Introduction
Historically, national emergencies (e.g., terrorist attack, war, natural disaster, economic crisis) have provided the occasion for democratically elected leaders to demand special powers in order to lead the country toward more stable footing. Sometimes these demands have been refused, like in the case of Franklin D. Roosevelt's attempt to pack the Supreme Court during the Great Depression (Epstein and Segal, Reference Epstein and Segal2005). Other times, they have been granted, as in the case of Adolf Hitler, Vladimir Putin, and Recep Tayyip Erdoğan concentrating executive power following various security threats attributed to terrorism (Paxton, Reference Paxton2004; Levitsky and Ziblatt, Reference Levitsky and Ziblatt2018). In the United States, fear and the “rally round the flag” effect helped George W. Bush's approval rating soar from 53 to 90% following the September 11 terrorist attacks. This enabled him to pass “The USA PATRIOT Act” the following month, which expanded surveillance powers in searches for possible terrorist threats (Ali and Abdullah, Reference Ali and Abdullah2016).
Former President Donald Trump also sought to subvert democratic norms, citing national emergency. During the height of the COVID-19 pandemic, on July 30, 2020, Trump proposed delaying the 2020 Presidential election, tweeting: “With Universal Mail-In Voting (not Absentee Voting, which is good), 2020 will be the most INACCURATE & FRAUDULENT Election in history. It will be an embarrassment to the USA. Delay the Election until people can properly, securely and safely vote???” Indeed, following his election loss in November 2020, several Republican lawmakers and Trump's own advisors encouraged him to “declare a national emergency,” evoke martial law, suspend civil liberties, and refuse surrendering the Presidency (Forgey et al., Reference Forgey, Montellaro and Oprysko2020; Shepherd, Reference Shepherd2020).
To be sure, most national Constitutions allow executives to expand their power during national emergencies. This is because leaders may need to act quickly and definitively, and, for a limited time, avoid the checks and balances that might hinder their effective leadership. Indeed, Abraham Lincoln, widely regarded as one of our nation's greatest Presidents (C-Span, 2021; Gallup, 2023), suspended habeas corpus and expanded other executive powers to end slavery and the Civil War. And FDR, also commonly ranked among the greatest Presidents (C-Span, 2021), famously violated democratic norms to lead the nation through the Great Depression and WWII. Yet as Levitsky and Ziblatt (Reference Levitsky and Ziblatt2018, 94) explain, “For a demagogue who feels besieged by critics and shackled by democratic institutions, crises open a window of opportunity to silence critics and weaken rivals. Indeed, elected autocrats often need crises” (emphasis theirs). This is because crises concentrate public opinion behind national leaders in a frenzy of fear and nationalism.
Drawing on data from a large, nationally representative survey fielded in Spring 2022, the current study examines what ideological and social characteristics might incline Americans to support leaders violating democratic norms during unspecified “national emergencies” in order to lead the country to safety. Building on research on the ideological and contextual antecedents of democratic backsliding, we anticipate that such support likely stems from different ideological and contextual factors. Ideologically, we expect ethno-nationalistic and authoritarian populism, represented in our study by Christian nationalist ideology, would see expanding executive powers as a way to thwart opposition and gain full control. And contextually, we anticipate that Christian nationalism will incline partisans whose leaders are currently in power to sacralize the current political status quo, thereby justifying leaders (implicitly their leaders) violating democratic norms to maintain the preferred order under the threat of national emergency.
Background
National emergencies and the ideological antecedents of democratic backsliding
The tension between maintaining both a central government with executive powers strong and nimble enough to handle crises on the one hand, and a distribution of powers to thwart tyranny on the other, lies at the foundations of American democracy (Cooper, Reference Cooper2014). And indeed, the tension has never been fully resolved. Less than a decade after the U.S. Constitution was ratified, President John Adams was granted unilateral authority by Congress in 1798 to deport non-citizens who were subjects of foreign enemies and curtail citizens' rights of free speech and free press. The Alien and Sedition Acts were passed citing concerns about the destabilizing influence of French anarchy, but were rooted in ethno-nationalism and partisan hostility (Berkin, Reference Berkin2017; Levitsky and Ziblatt, Reference Levitsky and Ziblatt2023). Indeed, the inter-partisan rancor between Federalists and Republicans was so great; it prompted Alexander Hamilton to support subverting the democratic process: “In times like these in which we live it will not do to be over-scrupulous. It is easy to sacrifice the substantial interests of society by a strict adherence to ordinary rules” (cited in Levitsky and Ziblatt, Reference Levitsky and Ziblatt2023, 18). As with Adams, Hamilton's goal was political control, but with the justification that “times like these in which we live” demanded it.
Just as in the founding generation, contemporary support for executive action outside the norm is often transparently partisan in nature. As cited above, after losing the 2020 Presidential election, Donald Trump received encouragement from advisors to “declare a national emergency” in order to retain the Presidency (Forgey et al., Reference Forgey, Montellaro and Oprysko2020; Shepherd, Reference Shepherd2020). And on the Democratic side, roughly two-thirds (66%) of Democratic-leaning Americans support expanding the number of Supreme Court justices, compared to only 27% of Republican-leaning Americans (Pew Research Center, 2023). These calls have been directed toward President Biden from his constituents, citing a crisis of confidence in the Supreme Court after Republicans' hardball tactics during the Obama Presidency and Trump's election resulted in a highly conservative court (Albertus and Grossman, Reference Albertus and Grossman2021a). In both instances, the violation of democratic norms would largely be done with partisan goals in mind, but citing a “crisis” demanding it.
Yet as scholars have recently shown, the psychological roots of support for democratic backsliding do not seem to emerge primarily from partisanship or even affective partisan polarization (see Hahl et al., Reference Hahl, Kim and Zuckerman Sivan2018; Albertus and Grossman, Reference Albertus and Grossman2021b; Druckman et al., Reference Druckman, Green and Iyengar2023). In their comprehensive analysis of the contemporary antecedents of democratic backsliding, Berlucchi and Kellam (Reference Berlucchi and Kellam2023, 829) conclude that “ideology and institutional contexts may concomitantly act as contributing factors to democratic decline by providing executives with motivation and means.” In particular, they identify populist ideology that “pits pure people against the corrupt elites” (2023, 819) and is often characterized by exclusion of “outsiders” such as ethnic or religious minorities and justifies radical political action “from a sense of enduring emergency” (2023, 820; see also Bonikowski and Zhang, Reference Bonikowski and Zhang2023; Foertsch et al., Reference Foertsch, Chakraborty and Joose2024). Other studies have also connected this sort of populist sentiment with times of political or social upheaval, or what is called the “politics of the extraordinary” (e.g., Kalyvas, Reference Kalyvas2009; Williams, Reference Williams2015; de la Torre, Reference de la Torre2016).
The combination of populist discontent (with its concomitant penchant for conspiracy theory, ethno-nationalism, and authoritarian “us vs. them” thinking) with legitimate national emergency can and has provided opportunity for strong-man demagogues who promise anti-elite “authenticity” as well as protection and restoration for the common people (Hahl et al., Reference Hahl, Kim and Zuckerman Sivan2018; Levitsky and Ziblatt, Reference Levitsky and Ziblatt2018; Foertsch et al., Reference Foertsch, Chakraborty and Joose2024). In their recent survey of psychological literature in this area, Jost et al. (Reference Jost, Goya-Tocchetto and Kay2023) explain populist support for democratic backsliding seems to center around authoritarian aggression, system justification (i.e., support for the social status quo), and the appeal of group-based dominance. Those citizens most inclined toward aggression against those who deviate from established norms (immigrants; ethnic, sexual, or religious minorities) or toward viewing the “other” as culturally inferior and thus subordinate, are more inclined to support strong men who violate democratic norms, with Donald Trump being a recent example (Djupe et al., Reference Djupe, Lewis and Sokhey2023).
Given that the appeal of group-based dominance and system justification might motivate support for democratic backsliding (Jost et al., Reference Jost, Goya-Tocchetto and Kay2023), we should not suspect right-wing ideological commitments as the only source of support for leaders being granted stronger executive powers during national crises. Americans on the ideological or partisan left may also be more inclined to favor stronger executive power when the issues they care about—including their own political influence—are at stake. Democrats, ideological liberals, and racial minorities, for example, are more likely to favor the government doing more to solve problems that they see as particularly threatening for them and the communities they identify with (including crises like climate change, gun violence, and COVID-19), while those on the partisan right and White Americans are more likely to favor limited government involvement on such issues (Pew Research Center, 2019, 2020, 2021; Perry et al., Reference Perry, Whitehead and Grubbs2021). Thus, the idea of unspecified “national emergencies” may tap into this favorability to see executive power step in when Americans assume it is in their own group interests for their leaders to do so. This might also include when their own political group or their representatives are currently (but tenuously) in power.
Despite these helpful insights, previous research in this area has given little attention to the potential role of religio-political ideologies also connected to these relevant factors. It is here that we transition to discuss the possible role of Christian nationalism.
Theorizing Christian nationalism and other factors shaping support for violating democratic norms during national emergencies
Though ostensibly about the supremacy of “Christian” identity and values in the American civil sphere (based on actual items used to measure it), research has connected Christian nationalist ideology to a variety of ethno-nationalist and authoritarian populist impulses that would lead us to expect its influence on support for leaders violating democratic norms during national emergencies. One of the most common findings in research on Christian nationalism is that is associated with xenophobic and racist attitudes, particularly among White Americans (McDaniel et al., Reference McDaniel, Nooruddin and Shortle2011; Sherkat and Lehman, Reference Sherkat and Lehman2018), but also among non-White Americans depending on who is classified as the “other” (Perry et al., Reference Perry, Schleifer, Whitehead and Frantz2024a). Americans who subscribe to Christian nationalism are also more inclined to view democratic participation as a privilege (not a right), see their religious and ethnic in-groups as under assault, believe conspiracy theories like QAnon, and support the hypothetical use of force to respond to political threats, such any-means necessary policing, torture, the death penalty, and the January 6 Capitol Riots (Armaly et al., Reference Armaly, Buckley and Enders2022; Gorski and Perry, Reference Gorski and Perry2022; Davis et al., Reference Davis, Perry and Grubbs2024; Djupe et al., Reference Djupe, Lewis and Sokhey2023).
Related to the issue of supporting leaders violating democratic norms, in their analysis of support for democracy or authoritarianism, Drutman et al. (Reference Drutman, Diamond and Goldman2018) found that “cultural conservatives” and those who felt it was important to have European heritage to be American were especially likely to support “a strong leader who does not have to bother with Congress and elections.” Christian nationalism seems to represent a form of “(ethno)cultural conservatism” (what Gorski and Perry, Reference Gorski and Perry2022, 22 call “authoritarian ethno-traditionalism”) not only on issues of culture and race, but voter access and election legitimacy, as well as persistent support for strongman leaders Donald Trump and Vladimir Putin (Gorski and Perry, Reference Gorski and Perry2022; Perry et al., Reference Perry, Whitehead and Grubbs2022, Reference Perry, Riccardi-Swartz, Davis and Grubbs2023; Djupe et al., Reference Djupe, Lewis and Sokhey2023). Building on these ideas, we propose Christian nationalist views will likely capture much of the ethno-nationalist and authoritarian-populist impulses associated with support for violating democratic norms during national emergency.
Recent research, however, has argued that Christian nationalism is not exclusively tied to right-wing ethno-nationalism or populism, but ultimately involves the sacralization of one's own in-group. This could theoretically include those typically on the partisan or ideological left. Perry et al. (Reference Perry, Schleifer, Whitehead and Frantz2024a), for example, found that when Black and Hispanic Americans were asked questions in which they were the out-group, they were more likely to reject assimilationist and ethno-centric views. However, when Black and Hispanic Americans were asked questions that implied another out-group, Christian nationalism inclined to them to hold assimilationist or ethno-centric views similar to White Americans. They argue this suggests that Christian nationalism sacralizes the in-group vis-à-vis some “other.” Similarly, Perry et al. (Reference Perry, Shortle, McDaniel and Grubbs2024b) found that Black and Hispanic Americans who subscribe to Christian nationalism were more likely to identify as “progressive” and Black Americans were more likely to identify with the term “woke,” again suggesting Christian nationalism inclines Americans toward in-group solidarity and interest (see also Perry and Whitehead, Reference Perry and Whitehead2019). Given that possibility, we theorize that Christian nationalism might incline those Americans whose national leaders currently hold political power (in this case Democrats) to sacralize the status quo with their group atop the political hierarchy, and thus give greater justification for leaders (implicitly their leaders) violating democratic norms to preserve their interests.
Lastly, Drutman et al. (Reference Drutman, Diamond and Goldman2018) also found that nonvoters and those who rarely consume news were especially likely to support strong leaders who can disregard Congress and elections, possibly reflecting a populist discouragement or disillusionment with business-as-usual politics. Thus, we also seek to account for the influence of political disengagement or disinterest in our analyses using nationally representative data.
Methods
Data
Data for this study come from the 2022 National Addiction and Social Attitudes Survey (NASAS) (Perry et al., Reference Perry, Riccardi-Swartz, Davis and Grubbs2023, Reference Perry, Shortle, McDaniel and Grubbs2024b). The 2022 NASAS was designed by the authors and fielded in March 2022 by YouGov, an international research data and analytics company. YouGov recruits a panel of respondents through websites and banner ads. These respondents are not paid directly but are entered into lotteries for monetary prizes. In order to draw a nationally representative sample, YouGov employs a method called “matching.” Drawing a random sample from the American Community Survey, YouGov then matches a respondent in the opt-in panel who is the closest to the Census respondent based on key sociodemographic factors. Because of the specific recruitment and sampling design used by YouGov, the company does not publish traditional response rates. However, YouGov develops sampling weights in order to ensure that the survey sample is in line with nationally representative norms for age, gender, race, education, and census region. The resulting original survey sample included 2,809 Americans that were matched and weighted. After removing the few cases with missing information, the final analytic sample in full models is 2,802 cases. For descriptive statistics of all measures included in our analyses, see Table 1.
Table 1. Descriptive statistics
Source: National Addiction and Social Attitudes Survey, Wave 1 (N = 2,802).
Support for leaders violating democratic norms during national emergencies
The primary outcomes for this study are three measures of support for leaders violating democratic norms during unspecified “national emergencies.” We analyze these combined as an index and separately to confirm robustness. We asked Americans to indicate their level of agreement with the following statements: (1) “National emergencies might require leaders to suspend elections so they can lead the country to safety.” (2) “National emergencies might require leaders to control or suppress groups who threaten the leader's ability to lead the country to safety.” And (3) “National emergencies might require leaders to disregard checks and balances in order to lead the country to safety.” Respondents could answer from 1 = strongly disagree to 5 = strongly agree. The three items together make an index with a Cronbach's α of 0.79 (initially ranging from 0 to 12) indicating acceptable reliability. Though we focus on this index for assessing moderating effects, we also test models with each item separately so we can ensure that the patterns we observe are not due to unique associations with one or two outcome measures. Because of the nature of these outcome variables, we use ordinary least squares (OLS) regression models as our estimation strategy.
Key predictors
The primary predictor of interest is Christian nationalist ideology. “Christian nationalism” has been operationalized in a variety of ways, and there is a growing critical literature (re)assessing its conceptualization and measurement (e.g., Djupe et al., Reference Djupe, Lewis and Sokhey2023; Foertsch and Pieper, Reference Foertsch, Pieper and Hiebert2023; Li and Froese, Reference Li and Froese2023). We construct an index using responses to four questions with high face validity that have been used across various studies (McDaniel et al., Reference McDaniel, Nooruddin and Shortle2011, Reference McDaniel, Nooruddin and Faith Shortle2022; Davis and Perry, Reference Davis and Perry2021; Gorski and Perry, Reference Gorski and Perry2022; Djupe et al., Reference Djupe, Lewis and Sokhey2023; Vegter et al., Reference Vegter, Lewis and Bolin2023). Americans are asked to indicate their level of agreement with the following statements: “America holds a special place in God's plan.” “I consider founding documents like the Declaration of Independence and the US Constitution to be divinely inspired.” “The federal government should declare the United States a Christian nation.” And “I consider being a Christian an important aspect of being truly American.”Footnote 1 Respondents could answer from 1 = strongly disagree to 5 = strongly agree. We combined the four questions into an additive index of Christian nationalism, ranging from 0 to 16 (Cronbach's α = 0.89).Footnote 2
Because the “national emergencies” in our outcome questions are unspecified, we anticipate support for leaders violating democratic norms can come from the political left or right. Thus, we include measures for both ideological identity and partisan identity. Ideological identity is a continuous measure ranging from 1 = very liberal to 5 = very conservative, and thus higher scores indicate greater identification with ideological conservatism. Similarly, strength of partisanship is measured as a continuous variable ranging from 1 = Strong Democrat to 5 = Strong Republican. (Clearly these two measures are correlated. However, variance inflation factors for both were well under acceptable bounds and thus both were included in models separately.)
Following the findings of Drutman et al. (Reference Drutman, Diamond and Goldman2018), we also included a measure for how often respondents followed politics in the news as an indicator of political interest. Responses were coded into binary variables with “Most of the Time,” “Some of the time,” “Only now and then,” “Hardly at all,” and “Don't know.” Because we expect that those who are more disengaged from political news will be more supportive of leaders violating democratic norms, we use “Most of the time” (the most politically interested group) as the reference category.
Controls
Religious controls were also included to ensure Christian nationalism was not simply a proxy for religious commitment or theological conservatism (Jost et al., Reference Jost, Goya-Tocchetto and Kay2023). We measure religious tradition with multiple categories, including Evangelical Protestant (reference), Non-Evangelical Protestant, Catholic, Other Christian, Non-Christian Religion, Atheist, Agnostic, and Nothing in Particular.Footnote 3 Religious commitment is measured by standardizing religious service attendance (1 = never to 6 = several times a week), prayer (1 = never to 7 = several times a day), and religious importance (1 = not at all important to 4 = very important) and combining them into an additive index (Cronbach's α = 0.85).
We also included a number of standard demographic controls for racial identity (White = reference, Black, Hispanic, Other Race), age (19–96), gender identity (man = reference, woman, non-binary), educational attainment (1 = less than high school, 6 = graduate school), household income (less than $30 K per year, $30–$60 K, $60–$100 K, $100–$200 K, $200 K or more, did not answer) and Southern residence.
Plan of analysis
For multivariate analyses, all variables were rescaled to range from 0 to 1, which makes interpretation more intuitive (see Table 1 for descriptive statistics with both ways of coding). Full OLS regression results are presented in Tables 2 and 3. These include models predicting our norm violation index and each outcome separately (Table 2) as well as models for interaction effects (Table 3). For simplicity's sake, in our results section below we focus our attention on forest plots and marginal effects presented in Figures 1–4. All our analyses use robust standard errors because of the survey weights for the NASAS, which sets a higher bar for obtaining statistical significance.
Figure 1. Predicted marginal effects of key predictors on indicators of support for leaders violating democratic norms during national emergencies.
Source: National Addiction and Social Attitudes Survey, Wave 1.
Note: Results correspond to full models in Table 2. All variables range from 0 to 1. Error bars are 95% confidence intervals.
Figure 2. Predicted marginal effects of partisan identities on support for leaders violating democratic norms across values of Christian nationalism.
Source: National Addiction and Social Attitudes Survey, Wave 1 (N = 2,802).
Note: Results from model 1, Table 3. All variables range from 0 to 1. Controls held at their means. Error bars are 95% confidence intervals.
Figure 3. Predicted marginal effects of Black or White racial identities on support for leaders violating democratic norms across values of Christian nationalism.
Source: National Addiction and Social Attitudes Survey, Wave 1 (N = 2,802).
Note: Results from model 2, Table 3. All variables range from 0 to 1. Controls held at their means. Error bars are 95% confidence intervals.
Figure 4. Predicted marginal effects of political news interest on support for leaders violating democratic norms across values of Christian nationalism.
Source: National Addiction and Social Attitudes Survey, Wave 1 (N = 2,802).
Note: Results from model 4, Table 3. All variables range from 0 to 1. Controls held at their means. Error bars are 95% confidence intervals.
Table 2. Ordinary least squares regression models predicting support for leaders violating democratic norms during national emergencies
RSE, robust standard errors.
Source: National Addiction and Social Attitudes Survey, Wave 1.
Note: Excluded categories are Most of the time, Man, White, Less than $30K per year, and Evangelical Protestant.
*p < 0.05, **p < 0.01, ***p < 0.001 (two-tailed tests).
Table 3. Ordinary least squares regression models predicting support for leaders violating democratic norms during national emergencies with interaction terms
RSE, robust standard errors.
Source: National Addiction and Social Attitudes Survey, Wave 1 (N = 2,802).
Note: Excluded categories are Most of the time, Man, White, Less than $30K per year, and Evangelical Protestant.
*p < 0.05, **p < 0.01, ***p < 0.001 (two-tailed tests).
Results
As can be seen from results presented in Table 2, whether our three outcomes of interest are predicted separately or as an index, Christian nationalism is a positive, significant predictor of each. Moreover, standardized regression coefficients also presented in Table 2 show Christian nationalism is by far the strongest predictor in the models (which is also reflected in the unstandardized betas because all variables have been scaled to range from 0 to 1).
Figure 1 presents four forest plots showing the marginal effects for key predictors on the three-item norm violation index and each item predicted separately. Effects to the left of the red “null effect” line are negatively associated with the outcome; effects to the right are positively associated. If the error bars (95% confidence intervals) do not touch the red line, they are statistically significant associations. Importantly, while Christian nationalist ideology shows the strongest positive association with any indicator of support for leaders violating democratic norms during national emergencies, identifying with Republican partisanship is negatively associated with each outcome (the effects of conservative ideological identity wash out with Republican partisanship included in models).Footnote 4 Moreover, Americans who follow political news less than “most of the time” are also more likely to support leaders violating democratic norms across all indicators.
Though not presented in the forest plots, the results in Table 2 show Black Americans and women compared to White Americans and men, respectively, are also more likely to support leaders violating democratic norms during national emergencies. Conversely, those with higher levels of education and higher incomes are less likely to support leaders violating democratic norms, while Catholics and Americans who affirm “Nothing in Particular” (in the full index model) are more likely to support norm violations compared to Evangelical Protestants.
Taken together, these patterns suggest different ideological and contextual factors contributing to support for leaders violating democratic norms during unspecified “national emergencies.” The strongest predictor, Christian nationalism, is typically understood to represent an ethno-nationalist and authoritarian populist element from the right, as might a lack of interest in political news (suggesting disillusionment with national politics characteristic of populist movements), while Democratic partisanship, being Black, or a woman, would suggest other contributing lines of thought, possibly reflecting the fact that a Democratic President was in office when the survey was conducted. How do these factors intersect with one another? As shown in Table 3, we tested interactions between Christian nationalism and partisanship, racial identity, gender identity, and political (dis)interest. Because interaction effects can be difficult to interpret we turn our attention to marginal effects presented in Figures 2–4.
Figure 2 shows the interactions by partisanship. First, it is clear that, regardless of partisan identity, Christian nationalism inclines Americans toward greater support for leaders violating democratic norms during national emergencies. In order to test that fact more directly, we estimated models for Democrats, Republicans, and Independents separately and found Christian nationalism was a significant positive predictor of our three-item index for each group (see online Appendix Table A2). However, as Figure 2 shows, while even strong partisans are statistically indistinguishable from one another at lower levels of Christian nationalism (completely overlapping confidence intervals), greater adherence to Christian nationalism amplifies support among Strong Democrats more than Independents and Strong Republicans.Footnote 5
We see a similar pattern for Black and White Americans in Figure 3. To be sure, Christian nationalism seems to incline White and Black (and Hispanic) Americans toward greater support for leaders violating democratic norms (which we confirm with separate models for each racial group in online Appendix Table A3). However, the interactions presented in Figure 3 show the slope is somewhat steeper for Black Americans, suggesting it is more strongly associated with their support. The pattern for Hispanic Americans was not substantively different from that of White Americans.
Moreover, as our interaction terms reflect in Table 3 (model 3), this racial variation is not reflected for women and men. Though women were more likely than men to support violating democratic norms during national emergencies, their slopes do not vary and they are statistically indistinguishable from one another at extreme ends of Christian nationalism (see the plotted average marginal effects in online Appendix Figure A1).
Lastly, in Figure 4 we look at how different levels of political news interest correspond to support for leaders violating democratic norms across levels of Christian nationalism. Once again, as Christian nationalism increases, support for leaders violating democratic norms during national emergencies increases for all groups. However, those who follow political news “most of the time” start relatively lower and their slope stays relatively shallow. However, among those Americans who are less engaged in political news (most particularly those who say they follow the news “Hardly at all”), their support for leaders violating democratic norms increases more sharply as Christian nationalism increases.
Ancillary analyses
Though the interaction effect for Christian nationalism and political disinterest makes sense in light of our populism argument, we suspect the positive interaction for Christian nationalism and Democratic partisanship or being Black (vs. White) makes more sense in light of arguments that Christian nationalism inclines Americans to sacralize their own in-group and thus justify extreme means to maintain a political status quo. Supporting this idea, we estimated regression models predicting partisan identity in Table 4 with interaction terms for Christian nationalism × racial identity. The interaction terms are all statistically significant and when we plot the marginal effects in Figure 5, we see that while Christian nationalism corresponds to stronger Republican partisanship for White Americans, Christian nationalism corresponds to stronger Democratic partisanship for Black Americans. Thus, the fact that Christian nationalism had a stronger association with support for leaders violating democratic norms during national emergencies among Black Americans (compared to White Americans) may have been due to the fact that a Democratic leader was in office, and Christian nationalism inclines Black Americans toward stronger Democratic partisanship, thus inclining them toward greater trust in their own leader (at the time Joe Biden) to violate democratic norms if necessary.
Table 4. Ordinary least squares regression models predicting republican partisanship with interactions for racial identity
RSE, robust standard errors.
Source: National Addiction and Social Attitudes Survey, Wave 1 (N = 2,805).
Note: Excluded categories are Most of the time, Man, White, Less than $30K per year, and Evangelical Protestant.
*p < 0.05, **p < 0.01, ***p < 0.001 (two-tailed tests).
Figure 5. Predicted marginal effects of racial identity on partisanship across values of Christian nationalism.
Source: National Addiction and Social Attitudes Survey, Wave 1 (N = 2,805).
Note: Results from model 2, Table 4. All variables range from 0 to 1. Controls held at their means. Error bars are 95% confidence intervals.
Discussion and conclusions
The tendency for democratically elected leaders to request temporary unilateral authority to lead their nations through crises is not new. Nor is it always disastrous for democracy (e.g., Lincoln, FDR). But as history has too often shown, it does involve the risk that authoritarian leaders will use their opportunity to permanently subvert democratic processes, sideline opposition, and entrench themselves in power (Paxton, Reference Paxton2004; Levitsky and Ziblatt, Reference Levitsky and Ziblatt2018). Given that possibility, this study examined what ideological and social characteristics might incline Americans to support leaders violating democratic norms during hypothetical and unspecified “national emergencies,” ostensibly to lead the country to safety. Examining three outcomes separately or combined, we find there are different factors that can work in combination.
First and foremost, Christian nationalist ideology (reflecting both a “deep story” of America's sacred origins and a “vision” that involves formally privileging Christian identity; see Gorski and Perry, Reference Gorski and Perry2022) is the leading predictor that Americans support leaders suspending elections, suppressing political opponents, and disregarding checks and balances in order to lead the country to safety. Moreover, across various moderating factors, we found no situation in which Christian nationalist ideology made Americans less inclined toward supporting leaders violating democratic norms. Given that previous research has found ethno-nationalistic and authoritarian populist ideology inclines citizens to support democratic backsliding and strongman leadership (Drutman et al., Reference Drutman, Diamond and Goldman2018; Berlucchi and Kellam, Reference Berlucchi and Kellam2023; Jost et al., Reference Jost, Goya-Tocchetto and Kay2023), we propose that Christian nationalist ideology captures these impulses. This would be entirely consistent with previous research on Christian nationalism showing that it is associated with anti-democratic, conspiratorial, ethnocentric, and authoritarian views, of the kind that fuel right-wing populist movements (McDaniel et al., Reference McDaniel, Nooruddin and Shortle2011; Armaly et al., Reference Armaly, Buckley and Enders2022; Gorski and Perry, Reference Gorski and Perry2022; Perry et al., Reference Perry, Whitehead and Grubbs2022, Reference Perry, Schleifer, Whitehead and Frantz2024a; Davis et al., Reference Davis, Perry and Grubbs2024; Djupe et al., Reference Djupe, Lewis and Sokhey2023).
Yet Christian nationalism was not the only contributing factor. In what also may reflect roots of populist sentiment, we find Americans who are more disengaged from political news are also more likely to support leaders violating democratic norms. Indeed, as we showed in Figure 1 (see also Table 2), marginal political interest was often among the strongest predictors of our outcomes, second only to Christian nationalism. The two strongest predictors of support for leaders violating democratic norms during “national emergencies,” in other words, were belief that the nation has and should always privilege a particular religio-political identity, on the one hand, and disengagement from politics, on the other. And interactions presented in Figure 4 (Table 3) showed these factors amplified one another. These findings are consistent with those of Drutman et al. (Reference Drutman, Diamond and Goldman2018) and we propose the strong effect of political disinterest also reflects an underlying element of populism, namely, frustration and disillusionment with politics in general and including democratic institutions. This may also help explain why persons with lower levels of education and lower incomes were also more likely to support leaders violating democratic norms (see Table 2). People who are socially and economically marginalized and disconnected from politics may find the idea of strong leaders “getting the job done” more attractive than preserving messier, slower, business-as-usual democratic politics (see Hochschild, Reference Hochschild2018).
Though the influence of Christian nationalism and political disinterest can be connected with ethno-nationalist and authoritarian populism on the right, Democratic partisanship, being Black (vs. White), and being a woman (vs. man) were also positively associated with support for leaders violating democratic norms. Given that Democratic partisanship in particular is negatively associated with Christian nationalism (r = −0.43, p < 0.001 in our data set), we propose these patterns reflect a different underlying factor leading to support for such undemocratic arrangements. Jost et al. (Reference Jost, Goya-Tocchetto and Kay2023) explain that, in addition to authoritarian aggression, support for the status quo and system justification incline people to support democratic backsliding. In this case, the fact that a Democratic President currently held power at the time of the survey might incline Democrats and those who often identify with the Democratic Party (Black Americans and women in particular) to support their implied leader (Joe Biden) violating democratic norms to help the nation through crisis. In addition, our hypothetical “national emergencies” were unspecified, and thus, it is possible that stronger Democrats and Black Americans assumed “national emergencies” like the Capitol Riots of January 6, or Hurricane Katrina, or the COVID-19 pandemic, which they would find more threatening than their counterparts on the partisan right or White Americans, respectively.
Importantly, however, while the influence of Democratic partisanship and racial minority status likely represent different ideological sources than Christian nationalism, their associations with support for leaders violating democratic norms were also amplified by Christian nationalism. We propose this reflects Christian nationalism's tendency to sacralize in-group power, whatever the group (Perry et al., Reference Perry, Schleifer, Whitehead and Frantz2024a, Reference Perry, Shortle, McDaniel and Grubbs2024b). Thus, because Democrats and Black Americans might have been more inclined to support their national leader violating democratic norms during national emergency, adhering to Christian nationalism could have made them even more likely to do so. This idea is supported by our analysis in Table 4 (Figure 5) where we found the more Black Americans subscribe to Christian nationalism the more likely they were to be stronger partisan Democrats (and as we found, stronger partisan Democrats were more likely to support leaders violating democratic norms). These patterns are somewhat inconsistent with Djupe et al. (Reference Djupe, Lewis and Sokhey2023) who argue that Christian nationalism represents an essentially “Republican project” and thus Democrats who score higher on Christian nationalism would be predicted to be less partisan. While this may often be the case, it may be more essential to Christian nationalism to support whatever in-group respondents have in view. Future research should further adjudicate between these two lines of argument.
On the subject of future research, several data limitations are worth acknowledging to chart a path for future studies. First, the outcome measures only ask about hypothetical “national emergencies.” This was intentional to remove possible partisan or left-right coding and assess Christian nationalism's association with support for leaders violating democratic norms to address a “crisis.” But it would be helpful for future studies on this topic to ask about specific national emergencies such as climate change, COVID-19, a terrorist or military attack from a foreign power, domestic rioting, a Constitutional crisis or coup from a domestic political actor or party, or a catastrophic natural disaster. It is possible that the partisan dynamics we observed as well as even the structure of the association with Christian nationalism would vary contingent on how the “crisis” is coded in partisan or left-right terms. Related to this, it would be important to replicate these same questions when the Republican party is in the White House to see if the patterns we observed among Democrats and Black Americans in particular (including their interactions with Christian nationalism) remain or reverse.
Second, while we use “Christian nationalism” items that correlate strongly and create a reliable measure (one in which all items predict the outcome individually in nearly identical ways; see Appendix Table A1), the concept itself has spawned not only a burgeoning literature, but it has made its way into popular vernacular (Perry, Reference Perry2024). With both developments, there is the possibility of overgeneralization and the canonization of measurements before it is scientifically justified (see Li and Froese, Reference Li and Froese2023). The very fact that Christian nationalism has been operationalized in a variety of ways with consistent effects (McDaniel et al., Reference McDaniel, Nooruddin and Shortle2011; Gorski and Perry, Reference Gorski and Perry2022; Perry et al., Reference Perry, Whitehead and Grubbs2022, Reference Perry, Schleifer, Whitehead and Frantz2024a; Vegter et al., Reference Vegter, Lewis and Bolin2023) validates the construct, but also suggests there is ongoing need for greater measure-testing and development. Other recent studies are testing even more precise measurements (e.g., Perry et al., Reference Perry, Shortle, McDaniel and Grubbs2024b) and we encourage further efforts toward that end.
Lastly, future studies should continue to explore the intersection of these factors in contributing to Americans' potential support for leaders violating democratic norms. Though this study accounted for demographic characteristics like income and educational attainment, more direct questions about economic frustrations could be used in future studies, along with more direct measures about support for a stronger central government and populist frustration with the political establishment. Other important factors worth considering would be indicators of social dominance orientation and perceptions of group persecution, which Djupe et al. (Reference Djupe, Lewis and Sokhey2023; see also Jost et al., Reference Jost, Goya-Tocchetto and Kay2023) have shown to amplify Christian nationalism's influence on certain political issues. Though it is unlikely these characteristics would wash out the effects of Christian nationalism, which seems to come from different cultural concerns, it would be important to observe whether these characteristics also work in combination to shape Americans' support for forms of democratic backsliding.
Supplementary material
The supplementary material for this article can be found at https://doi.org/10.1017/S1755048324000208.
Competing interests
None.
Data availability
Data and code for replication will be made available.
Samuel L. Perry is the Sam K. Viersen Presidential Professor of Sociology at the University of Oklahoma. His work examines links between American religion and politics, race, families, and sexual behavior. His forthcoming book is entitled Religion for Realists: Why We All Need the Scientific Study of Religion (Oxford, 2024).
Joshua B. Grubbs is an Associate Professor of Psychology and an investigator at the Center on Alcohol, Substance Use, and Addictions at the University of New Mexico. Broadly interested in compulsive sexual and gambling behavior, Grubbs's work focuses largely on the social construction of “addiction” and its consequences for personal well-being.
Open access
