal1944; Sniderman and Piazza, Reference Sniderman and Piazza1993). There are trigger moments when views about race come to the fore such as the bombing of the 16th Street Baptist Church in Birmingham, Alabama, in the fall of 1963, a prelude to the Civil Rights Act of 1964; the murder of Martin Luther King in the spring of 1968, with subsequent protests across the country; and the epidemic of killings of Black men at the hands of police in 2014 (e.g., Eric Garner in New York City on July 17, 2014; Michael Brown in Ferguson, Missouri, on August 9, 2014), spurring the growth of the Black Lives Matter movement.

In the context of the 2020 campaign, views about race were inevitably going to be salient. The sitting president had a history of controversial behavior regarding race. In the 1970s, the US Department of Justice sued Trump and his father for violating the Fair Housing Act of 1968 by refusing to rent apartments to people “because of race and color” (Dunlap, Reference Dunlap2015). Decades later, Trump became a political force when he embraced “birtherism” with the false charge that the nation’s first Black president, Barack Obama, was not born in the United States (Abramson, Reference Abramson2016). And approximately six months into Trump’s presidency, hundreds of neo-Nazis and white supremacists marched in Charlottesville to protest the removal of a statue of Robert E. Lee. On the day of these protests, August 12, 2017, a Black man named DeAndre Harris was beaten by at least four white supremacists. On the same day, a twenty-year-old white supremacist from Ohio drove his car into a crowd of counterprotesters, killing thirty-two-year-old Heather Heyer and injuring thirty-five others.Footnote ¹⁸

Speaking a few days later at a news conference, Trump said, “I think there’s blame on both sides. If you look at both sides – I think there’s blame on both sides…and you had some very bad people in that group, but you also had people that were very fine people, on both sides.” Joe Biden said Trump’s comments at the August press conference motivated him to challenge Trump in 2020: “With those words, the president of the United States assigned a moral equivalence between those spreading hate and those with the courage to stand against it” (Holan, Reference Holan2019).

Based on the citizen-centered theory of campaigns, we expect people’s views of race to influence their search for information and how they subsequently interpret and evaluate the candidates, campaign issues, and campaign events. As an illustration, Trump launched his successful presidential campaign in the summer of 2015 saying, “When Mexico sends its people, they’re not sending the best. They’re not sending you; they’re sending people that have lots of problems and they’re bringing those problems. They’re bringing drugs, they’re bringing crime. They’re rapists and some, I assume, are good people” (Gamboa, Reference Gamboa2015). Citizens who were sympathetic to Trump’s message may have sought to acquire additional information confirming the link between immigration and crime and may have turned to news sites and social media platforms producing such content.Footnote ¹⁹ We argue that the easy availability of information supporting divergent viewpoints makes it possible for people to acquire communications consonant with their predilections, thereby affecting people’s interpretation of racial messages during political campaigns.

We know that racial attitudes influence voting preferences (e.g., Valentino, Hutchings, and White, Reference Valentino, Hutchings and White2002), support for government policy (e.g., Gilens, Reference Gilens1995), and approval of the president (e.g., Tesler, Reference Tesler2013). While scholars have offered alternative measures of racism (for a review, see Huddy, Feldman and Sen, 2023), we utilize the racial resentment scale.Footnote ²⁰ Kinder and Sanders (Reference Kinder and Sanders1996: 105–106) explain that racial resentment is a form of racism that rests on the view “that Blacks do not try hard enough to overcome the difficulties they face and they take what they have not earned.” We rely on the following four items to measure racial resentment: (1) “Irish, Italians, Jewish, and many other minorities overcame prejudice and worked their way up. Blacks should do the same without any special favors.” (2) “Over the past few years Blacks have gotten less than they deserve.” (3) “It’s really a matter of some people not trying hard enough; if Blacks would only try harder, they could be just as well off as whites.” (4) “Generations of slavery and discrimination have created conditions that make it difficult for Blacks to work their way out of the lower class” (Henry and Sears, Reference Henry and Sears2002: 260–261). Respondents are asked to agree or disagree with each of the statements on a five-point scale.Footnote ²¹

Hostile Sexism

The second psychological predisposition we examine recognizes the history of sexism in the United States. The dynamic nature of public opinion about women’s rights and opportunities have formed the backbone of several large mass movements working to acquire greater equality for women in society, including the Women’s Suffrage Movement as part of the Progressive Movement (1890s–1920s), the Second Wave Feminist Movement (1960s–1970s), and the #MeToo Movement (2006–present).Footnote ²²

While women’s role in society has improved over the arc of US history, sexism remains a pernicious fact in contemporary society. For instance, women are still paid less than men, regardless of their level of education (Barroso and Brown, Reference Barroso and Brown2021). Further, women hold less than 30 percent of the seats in the US Congress and only 18 percent of the governorships (CAWP, 2021), and only 8 percent of the 500 largest corporations in the United States have women CEOs (Mazzoni, Reference Mazzoni2021). A poll conducted in the summer of 2020 showed that a majority (57 percent) of respondents believe “the U.S. hasn’t gone far enough when it comes to giving women equal rights with men” (Barroso, Reference Barroso2020). Among respondents who think the country has work to do in achieving gender equality, 77 percent say sexual harassment is a major obstacle to women’s equality; 67 percent say women do not have the same legal rights as men; 66 percent say societal expectations are different for men and women; and 64 percent say women are less likely to be in positions of power compared to men.

With this backdrop, the candidacy and election of Donald Trump in 2016 was seen by some as an assault on women, spurring the Women’s March on the day after Trump’s inauguration in January 2017. Hundreds of thousands gathered in Washington and thousands more protested in cities across the country (Hartocolli and Alcindor, 2017). Trump has a long history of misogyny, sexual harassment, sexism, and unwanted sexual advances (Nelson, Reference Nelson2016). A total of eighteen women publicly accused him of sexual misconduct (Keneally, Reference Keneally2020). Further, he repeatedly disparaged women’s looks, such as calling comedian Rosie O’Donnell “a big fat pig,” “disgusting,” “a slob,” and “a very unattractive person.”

Biden’s history with women was not unblemished. For instance, he faced sharp criticism for his treatment of Anita Hill during the Clarence Thomas hearings in 1991. However, since those hearings more than thirty years ago, Biden worked to “redeem himself with women” by advocating for women, such as introducing the Violence against Women Act in 1994 (Kranish and Viser, Reference Kranish and Viser2020). In March 2020, he committed to picking a woman to be the vice president if he secured the Democratic nomination. A few months later, he made good on his promise by picking Kamala Harris as his running mate (Schwartz, Reference Schwartz2020). Harris was a historic choice: the first Black woman and the first person of Indian descent to be nominated for national office by a major party (Burns and Glueck, Reference Burns and Glueck2020).

A growing literature suggests that people’s attitudes toward women influence their views of politics, including assessments of presidential candidates (e.g., Filindra, Kaplan, and Buyuker, Reference Filindra, Kaplan and Buyuker2021; Glick, Reference Glick2019; Ratliff et al., Reference Ratliff, Redford, Conway and Smith2019; Schaffner, MacWilliams, and Nteta, Reference Schaffner, MacWilliams and Nteta2018). While sexist attitudes have been measured in a number of ways (e.g., Larsen and Long, Reference Larsen and Long1988; Spence and Hahn, Reference Spence and Hahn1997), we turn to Glick and Fiske’s (Reference Glick and Fiske2001: 116) ambivalence sexism theory that highlights the “coexistence of power difference and intimate interdependence between the sexes.” This conceptualization produces two dimensions of sexism: hostile sexism and benevolent sexism. Hostile sexism justifies men’s power through sexist antipathy and predicts hostility toward women who are seen as challenging men’s power. Benevolent sexism legitimizes men’s power by promising women that men will take care of them, predicting benevolence toward women who behave in a way consistent with conventional gender norms.

We focus on hostile sexism, characterized by negative attitudes that demean women, including viewing women as untrustworthy, power-seeking, and manipulative (Glick and Fiske, Reference Glick and Fiske1996). Glick (Reference Glick2019: 715) explains that given Trump’s “misogynistic behavior and heteronormative masculinity, hostile sexists would likely view Trump as holding similar attitudes to themselves.” We rely on the following four items developed by Glick and Fiske to assess hostile sexism: (1) Many women are actually seeking special favors, such as hiring policies that favor them over men, under the guise of asking for “equality.” (2) Feminists are not seeking for women to have more power than men. (3) Most women interpret innocent remarks or acts as being sexist. (4) Feminists are making entirely reasonable demands of men.Footnote ²³

According to the citizen-centered theory of campaigns, people’s level of hostile sexism will influence their search for campaign information, thereby reinforcing their views of women’s role in politics and society. We expect hostile sexists to accumulate information supporting their stereotypical views of women, potentially altering how they view the political candidates and the campaign. For example, we hypothesize that hostile sexists will agree with Trump’s characterization of Kamala Harris, the Democratic vice-presidential nominee, as “nasty,” “mad,” “angry,” and “a monster” (Clifton, Reference Clifton2020; Summers, Reference Summers2020). Further, these individuals may pursue sources and stories portraying Harris in a negative light, leading to the development of even more critical views of Harris over the course of the campaign.

Authoritarianism

The United States was founded on rebellion against the actions of a remote and unaccountable king. Charles Edel (Reference Edel2021) explains that the Declaration of Independence articulates “American opposition to the closed, authoritarian models of governance that they knew from abroad and feared would develop in their own country.” Nevertheless, the country has dealt with authoritarian figures throughout its history, including Joseph McCarthy who stoked national fears during the Cold War and presented himself as the arbitrator of truth surrounded by deceitful enemies.

Authoritarianism among the public, too, is not new to US politics. The study of the authoritarian personality began with Adorno et al. (Reference Adorno, Frenkel-Brunswik, Levinson and Sanford1950), who described an authoritarian as someone who admires figures of strength, disdains those who are weak, prefers conventionalism, rigidity, and stereotypical thinking, and exhibits aggression toward out-groups. More recent scholars confirm that conformity, obedience to authority, and out-group aggression are central to understanding authoritarianism (e.g., Altemeyer, Reference Altemeyer1981; Feldman and Stenner, Reference Feldman and Stenner1997; Stenner, Reference Stenner2005). Finally, Butler shows (2013) that authoritarians, compared to nonauthoritarians, are more fearful of both social threats (e.g., social disorder) and personal threats (e.g., confronting a dangerous situation).

Trump, during his campaigns and while governing, showed signs of authoritarianism by embracing autocratic rulers abroad, stoking fears about immigrants (e.g., calling Latinos “animals” and “invaders”),Footnote ²⁴ and damaging democratic norms at home, including interfering with judicial independence, attacking the news media, and questioning the integrity of the election. Also, he flattered authoritarian strongmen like Vladimir Putin of Russia, Rodrigo Duterte of the Philippines, and North Korean dictator Kim Jong Un. According to senior aides, in a phone call with Putin in the early days of the administration, Trump was “obsequious” and “fawning,” even apologizing to the dictator for not calling him sooner (Leonnig, Harris, and Dawsey, Reference Leonnig, Harris and Dawsey2019).

Research on authoritarianism and politics has consistently demonstrated a strong relationship between level of authoritarian thinking and generalized prejudice, nationalism, opposition to civil liberties, and support for aggressive foreign policy (for a review, see Caprara and Vecchione, Reference Caprara, Vecchione, Huddy, Sears and Levy2013). In addition, some researchers find that authoritarians are more likely to support prejudicial and restrictive government policies in response to threat (e.g., Feldman, Reference Feldman2003; Feldman and Stenner, Reference Feldman and Stenner1997; Lavine, Lodge, and Freitas, Reference Lavine, Lodge and Freitas2005, but also see Hetherington and Suhay, Reference Hetherington and Suhay2011). Finally, in the 2016 election, a number of studies demonstrate a positive correlation between support for Donald Trump and level of authoritarian thinking (e.g., Choma and Hanoch, Reference Choma and Hanoch2017; Knuckey and Hassan, Reference Knuckey and Hassan2020; Womick et al., Reference Womick, Rothmund, Azevedo, King and Jost2019).

When Lane (Reference Lane1955: 173) explored the impact of the “authoritarian personality” in the 1952 presidential election, he argued that “in an electoral situation, as in any other situation, personality factors play a double role: (1) they affect the perceptions of the individual, screening out some stimuli, distorting others, and leaving others intact; and (2) they shape the responses of a person, selecting among the various possible responses those which are more serviceable to basic personality needs.” His argument is entirely consistent with the citizen-centered theory of campaigns.

During Trump’s 2016 campaign, his presidency, and his reelection campaign, he presented himself as the strong leader who was needed at this moment in US history to protect Americans from dangerous elements inside and outside their borders. We expect that Trump’s rhetoric will resonate differently depending on people’s level of authoritarianism, thereby influencing how they obtain and process campaign news and altering assessments of the competing candidates.

To measure authoritarianism, we rely on Feldman and Stenner’s (Reference Feldman and Stenner1997) measure that likens hierarchical thinking in child-rearing with hierarchical thinking in society.Footnote ²⁵ Respondents are given the following instructions: “Although there are a number of qualities that people feel that children should have, every person thinks that some are more important than others. Read the following pairs of desirable qualities. Please indicate which one is more important for a child to have: (1) Respect for Elders or Independence; (2) Obedience or Self-Reliance; (3) Curiosity or Good Manners: (4) Being Considerate or Well-Behaved.”Footnote ²⁶ In a recent examination, Engelhardt, Feldman, and Hetherington (Reference Engelhardt, Feldman and Hetherington2021) show that the child-rearing measure of authoritarianism is related to Right-Wing Authoritarianism (RWA), suggesting both measures are tapping the same underlying concept. Second, the authors demonstrate that the child-rearing measure assesses authoritarianism equally well for whites and nonwhites. Third, the authors find no evidence that the child-rearing measure of authoritarianism is endogenous to political and social attitudes, a critique leveled against the RWA measure. Finally, the child-rearing measure is temporally stable as one would expect from a measure tapping a “personality adaptation.”

Conflict Avoidance

We turn next to discussing the psychological predisposition of conflict avoidance. E. E. Schattschneider (Reference Schattschneider1960: 3), in his classic book The Semi-Sovereign People: A Realist’s View of Democracy in America, suggests politics is about conflict: “At the nub of all politics are: first, the way in which the public participates in the spread of the conflict and, second, the processes by which the unstable relation of the public to the conflict is controlled.” It is no secret that conflict is baked into electoral campaigns with competition around the development of public policies, representation, and resources. Further, the conflictual nature of politics is highlighted in news coverage. In particular, news media organizations embrace the “conflict frame” by emphasizing “conflict between individuals, groups, or institutions as a means of capturing audience interest” (Semetko and Valkenburg, Reference Semetko and Valkenburg2000: 95). For example, clashes between political elites of opposing parties in Congress, intraparty disagreements over policy, or news descriptions of elections often employ metaphors like “battle,” “fight,” and “brawl.”

It is also not surprising that people vary in their comfort with conflict, with some people hoping to avoid confrontation altogether while others may actually seek it out (Ulbig and Funk, Reference Ulbig and Funk1999). People’s tolerance of conflict is important in family relationships, in the workplace, and in educational settings (e.g., Barsky and Wood, Reference Barsky and Wood2005; Koerner and Fitzpatrick, Reference Koerner and Fitzpatrick1997; Weider-Hatfield and Hatfield, Reference Weider-Hatfield and Hatfield1995). Tolerance toward conflict has been conceptualized as a stable disposition resulting from early childhood socialization (e.g., Bresnahan et al., Reference Bresnahan, Donohue, Shearman and Guan2009; Mutz, Reference Mutz2002; Testa, Hibbing, and Ritchie, Reference Testa, Hibbing and Ritchie2014).

A number of scholars have examined the impact of conflict avoidance on a range of political attitudes and behaviors. For example, Mutz and Reeves (Reference Mutz and Reeves2005) show that exposure to incivility lowers political trust among people high in conflict avoidance. Sydnor (Reference Sydnor2019) finds that conflict-avoidant individuals are more likely to experience negative affect (e.g., disgust, anger) when exposed to incivility. And several studies have examined the link between level of conflict avoidance and political participation (e.g., Bjarnøe, de Vreese, and Albæk, Reference Bjarnøe, de Vreese and Albæk2020; Mutz, Reference Mutz2002; Sydnor, Reference Sydnor2019; Ulbig and Funk, Reference Ulbig and Funk1999; Wolak, Reference Wolak2020).

The 2020 presidential campaign was rife with conflict; the toxic mix of racial division, polarized politics, and a competitive presidential election produced a highly contentious campaign. Nevertheless, the two candidates differed in their penchant for discord. In a New York Times analysis published at the end of August 2020, Peter Baker and Maggie Haberman (Reference Baker and Haberman2020b) write, “President Trump has been throwing accelerant on the fire of the nation’s social unrest rather than trying to put it out, seeking confrontation rather than calm… From his time as a celebrity real estate developer, Mr. Trump has never been a conciliator, and he has long gravitated toward conflict and sought to escalate it.”

Also in August 2020, when accepting the Democratic nomination for president in Milwaukee, Biden emphasized unity over division. He said,

Given the conflictual nature of the 2020 presidential election as well as differences in the two candidates’ preference for divisiveness, we hypothesize that people’s level of conflict avoidance will serve as a lens for people to receive, interpret, and draw conclusions about the candidates and events of the campaign. To measure conflict avoidance, we rely on the “approach/avoidance component” of Goldstein’s (Reference Goldstein1999) Conflict Communication Scale. This conflict avoidance scale includes five items where people are asked to indicate how much they agree with the following items on a four-point scale ranging from disagree strongly to agree strongly: (1) I hate arguments; (2) I find conflicts exciting; (3) I feel upset after an argument; (4) I enjoy challenging the opinions of others, and (5) arguments do not bother me.Footnote ²⁸

Conspiracy Thinking

We turn to the fifth and final psychological predisposition: conspiracy thinking. Conspiracy thinking (also called conspiracy ideation or conspiracy mentality) is a stable predisposition that leads individuals to attribute “events to a secret plot by a covert alliance of powerful individuals or to clandestine organizations rather than to more mundane human (in)activity or natural forces” (Imhoff and Bruder, Reference Imhoff and Bruder2014: 25). Conspiracy theories are universal and occur across time and cultures and in a wide variety of social settings (e.g., government, workplace, school).

One of the most enduring political conspiracies in the United States involves the assassination of President John F. Kennedy. The Warren Commission, authorized by President Lyndon Johnson and led by Chief Justice Earl Warren, investigated the assassination and concluded Lee Harvey Oswald acted alone in the assassination of Kennedy. These findings were supported, by and large, by additional investigations by the US Attorney General’s Office in 1968, by the Rockefeller Commission in 1975 and by the US House of Representatives from 1978 to 1979. Nevertheless, people’s suspicions about Kennedy’s assassination persist more than fifty years after the event. For example, a survey commissioned by FiveThirtyEight in 2017 found 61 percent of respondents believed that Lee Harvey Oswald did not act alone.Footnote ²⁹

Additional conspiracies are widely circulated in the United States including: the moon landing in 1969 was faked, the George W. Bush Administration knew about the planned 9/11 terrorist attacks and let it happen, and President Obama was not born in the United States. More recent examples of conspiracy theories include the QAnon conspiracy claiming the world is run by a cabal of Satan-worshiping pedophiles. QAnon followers believe that this cabal includes top Democrats like Joe Biden, Hillary Clinton, Barack Obama and George Soros. According to the QAnon theory, “Donald J. Trump was recruited by top military generals to run for president in 2016 to break up this criminal conspiracy and bring its members to justice.”Footnote ³⁰

During the COVID-19 pandemic, a number of conspiracies were disseminated regarding the virus, such as (1) COVID-19 is no worse than the flu; (2) wealthy elites intentionally spread the virus to win power and profit; and (3) wearing a mask makes people more susceptible to COVID-19 (Lewis, Reference Lewis2020). Additional conspiracy theories developed surrounding the safety and efficacy of the COVID-19 vaccines, including (1) the COVID-19 vaccine makes you magnetic; (2) the COVID-19 vaccine makes you infertile; (3) the government put a microchip in COVID-19 vaccines to track you; and (4) the COVID-19 vaccines rewrite your DNA (Cassata, Reference Cassata2020).

There is a growing literature across a number of disciplines aimed at understanding the determinants of conspiratorial thinking among individuals (for a review see Douglas et al., Reference Douglas, Uscinski, Sutton, Cichocka, Nefes, Ang and Deravi2019). Research suggests that conspiracy thinking is higher among people with lower levels of trust and higher levels of alienation (e.g., Goertzel, Reference Goertzel1994; Imhoff and Bruder, Reference Imhoff and Bruder2014). Further, people’s belief in conspiracy theories is highly sensitive to social context. For example, Republicans are more likely to believe in governmental conspiracies when there is a Democrat in the White House and Democrats are more likely to believe in governmental conspiracies when a Republican is president (e.g., van Prooijen and Douglas, Reference Van Prooijen and Douglas2018).

The media may play an important role in increasing levels of conspiracy thinking. Hollander (Reference Hollander2018) reports exposure to Fox News is positively related to belief in conspiracy theories; Stempel, Hargrove, and Stempel (Reference Stempel, Hargrove and Stempel III2007) find that the use of blogs and YouTube news channels stimulates conspiracy thinking; Walter and Drochon (Reference Walter and Drochon2020) show that conspiracy thinking is higher when people turn to blogs and nonmainstream social media, compared to newspapers, for their news.

Regarding political attitudes and political behavior, researchers have linked conspiracy thinking to attitudes toward climate change (Uscinski and Olivella, Reference Uscinski and Olivella2017) and immigration (Gaston and Uscinski, Reference GaSton and Uscinski2018). Further, some studies demonstrate a positive relationship between conspiracy thinking and levels of political engagement (e.g., Imhoff, Reference Imhoff, Bilewicz, Cichocka and Soral2015; Imhoff and Bruder, Reference Imhoff and Bruder2014; Kim, Reference Kim2019). However, Uscinski and Parent (Reference Uscinski and Parent2014) report that conspiracy thinking decreases people’s likelihood of participating in conventional political activities while people higher in conspiracy thinking are more likely to support violence against the government.

During the 2020 campaign, Trump was aggressively promoting a conspiracy theory regarding voter fraud, especially mail voting. On July 2, 2020, Trump tweeted, “Mail-In Ballots will lead to massive electoral fraud and a rigged 2020 Election.” By September 11, 2020, a fact-check by the Washington Post reported that Trump had “peddled imaginary threats about voting by mail” more than 100 times (Rizzo, Reference Rizzo2020). By Election Day, Trump had made more than 150 false claims concerning “fraudulent ballots or the alleged dangers of mail-in voting” (Kessler and Rizzo, Reference Rizzo2020). In the modern era, there has never been an incumbent president who openly and repeatedly disseminated conspiracy theories regarding the security of a US election. Therefore, we believe the 2020 presidential election is an ideal setting for examining how people’s level of conspiracy thinking influences their interpretation of campaign information and evaluations of the ongoing political campaign.

To measure conspiracy thinking, we rely on the following three items from the Conspiracy Mentality Questionnaire: “I think that government agencies closely monitor all citizens”; “I think that events which superficially seem to lack a connection are often the result of secret activities”; and “I think that there are secret organizations that greatly influence political decisions” (Bruder et al., Reference Bruder, Haffke, Neave, Nouripanah and Imhoff2013: 5).Footnote ³¹ Respondents are asked to indicate their level of agreement with each of these statements on an eleven‐point scale ranging from 1 = 0 percent – certainly not to 11 = 100 percent – certain.

To review, we argue that the availability of information from a variety of sources allow citizens to actively search for and assimilate information consistent with their psychological and political predispositions. We expect these predispositions to consistently influence how people interpret campaign information and evaluate campaign events, issues, and candidates. We also contend that people’s psychological predispositions do not simply mirror their political proclivities. Instead, we expect that partisan attitudes will be orthogonal to psychological attitudes.

Relying on our original dataset, we examine the correlation between party identification and each of the five psychological predispositions.Footnote ³² The findings in Table 1.1 indicate that psychological predispositions are not simply a proxy for partisanship. We find the strongest correlations between partisanship and racial resentment and hostile sexism (.42 and .35, respectively). However, these correlations do not suggest that party identification is simply collinear with level of prejudice.Footnote ³³ Further, the relationship between partisanship and authoritarianism, conspiracy thinking, and conflict avoidance is weak to nonexistent (i.e., .16, .10, .01, respectively).

We also look at the correlations among the five psychological predispositions. We see that the correlations between these constructs vary from a low of .03 (i.e., racial resentment and conflict avoidance) to a high of .47 (i.e., hostile sexism and racial resentment), with an average correlation of less than .18. The data in Table 1.1 suggest that the five psychological characteristics are tapping something different from partisanship, and these psychological predispositions are not highly correlated with one another, with the exception of racism and sexism.

Throughout the book, we test the citizen-centered theory of campaigns with data from a three-wave panel study of more than 4,000 people interviewed in September, October, and immediately after Election Day in November 2020. We turn next to a description of the design and measurement of the panel survey.

The Panel Design and Sample

We designed a panel study where information is gathered from the same individuals at different points during the campaign to study the dynamics of the 2020 presidential campaign. According to Finkel (Reference Finkel1995: 1), an important advantage of panel studies is that “change is explicitly incorporated into the design so that individual changes in a set of variables are directly measured.” Brady, Johnston, and Sides (Reference Brady, Johnston, Sides, Brady and Johnston2006) explain that panel designs can isolate change when the waves of the panel straddle a “treatment” such as a presidential debate.Footnote ³⁴ For untangling causal relationships, such as how campaign events alter views about the competing candidates, panel designs have distinct advantages over cross-sectional designs. Precisely because we are measuring political attitudes for the same individuals across time, we can model how campaign events change people’s attitudes.Footnote ³⁵

We conducted a three-wave panel survey during the 2020 presidential campaign with recruitment done by Dynata (formerly SSI) using an opt-in internet panel. Dynata maintains a large online panel of US adults. It uses invitations (e.g., email, web banners, phone alerts) to enroll people in research panels.Footnote ³⁶ It screens and recruits participants with quotas (e.g., sex, age, ethnicity) in order to obtain a sample with demographics representative of the United States (Shaverdian et al., Reference Shaverdian, Yeboa, Gardner, Harari, Liao, McCloskey, Tuli, Vapiwala and Jagsi2019). The opt-in sample is a nonprobability sample, and studies have shown that nonprobability samples can produce accurate results (e.g., Twyman, Reference Twyman2008; Vavreck and Rivers, Reference Vavreck and Rivers2008).

In Table 1.2, we provide a comparison of the demographic profile of the panel respondents with recent census data. The data in Table 1.2 show that the panel respondents are generally representative of the nation. However, they tend to be somewhat more educated, older, and more male compared to census numbers. In terms of partisanship, the panel respondents are somewhat more likely to identify with the Democratic party and less likely to call themselves independents, compared to a Gallup survey in the field during the 2020 presidential campaign.Footnote ³⁷

Respondents completed each wave of survey on the Qualtrics platform. We collected the first wave of the panel survey between September 1 and September 22, 2020, with 3,013, respondents completing the questionnaire (see Table 1.3). The second wave was conducted between October 4 and October 7, 2020, with 1,510 respondents completing the October wave of the panel. We also collected a fresh sample of 1,298 respondents in October between October 5 and October 18, 2020. Finally, we recontacted all of the respondents for a final wave of the survey between November 4 and November 6, 2020, with 2,153 completed surveys. During the fall campaign, we collected survey data from 4,311 respondents, with 1,040 respondents completing each of the three waves of the panel survey. The September wave of the survey took about 23 minutes to complete (i.e., an average of 23.10 minutes). The October wave of the survey took somewhat less time, averaging almost 18 minutes (i.e., an average of 17.95 minutes), while the new wave of October respondents averaged just under 22 minutes (i.e., an average of 21.63 minutes). The November wave was the shortest survey, taking an average of 15.13 minutes to complete.

The Survey Questionnaire

We ask a variety of different questions about politics and government, such as party identification, ideology, political knowledge, political trust, civic duty, and political engagement.Footnote ³⁸ We also pose a series of questions about people’s attention to news as well as their preference among different news outlets. Respondents are also queried about the most important problem facing the nation, the state of the economy, and concern about the COVID-19 pandemic and worries about the integrity of the election. Questions about support for racial justice protests and support for local police are also included in each wave of the survey.

The survey instrument also contains a number of questions about people’s views of Trump and Biden, including approval of Trump’s performance in office, feeling thermometer ratings for each candidate, assessments of the candidates’ traits (i.e., leadership, integrity, empathy, and temperament), tone of each candidate’s campaign, and views of each candidate’s ability to deal with a variety of issues (e.g., COVID-19, race relations, the integrity of the election).

As discussed earlier, we include established measures to assess levels of racism (racial resentment) and sexism (hostile sexism), conflict avoidance, authoritarianism, and conspiracy thinking.Footnote ³⁹ Finally, questions about standard demographic variables, including age, gender, education, income, religion, race, and ethnicity, are asked during the first wave of the survey.Footnote ⁴⁰

The Campaign Setting

We focus explicitly on two important events during the 2020 general election campaign. The initial event is the first presidential debate between Trump and Biden. Presidential debates have become routine events during campaigns, with debates occurring each presidential election year since 1976. Although they are highly scripted by the campaigns, detailed preparation is not always a strong predictor of how potential voters view the outcome. The first debate of 2020 was a doozy, and we examine how psychological and political predispositions influence people’s views of the candidates’ performances in the debate. Furthermore, since the debate occurred days before the launching of the second wave of our panel, we examine how views of the debate change people’s overall evaluations of the candidates from September to October.

The second significant event is Trump’s COVID-19 diagnosis and hospitalization. Trump’s COVID-19 infection, unlike the debate, was unscripted and occurred well before the advent of vaccines, generating a great deal of discussion and concern about the health of the incumbent president. His diagnosis also fell between the September and October waves of our panel survey. This serendipitous timing provides us with analytical leverage to examine how Trump’s contraction of the virus alters people’s views about the COVID-19 pandemic as well as changing assessments of his ability to deal with the pandemic.

In addition to the COVID-19 pandemic, we examine two more issues during the fall campaign. The first issue, the protest movement against police brutality, was triggered by the murder of George Floyd by the Minneapolis police in May 2020. This event reignited the sustained, emotional, and complex issue of race in America in the midst of a presidential campaign and generated a great deal of discussion by the candidates about racial justice and policing. We measure attitudes toward the social justice movement as well as views of police during each wave of the panel, allowing us to examine dynamics in people’s opinions and explore how views about racial justice and policing influence evaluations of the candidates over the course of the campaign.

Finally, we examine the issue of election integrity. This issue surfaced during the 2016 presidential campaign, but it became a focus of the 2020 campaign. For instance, during the course of the campaign, Trump and some members of the GOP questioned the legitimacy of voting by mail. While Trump’s rhetoric about a “rigged election” was persistent and accelerated as Election Day approached, Biden and Democratic Party surrogates responded by emphasizing the safety and security of convenience voting. We explore the determinants of people’s views about election integrity as well as citizens’ assessments of each candidate’s ability to ensure the legitimacy of the election results. Further, we explore how confidence in election integrity influences how people decide to cast their ballot: on Election Day, by mail, or via in-person early voting.

An Analytical Road Map

In each of the next four chapters, we focus on one issue or event. In particular, Chapter 2 examines the September presidential debate; Chapter 3 centers on President Trump’s COVID-19 diagnosis; Chapter 4 looks at attitudes toward social justice protests and policing; and Chapter 5 addresses the issue of election integrity. In each chapter, we employ the same basic analytical strategy.

First, we develop models to understand people’s attitudes about each event and issue. So, for example, in Chapter 2, we present models explaining people’s views about the performances of Trump and Biden in the first debate. Based on the citizen-centered theory of campaigns, we expect people’s psychological predispositions and political proclivities (e.g., partisanship) to powerfully influence views about the event or issue. In these models, we also control for rival factors that may influence assessments. For instance, when we predict people’s worries about COVID-19 in Chapter 3, we include demographic factors that may influence people’s concern about the coronavirus. Second, we rely on the panel design to estimate how political and psychological predispositions, along with relevant rival factors, influence changes in people’s views of each issue or event from September to October.Footnote ⁴¹ We conclude each chapter by looking at how assessments of the issue or event influence evaluations of the competing candidates in a baseline model and in a change model. For example, in Chapter 4, we examine how attitudes toward the social justice movement and views toward policing affect attitudes toward Trump and Biden in September. Then we develop change models to estimate how beliefs about police and racial justice protests influence changes in evaluations of Trump and Biden from September to October.Footnote ⁴²

In Chapter 6, we develop a comprehensive model to examine how each of the campaign issues and events, along with rival factors, predict evaluations of Trump and Biden in November. In these models, for example, we see that views of the candidates’ debate performance in September continue to impact views of Trump and Biden in November. In addition, we estimate changes in people’s evaluations of Trump and Biden from September to November and demonstrate that campaign issues and events powerfully alter views of the candidates. We conclude Chapter 6 by examining how the major elements of the campaign produce changes in vote intention from September to November.

We conclude the empirical chapters by demonstrating that the citizen-centered theory of campaigns improves our understanding of who voted in the 2020 election as well as how people decided to cast their vote (i.e., voting on Election Day or relying on convenience voting). In Chapter 7, we begin by predicting turnout where we include psychological predispositions as explanatory variables. Based on the vast literature on political engagement, we also include a series of additional factors (e.g., strength of partisanship, civic duty, political attention). We also predict people’s likelihood of voting by mail (compared to on Election Day) and voting early (compared to on Election Day). We find that psychological predispositions play a powerful role, influencing people’s decision to vote as well as affecting their reliance on convenience voting.

Chapter 2. “A Hot Mess Inside A Dumpster Fire Inside a Train Wreck”: Understanding the Impact of the First Presidential Debate

Presidential debates are now a fixture in the landscape of fall campaigns for the presidency. They attract worldwide media attention, as well as the interest of tens of millions of potential voters, and are held in close proximity to Election Day. In 2020, the first general election debate was a donnybrook. We find citizens develop clear opinions about who won the debate and who performed well; more people viewed Biden as the winner of the first debate and his performance ratings were significantly higher than Trump’s ratings, except among Republicans. We also demonstrate that people who have low tolerance for conflict develop significantly more negative views of Trump’s performance and are significantly more likely to consider Biden the winner of the debate. Furthermore, people’s level of racism and conspiratorial thinking shape views of Trump’s and Biden’s performances during the first presidential debate. Finally, evaluations of the candidates’ performance in the debate as well as people’s views of who won the debate influence overall evaluations of Trump and Biden and produce significant changes in the ratings of Trump and Biden from September to October.

Chapter 3. The Priming of COVID-19 during the Campaign: The Consequences of Trump’s Coronavirus Diagnosis

The 2020 presidential campaign occurred in the midst of the first worldwide pandemic in 100 years. The pandemic engulfed the United States for the entire length of the campaign and the incumbent president was hospitalized with the virus at the height of the fall campaign. We show that people’s concern about the coronavirus pandemic increased significantly after Trump contracted COVID-19. Furthermore, and consistent with the citizen-centered theory of campaigns, we find that psychological predispositions, along with political and demographic characteristics, substantively and significantly predict changes in worry about the coronavirus from September to October. For instance, people high in authoritarianism and conflict avoidance become significantly more worried about the coronavirus pandemic from September to October. Finally, we show that people are more likely to consider assessments of the candidates’ competence for dealing with the coronavirus when developing overall evaluations of the candidates in October – after Trump’s COVID-19 diagnosis – compared to September.

Chapter 4. Protests against Police Brutality: How Attitudes about Racial Injustice and Policing Affected Campaign 2020

The murder of George Floyd at the hands of Minneapolis police over Memorial Day weekend ignited sustained protests across the country and placed the issue of race front and center. By September, more than two-thirds of our survey respondents report positive views of the Black Lives Matter movement. While the salience of race began to fade as the general election campaign unfolded, we find that political characteristics of citizens, such as party attachment and partisan media exposure, influence support for the social justice movement and support for law enforcement. Further, psychological predispositions consistently and significantly influence views of social protests and policing. For example, people’s level of racial resentment produces powerful changes in their views of the protests and police from September to October. Finally, attitudes about racial justice and policing influence overall impressions of Biden and Trump, producing significant changes in people’s views of the candidates during the first months of the fall campaign.

Chapter 5. “A Rigged Election”: How Views about Election Integrity Altered the Campaign

The incumbent president consistently and systematically sowed doubts about the integrity of the American electoral process throughout the 2020 presidential campaign. Trump’s campaign tactic had effects on voters. We show that public confidence in the integrity of the election is much lower for Republicans and for people paying attention to conservative news compared to Democrats and consumers of left-leaning news. Further, a propensity to believe in conspiracy theories fuels doubts about the security of the election. In addition, we show that a number of psychological predispositions consistently influence people’s assessments of Biden’s and Trump’s ability to safeguard the election, including people’s level of racial resentment and level of hostile sexism. Finally, people’s confidence in the security of the election is associated with positive changes in overall evaluations of Biden and negative changes in overall evaluations of Trump from September to October.

Chapter 6. How the Campaign Shapes Voters’ Decisions about the Candidates

We develop a comprehensive model where we include assessments of each campaign event (e.g., September debate) and issue (e.g., election integrity, worries about COVID-19) when predicting overall evaluations of Biden and Trump in November as well as changes in feeling thermometer scores from September to November. These models show that views about the first presidential debate and attitudes toward major campaign issues (i.e., election integrity, COVID-19, social justice protests) explain views of the candidates in November and predict shifts in evaluations over the length of the campaign. Finally, we estimate changes in vote preference from September to November and we find that elements of the campaign (e.g., views about the presidential debate, support for social justice protests) produce important changes in vote preferences.

Chapter 7. The Impact of Campaign Messages on the Decision to Vote

We apply the citizen-centered theory of campaigns to help improve our understanding of participation in the 2020 election. We find a strong positive relationship between conflict avoidance and turnout, with people who dislike conflict participating in the election at a much higher rate than people who are more tolerant of conflict. We also demonstrate the significance of the campaign for understanding turnout; people who watched the September presidential debate, people who have higher levels of confidence in the election results, and people with more polarized views of the social justice movement are significantly more likely to vote in the general election. The citizen-centered theory of campaigns also informs our understanding of convenience voting. People who are more sympathetic to Trump (i.e., Republicans, people with less progressive views on race and gender) are more likely to heed his message of forgoing mail voting and going to the polls on Election Day. Further, people who dislike conflict are significantly more likely to rely on mail voting compared to voting on Election Day. Finally, views about the important issues of the campaign affect how people choose to cast a ballot; people who are more concerned about the COVID-19 pandemic and people with more confidence in the integrity of the election are more likely to vote by mail than in person on Election Day.

Chapter 8. How Campaign 2020 Matters

We begin by highlighting the impressive evidence for the citizen-centered theory of campaigns. In particular, we find that psychological predispositions do not simply reinforce partisan orientation. Instead, these predispositions tap distinct characteristics, influencing how people view the events and issues of the campaign. We also make suggestions about how to study campaigns in the future. While the electoral context of 2020 highlighted particular psychological predispositions, future elections are likely to put a premium on alternative psychological predispositions (e.g., benevolent racism, need for affect). We encourage researchers to be more exhaustive, systematic, and consistent in exploring the impact of people’s psychological predispositions during campaigns. We also review and speculate about how candidates’ campaign strategies may have helped shape the outcome, especially when we consider the razor thin vote margins in a few key states. Specifically, it appears Trump’s actions worked to his detriment both in who voted and in who they supported. Finally, given the events and rhetoric associated with the 2020 campaign, we conclude by assessing the health of our representative system of government where elections play a vital role.