Polarization between Democrats and Republicans has not produced unity within the contending parties. Divides within each party are almost as deep as those between them, and these divides are not just policy differences, but disagreements about the future of democracy in the United States. Majorities among White Christians and Republicans believe that the “American way of life” has broken down. According to the Public Religion Research Institute, nearly a third of White evangelicals believe that violence may be necessary to put the country back on track.Footnote 1
This paper examines the divide among Republican and Republican-leaning pastors on the legitimacy of the use of force to achieve political goals. North Carolina is an instructive case (Rozell and Wilcox, Reference Rozell and Wilcox1997; Cooper and Knotts, Reference Cooper and Knotts2008; Ferguson, Reference Ferguson2017; Gilmore, Reference Gilmore2019). The hard right turn of North Carolina's Republican Party in the 2010s drew heavily on the moral influence of evangelical churches, two-thirds of which are affiliated with the Southern Baptist Convention (Tervo, Reference Tervo, Skocpol and Tervo2020, 56, 73, note 10). Yet, even the most conservative denominations, such as the Southern Baptist Convention and the Baptist State Convention, are riven on gender and LGBTQ+ rights, critical race theory, and perhaps most of all, the legitimacy of the 2020 presidential election. Beneath divisions on issues runs a more profound divide over core questions of democratic legitimacy and the rule of law (Bartels, Reference Bartels2020; Helmke et al., Reference Helmke, Kroeger and Paine2022). While most pastors accept competition and compromise within democratic channels, others believe it is justifiable to use extra-legal means on the ground that “we may have to use force to defend the traditional American way of life.”
We examine this divide using an original survey of 145 North Carolinian pastors who lead Baptist, Methodist, and Presbyterian churches and identify as Republican or Republican-leaning. We complement the survey with semi-structured interviews to probe individual motivations more deeply. Contrary to what one might expect, we do not find that theological fundamentalism—operationalized here as biblical literalism—predisposes pastors to support the use of force. Instead, we find that Christian nationalists and those who harbor systemic White grievance are most likely to support the use of force. We also find that distrust of elections intensifies the effect of these attitudes.
Intraparty polarization has received limited scholarly attention by comparison to the flow of research on intergroup polarization. However, such conflict can be severe and consequential. Our focus here is on intraparty conflict among Christian community leaders who support or lean toward the Republican Party. While some scholars have argued that Christian nationalism is contributing to polarization between parties (Gorski and Perry, Reference Gorski and Perry2022), we contend that Christian nationalism, among other socio-political factors, also divides the party where this ideology is strongest.
Christianity and American politics
There is renewed interest in the role of Christianity in American democracy in light of the January 6 insurrection. Clergy face complex decisions about whether to engage political topics from the pulpit (Boussalis et al., Reference Boussalis, Coan and Holman2020), how to arbitrate political conflict among their congregations (Glazier, Reference Glazier2015; Hafner and Audette, Reference Hafner and Audette2022), and where to stand on the issues that divide their association (Ammerman, Reference Ammerman1990; Lewis, Reference Lewis2014).
Research in this area is challenging because conservative religious pastors are often unwilling to participate in academic surveys. Perceptions of social scientists can make it difficult to build trust for honest disclosure of politically sensitive opinions. However, gaining such participation is valuable both for understanding religious conservatism and the nature, extent, and sources of intraparty division among opinion leaders.
Our inquiry focuses on Republican and Republican-leaning Protestant pastors in North Carolina because of their strategic position in navigating political discourse in their congregations at a time of transformational change in the Republican Party. Research suggests that, over the past few decades, clergy have increasingly incorporated political messages in their work (Guth et al., Reference Guth, Green, Smidt, Kellstedt and Poloma1997; Smidt, Reference Smidt2016; Guth and Smidt, Reference Guth and Smidt2021). While their direct impact on congregant political attitudes through sermons appears circumscribed, their indirect influence by virtue of being “the principal architects in creating the small group infrastructure of congregations” is substantial (Djupe and Gilbert, Reference Djupe and Gilbert2008, 244–45). It is in these congregant groups where theological and political norms are enforced and where recruitment into politics may take place (Bean, Reference Bean2014).
Compared to lay congregants, clergy have more coherent belief structures and they also sort sharply along partisan lines (Malina and Hersh, Reference Malina and Hersh2021). Republican pastors might be expected to have similar worldviews. Almost all were born into Christian households. Those in the sample we consider live in NC, and most were raised there. They are overwhelmingly male, White, and middle-aged or older. Yet, as we show in this paper, they are divided on one of the most urgent questions of our time: the legitimacy of force in achieving their goals.
In the next section we set out expectations relating the use of force to four factors: theological beliefs, Christian nationalism, White grievance, and electoral distrust. We then introduce original data to evaluate our expectations along three paths. We begin with a multivariate logistic analysis which allows us to evaluate the predictive power of each of these factors under controls. We then employ mediation analysis to probe the causal pathways. Finally, we use conjoint analysis and evidence from interviews to provide nuance about what motivates pastors, paying special attention to Christian nationalism and White grievances.
Expectations on the use of force
We theorize three sources of division among Republican pastors on the use of force. One problematizes whether a Christian life can be meaningful if the United States remains a pluralistic country where church and state are separate: How fundamentalist and nonnegotiable is the version of Christian life that Republican pastors champion; and how vital is it for them to impose their values on the country to forge an expressly Christian nation? The second source of division concerns the fears of White conservatives that they are losing status and hegemony over public affairs. This debate takes place in the shadow of absolute and relative White population decline, the rise of Black Lives Matter, and out-group antagonism. Christian nationalism and White grievance bear directly on attitudes over force for Republican pastors who believe that these goals are nonnegotiable yet impossible to achieve in a pluralistic democracy. The third source of division concerns distrust of the rules of the game, and more precisely, whether elections can be fair under conditions of high affective polarization. As the aftermath of the 2020 election demonstrated, those who saw “their party” defeated may falsely conclude that elections are fraudulent, which sets the door further ajar for extra-legal behavior.
Biblical literalism
A willingness to use force may rest on the conviction that one's beliefs are nonnegotiable, that is, absolute and impregnable to dialogue. Applied to Christianity, this is most clearly expressed in biblical literalism, the theological view that the Bible is the authoritative and complete word of God, and consequently an unchallengeable source of truth (Smith, Reference Smith2011). Survey research has found that biblical literalists tend to hold their views more dogmatically and embrace absolute conceptions of right and wrong (Pew, 2008; Meulemann, Reference Meulemann2013, note 4). In the words of Hannah Arendt (Reference Arendt1978, 181–82) this mindset is one of “God-like certainty which stops all discussion.”
Whether this also actually paves the way to condoning force is contested. Studying literalism in Islam, Matusitz (Reference Matusitz2015, 183) claims that religious literalism sustains religious violence because it offers “no prospect for any compromise or other types of negotiation,” and Collins (Reference Collins2003, 21) points out that the same can apply to Christianity: “The Bible has contributed to violence in the world precisely because it has been taken to confer a degree of certitude that transcends human discussion and argumentation.” Against this stands the view that a dogmatic belief can be introspective, a personal code intruding little into the wider society. Historically, many evangelical churches, where biblical literalism is entrenched most deeply, chose to focus on spiritual wellbeing on the sidelines of public life (Gorski, Reference Gorski2020). From that perspective, while a theological fundamentalist is more likely to consider his beliefs as nonnegotiable, this does not necessarily imply that he wishes to forcefully impose his beliefs on others.
H1a: Biblical literalist beliefs increase the likelihood of finding the use of force legitimate.
H1b: Biblical literalist beliefs do not increase the likelihood of finding the use of force legitimate.
Christian nationalism
The prospect of force grows when a minority belief becomes impositional. Christian nationalism has two key features that might drive its followers to legitimize the use of force to achieve political goals: it is a minority ideology in the face of declining evangelical observance, and it asserts a code for the society as a whole.
The core of Christian nationalism is that the United States is a country with Christian heritage, values, and calling (Whitehead and Perry, Reference Whitehead and Perry2020; Davis, Reference Davis2023, 3–4). It is incumbent upon its defendants to “take America back for God” from liberal forces by imposing a traditionalist (White and patriarchal) view of racial order and social hierarchy on American public policy, public symbols, and national identity (Gorski and Perry, Reference Gorski and Perry2022; Perry, Reference Perry2022, 94). In contrast to biblical literalism, Christian nationalism's theology is less about morality or virtue and more about political power and who should wield it (Davis, Reference Davis2023, 3). Here we consider Christian nationalism as both an explicitly governmental project (Davis, Reference Davis2023; Vegter et al., Reference Vegter, Lewis and Bolin2023) and an ideology that imposes a religious-nationalist public sphere (Whitehead and Perry, Reference Whitehead and Perry2020; Djupe et al., Reference Djupe, Lewis and Sokhey2023).
In both respects, Christian nationalists face an uphill struggle in an increasingly diverse, secular United States. White evangelicals in particular have declined sharply—from 23% of the adult population in 2006 to 13.6% in 2022—and their adherents are now the oldest religious group in the country. The largest and youngest group are the so-called “nones,” the religiously unaffiliated, who in 2022 made up 26.8% of the adult population.Footnote 2 Campbell et al. (Reference Campbell, Layman and Green2021) describe this as the Secular Surge—deeply threatening from the standpoint of Christian nationalism.
Christian nationalism became aligned with right-wing Republicanism in reaction to a series of perceived defeats—the Civil Rights Movement, the progressive 1960s, growing secularism (Putnam and Campbell, Reference Putnam and Campbell2010; Gorski, Reference Gorski2020), and—with enhanced significance—electoral victory and rule by a Black president (Hooghe and Dassonneville, Reference Hooghe and Dassonneville2018). Under the Trump administration, Christian nationalism has become more deeply intertwined with the Republican Party (Djupe et al., Reference Djupe, Lewis and Sokhey2023). Pastors steeped in Christian nationalism became “local captains for culture war politics” motivated by a “thick coherence” between religious and political conservatism (Bean, Reference Bean2014, 133).
Christians who favor the integration of Christianity and governance have, for many years, been more active in mobilizing support than those who are opposed. However, in recent years, Christian opponents of Christian nationalism have begun to mobilize, too, and in 2019, hundreds of pastors with support from the Baptist Joint Committee for Religious Liberty endorsed Christians Against Christian Nationalism.Footnote 3 Following January 6, 2021, nearly 3000 evangelical pastors released an open letter against Christian nationalism's role in the insurrection.Footnote 4 Given its current salience, we expect Christian nationalism to be a critical factor for understanding division among Republican pastors on the legitimacy of force.
H2: Christian nationalist beliefs increase the likelihood of finding the use of force legitimate.
White grievance
The second source of division among Republican pastors concerns the fear of Whites losing status and hegemony in public affairs. Bartels (Reference Bartels2020, 3) finds that antidemocratic sentiment among Republican voters is predicted by a latent dimension, ethnic antagonism, summarizing “not only unfavorable feelings toward Muslims, immigrants, and other out-groups, but also—and especially—concerns about these groups’ political and social claims.” Kalmoe and Mason (Reference Kalmoe and Mason2022) observe that, among White Republicans, vilifying one's political opponents as a threat to society and having potentially violent attitudes toward them are predicted by racial resentment.
White grievance grew as progressive federal policies challenged the racism of daily routine such as taking the bus, eating in a diner, or sending one's children to school. Initially, resentment was voiced in churches, chambers of commerce, and country clubs, and in the South, it became the credo of many local and state governments. As the federal government became an active player, White identity was reconfigured as White victimhood by Jerry Falwell, NC native Franklin Graham, and other conservative evangelicals (Tarrow, Reference Tarrow2022).
Falwell and his allies coined the term “Moral Majority” to present a united front opposing liberalism, secularism, and compulsory desegregation. However, White victimhood is divisive among conservative Christians. At the 2021 Southern Baptist Convention, the conservative leadership narrowly fought off an insurgency seeking to denounce critical race theory (Graham, Reference Graham2021), and a majority re-affirmed its landmark resolution of 1995 that includes the language: “We apologize to all African Americans for condoning and/or perpetuating individual and systemic racism in our lifetime; and we genuinely repent of racism of which we have been guilty, whether consciously (Psalm 19:13) or unconsciously (Leviticus 4:27).”Footnote 5
The chief circumstance that may help propel those harboring White grievance to the use of force is a perception that things are bound to get worse. Several concerns converge: the belief that secularism, LGBTQ+ rights, and critical race theory are diluting American values; that “Blacks, women, immigrants, refugees, brown pelicans—all have cut ahead in line” of conservative White families (Hochschild, Reference Hochschild2016, 137–38); and that Whites are marginalized in a country they should own. Such attitudes, which we label “White grievance,” may justify the use of force.
H3: White grievance beliefs increase the likelihood of finding the use of force legitimate.
Electoral distrust
Finally, we theorize that fear of democratic failure can be a powerful motivation for extra-legal means. Distrust in elections appears to intensify for those who suffer electoral loss (Grant et al., Reference Grant, Flores, Pedersen, Sherman and Van Boven2021; Kalmoe and Mason, Reference Kalmoe and Mason2022), those living under an out-party government (Graham and Svolik, Reference Graham and Svolik2020), and those susceptible to unsubstantiated claims of election fraud (Berlinski et al., Reference Berlinski, Doyle, Guess, Levy, Lyons, Montgomery, Nyhan and Reifler2023). Moreover, once in place, these beliefs appear resistant to change among partisans (Van Bavel and Pereira, Reference Van Bavel and Pereira2018; Botvinik-Nezer et al., Reference Botvinik-Nezer, Jones and Wager2023). On the one hand, electoral distrust may directly lead to a conviction that force is warranted; on the other, it may mediate the effect of Christian nationalism or White grievance. A person who believes that Christian nationalism or White grievance is existential but the project has no fair chance of winning at the ballot box may be pushed to endorse extra-legal means.
H4a: Electoral distrust increases the likelihood of finding the use of force legitimate.
H4b: Electoral distrust mediates the effect of Christian nationalism and White grievance on the likelihood of finding the use of force legitimate.
Data and methods
The sample
The data are derived from the authors' online survey among pastors of Protestant denominations in North Carolina. North Carolina is instructive for evaluating the intersection between Republicanism and organized religion. It is a textbook case for systematic coordination between Republican donors, Christian conservative advocacy groups, and evangelical churches (Tervo, Reference Tervo, Skocpol and Tervo2020), and it became, alongside WI, an influential battleground state in which the Republican Party shifted to the hard right (Skocpol and Hertel-Fernandez, Reference Skocpol and Hertel-Fernandez2016; Grumbach, Reference Grumbach2023).
The sample is constructed through purposive sampling, “a selection method where the study's purpose and the researcher's knowledge of the population guide the process” (Tansey, Reference Tansey2007, 770). Given the sensitivity of religion and politics in the United States and of the cultural distance in the public imagination between UNC-Chapel Hill and evangelical pastors, we wished for both scientific and ethical reasons to build trust with a community where we are outsiders. Had we used sampling methods with more indirect contact, such as working with a third-party firm to distribute the survey, we would have expected greater skepticism of the research and fewer responses. We prioritized depth of thoughtful responses over breadth of population coverage by establishing rapport with pastors in leadership positions prior to distributing the survey. Three channels were used to build the pastor sample. First, we contacted the leaders of the conferences or associations of the larger denominations, namely the Baptist State Convention of North Carolina, the United Methodists, the Cooperative Baptist Fellowship, and the four Presbyterian associations (Associate Reformed Presbyterian Church, Evangelical Presbyterian Church, Presbyterian Church in America, and Presbyterian USA). Those who consented to meet in person or more usually over Zoom were asked to distribute a personalized secure survey link to their individual members. We reasoned that gaining the cooperation of the major associations would increase the survey's coverage and response rate. Second, where the state-wide associations declined to collaborate, we contacted the regional leaders of denominational associations and chapters, including 76 lower-tier associations of the largest association, the Baptist State Convention of North Carolina. Around one-third cooperated by sending the survey link to their pastors. And finally, we used Google Maps to identify and send direct emails (with two reminders) to 523 nondenominational churches, Black churches, and smaller denominations (such as Pentecostal churches) using email addresses from their official websites. The survey was in the field from June through October 2021 and produced 467 responses; 134 of whom provided contact information for a subsequent interview. This is around 4–10% of the estimated 5,000 to 10,000 Protestant pastors in North Carolina. As expected, this group is somewhat biased toward the larger associations. Table 1 compares the sample distribution with the distribution of Christian adults in North Carolina as estimated by the most recent Religious Landscape Study by the Pew Foundation.Footnote 6
Table 1. Distribution of pastors across denominations compared to North Carolina churchgoers
Note: The evangelical/mainline distinction cuts across denominations. In NC, most Baptists and Pentecostals consider themselves evangelical and most United Methodists and Presbyterians consider themselves mainline.
Source: Percentages for the NC population are calculated from https://www.pewforum.org/religious-landscape-study/state/north-carolina/.
The focus of this study is on the subset of Protestant pastors who are Republican or Republican-leaning, nearly all of whom are socially and theologically conservative. One hundred and forty-five pastors answered that they were closer to the Republican Party, and 143 of these responded to the question that informs our dependent variable: the use of force to defend the traditional American way of life. As Table 2 reveals, the sample, as expected, is predominantly male, White, rural, educated, and middle-aged. It is worth noting at the outset that the usual drivers of political division will not help us explain the division on force because our sample takes them off the table: education, occupation, social status, race or ethnicity, partisanship, and living in the American South. To the extent that our sample is biased toward moderate Republicans, so the sample will under-estimate the willingness to legitimize force.Footnote 7
Table 2. Demography of Republican pastors in the sample
Note: Data from the authors' survey.
Variables
The dependent variable is the response to a statement adapted from Bartels (Reference Bartels2020): “The traditional American way of life is disappearing so fast we may need force to save it,” a four-category item ranging from strongly agree to strongly disagree. We dichotomize the variable and reverse the order, so that the variable takes on a value of 1 if pastors agree strongly or somewhat, and a value of zero if they disagree strongly or somewhat.
The chief independent variables are biblical literalism, Christian nationalism, White grievance, and election distrust (see the online Appendix for wording).
• Biblical literalism is a principal-components factor of two items that probe creationist beliefs and belief in the Scripture as the actual word of God (Smidt, Reference Smidt2016; Oberlin and Scheitle, Reference Oberlin and Scheitle2019; Margolis, Reference Margolis2020). Higher values indicate biblical literalism. Results are similar with a simplified measure that uses just belief in Scripture as the actual word of God.
• Christian nationalism is an additive 0–18 scale derived from six four-category statements about the United States as Christian nation, U.S. success as part of God's plan, federal government advocacy of Christian values, prayer in public schools, religious symbols in public places, and separation of church and state (reversed) (Whitehead et al., Reference Whitehead, Perry and Baker2018; Whitehead and Perry, Reference Whitehead and Perry2020). Higher values indicate higher Christian nationalism. Measurement of Christian nationalism is an ongoing debate, with some scholars favoring the Whitehead and Perry scale and others advancing narrower conceptualization and measurement. Djupe et al. (Reference Djupe, Lewis and Sokhey2023) find that explicitly cueing Christianity in the three Whitehead and Perry items that use “religion” makes little substantive difference in effects, and they further observe overall consistency of the effects of Christian nationalism on political and religious behavior across 1-, 2-, 4-, and 6-item scales. In a robustness check, we adopt Davis' (Reference Davis2023, 19) recommendation of using two items about the United States as Christian nation and advocacy of Christian values. In our sample, this measure is highly correlated with the six-item index (0.88), and it produces nearly identical results (C.5 in the online Appendix).
• White grievance is an additive 0–9 scale derived from three four-category statements that tap attitudes on racial discrimination, border defense against illegal immigration, and immigrants' contribution to American society and culture. Higher values indicate higher White grievance.
• Election distrust is a four-category item that reads, “It is hard to trust the results of elections nowadays” (Bartels, Reference Bartels2020). Response categories have been reversed so that higher values indicate distrust.
We use logistic regression that produces odds ratios. All models include dichotomized controls for gender (woman = 1), age (younger than 55 = 1), education (B.A. or less = 1), and rural (living in rural countryside or rural town = 1). We do not control for race/ethnicity because our sample is nearly all White (just 6% identifies as non-White); results hold when controlling for race.
Republican pastors and the use of force
The United States is no stranger to systemic violence, and North Carolina, a former Jim Crow state, has a long history of racist and partisan violence (Christensen, Reference Christensen2008; Eamon, Reference Eamon2014). Scholars of religion have for decades documented how Christians have sorted into religious conservative and liberal camps (Wuthnow, Reference Wuthnow1988), which fueled what Pat Buchanan hailed as “a religious war … for the soul of America,” vividly captured in Hunter's Before the Shooting Begins (Reference Hunter1994). This history is suggestive of a widely held view that the battle lines are chiefly drawn between conservatives and liberals. Here we seek to refine this view by looking inside the conservative faction. We find that Republican pastors are divided on the use of force, and we examine why.
We begin by describing the sample distribution on force and assessing our expectations regarding biblical literalism, Christian nationalism, White grievance, and election distrust. We then draw on interviews,Footnote 8 conjoint analysis, and mediation analysis to investigate direct and indirect effects.
The notion that America's traditional way of life is under threat is pervasive among conservative Christians. Sixty-four percent of Republican North Carolina pastors in our sample agree that moral values are “very much” threatened, and 87% report that “a majority” or “very many” of their congregations feel the same way.
Thirty percent of those in our sample report that force may have to be used to save America's traditional way of life (Table 3). It is worth pointing out that a pastor who agrees that “we may have to use force” is not saying that he is likely to actually engage in violence. “May” does not mean “will.” Yet under specific circumstances “may” can slip into “will.” Follow-up interviews revealed that pastors considered their response to this question with care and, in some cases, trepidation. The Civil War left a deep imprint on the politics of NC and the consciousness of pastors (Christensen, Reference Christensen2008), yet one pastor (003F) explained that he still agreed that we may have to use force: “I remember answering that question with that in mind, thinking, O God, have mercy. Because it would destroy the fabric of our culture.”
Table 3. Republican pastors and the use of force
Note: Data from the authors' survey. N = 141 pastors.
We now assess our expectations on biblical literalism, Christian nationalism, White grievance, and election distrust. Model 1 shows that biblical literalism, an indicator for religious dogmatism, is positively associated with the legitimation of force. As we expect, the estimate for religious dogmatism loses significance when we specify a model that includes Christian nationalism (model 5), consistent with H1b. Christian nationalism is strongly predictive of legitimizing force (model 2, H2), as is White grievance, the heightened fear about White marginalization in a multicultural society (model 3, H3). Model 4 corroborates that distrust in elections legitimizes force.
The results of our full model, model 4, are robust to the inclusion of support for Trump in the 2020 election, which is positively associated with, but not a statistically significant predictor of, support for the use of force. Nor does support for Trump mediate Christian nationalism in a statistically significant way (see online Appendix C.3). One may speculate about the reasons for this nonfinding. Perhaps the strongest reason is that Republican tolerance for force to protect America's traditional way of life predates the 2020 general election and the January 6 events. Bartels' survey of 2019 among Republican voters, which informed our study, provides incontrovertible evidence. Kalmoe and Mason's (Reference Kalmoe and Mason2022) book documents a long-standing rising trend among a minority of Republicans (and to a lesser extent, Democrats) toward endorsement of partisan violence. Additionally, Trump has largely consolidated his support within the Republican Party, so intra-party divides occur along other dimensions.
In a multivariate model, Christian nationalism, White grievance, and electoral distrust explain 29% of the variance. The substantive effects are large: a one standard deviation shift in Christian nationalism, from a sample average of 11 on the Whitehead and Perry 0–18 scale to a score of 15, increases the odds by 89.6%; translated into probabilities of legitimizing force, this represents an increase from 15.5 to 29%. The equivalent substantive effect of a one-standard shift on White grievance, from a sample average of 5 on the 0–9 scale to a score of 7, increases the odds by 78%, which shifts the propensity to legitimize force from 14.2 to 23.6%. A one-standard shift on election distrust, from a sample average of 2.9 to 3.9, increases the odds by 103%, or an increase in propensity to legitimize force from 19.2 to 33.4%.
The effect of Christian nationalism and White grievance appears conditional on how pastors judge the electoral process, as Figure 1 shows (H4b). The left panel models the effect of a pastor's Christian nationalism on the probability of legitimizing force conditional on whether they strongly distrust (dark gray) or strongly trust (light gray) elections. The right panel does the same for White grievance. The pattern is similar: as support for Christian nationalism becomes stronger or White grievance intensifies, election distrust drives a wedge between the pastors.
Figure 1. Christian nationalism, White grievance and distrust in elections. Note: The left panel plots how the effect of Christian nationalism on the probability of legitimizing force varies among those who strongly distrust elections (dark gray slope) and those who trust elections (light gray slope). The right panel does the same for White grievance. The X-axis registers the minimum value, the 25th, 50th, and 75th percentiles, and the maximum value on each scale. The Y-axis plots probabilities with 95% confidence intervals. Full models are available in online Appendix C.6.
Discussion
A statistical model in which covariates are tightly associated (see online Appendix Table B.2) is a leaky adjudicator of causality. This section complements these models with mediation and conjoint analysis, and with illustrations from face-to-face interviews of the deep consideration pastors gave to their responses in the survey. This strengthens the finding that (a) Christian nationalism—not biblical literalism—is a driver for endorsing violence, (b) White grievance is consistent with systemic but not interpersonal racism, and (c) electoral distrust is the major mediating force transforming Christian nationalism and, especially, White grievance, into condoning the use of force. Holding partisan identity constant, these factors shed light on why Republican pastors may share theological and social values but differ substantially in their support for democratic norms.
Christian nationalism in a “Broken world”
How might a Christian pastor legitimate violence? First, it is important to distinguish between religion in its original sense as piety or devotion, and religion in its modern meaning as a socio-cultural system of designated behaviors and practices (Droogers, Reference Droogers and Clarke2011). The first is a relation between a person and God; the second is a relation between a person and the wider society. In the context of contemporary Christianity, this is the distinction between biblical literalism and Christian nationalism.
Forty-nine percent of Republican pastors in our sample report that they believe Scripture is the actual word of God and is to be taken literally, word for word. An evangelical biblical literalist will seek to proselytize, but this is not the same as seeking to make Christianity the law of the land. One biblical literalist (001F) explained this important distinction:
We're living in a broken world, I know that. But I also know that one day it is not going to be like that. And that's what gets me through every day I sit here. … I can't change that destiny. I can't. I can't change this world but just one person at a time.
Christian nationalism, on the contrary, is political, but even among Republican pastors in our sample, it is a minority project. Forty-two percent agree somewhat or strongly that the federal government should declare the United States a Christian nation. This topic elicits strong sentiment. Conflict can take place within, as across, the poles of a polarized society. One low-key pastor (003F) erupted, “Christian nationalism makes me sick,” yet like the majority of conservative pastors, this pastor supports school prayer and religious symbols in public spaces. Another pastor, who strongly rejects force (004NF), links the danger of Christian nationalism to the January 6 events:
I would have never thought that what you put forth [the use of force as legitimate] would actually ever take place in this nation until the insurrection in January where a group of people were so convinced that their presidential candidate lost because of a rigged election. And they were so convinced of that they actually took up arms against our nation. I mean, I have never known in the history of the church when the church took up arms that it ever gained anything. […] I do understand the frustration. But to resort to violence so that your particular viewpoint wins the day is against everything Jesus stands for.
Viewed in isolation, we find that biblical literalism affects willingness to use force under controls (Table 4, model 1), but it loses significance when Christian nationalism is present (Table 4, model 5). Mediation analysis reveals that the effect of biblical literalism on legitimizing force is absorbed by Christian nationalism (Table 5).
Table 4. Logistic models explaining the use of force among Republican pastors
Note: This table reports odds ratios (probability of success/probability of failure) with standard errors in parentheses. A coefficient of 1.22 on Christian nationalism (model 5) means that the odds of someone legitimizing force (versus not legitimizing force) increase by a factor of 1.22 for each one-unit increase. This means that the chances of someone with a score of 16 on the Whitehead/Perry scale of Christian nationalism endorsing force are 22.4 × 10 = 224% higher than for someone with a score of 6. Translated in probabilities, this produces a probability of 33.4% against 6.2%, holding other variables at their means. N = 141. ***p < 0.01, **p < 0.05, *p < 0.1. The results are robust when controlling for partisanship (online Appendix C.2) and vote for Trump (online Appendix C.3), and when using the four-category version of force (online Appendix C.3). Diagnostics tests show that these results are robust to outliers with high residuals and high leverage (analyses available from authors).
Table 5. Christian nationalism mediating Biblicism on force
Note: This table shows standardized regression coefficients from a mediation analysis. The coefficients are obtained with the Zhao et al. (Reference Zhao, Lynch and Chen2010) Monte Carlo simulation (500 iterations). Control variables are set at their means.
Biblical literalism can add stridency to Christian nationalism, but it does not produce the desire for a Christian nationalist project. The stridency stems from casting political opponents as agents of evil and infusing graphic imagery into political discourse. Recent sociological research documents how the language of conquest, war, and apocalyptic thinking has become intertwined with much contemporary evangelicalism (Whitehead and Perry, Reference Whitehead and Perry2020). However, a literalist takes this imagery at face value, and does not reflexively conclude that he should fix the political world. One of the pastors (005NF) reflected on how he assesses his congregation's understanding of the links between Revelation and contemporary American politics:
I can count on perhaps two hands [in a church community of over 2000 people] the people who hold to [the] QAnon conspiracy theory. I am talking about something different. I am talking about—and this actually ties into what our theology, specifically our eschatology, the way we view the end times, says—because we believe that God tells us in Scripture that [the times] will continue to get worse and worse but that we are to hold on and persevere. So there is this very unique tension—I haven't made this connection till this conversation with you right now—between being actually very irate over the direction our country is going and yet at the same time to hold this Book that we believe is inspired by God and recognize that He told us this would happen. So are we angry at the direction of our country? Absolutely, one hundred percent. But it is not a shock.
Framing contemporary political conflict in the language of biblical cosmic struggle between good and evil can be a powerful mobilizing tool for those wanting to “take back America”—Christian nationalists. Republican Christian nationalists who perceived persecution of Christianity were the most polarized segment of the public ahead of the 2020 presidential election (Djupe et al., Reference Djupe, Lewis and Sokhey2023). In the days before the January 6 insurrection, Trump supporters organized a “Jericho March” against alleged election fraud (Armaly et al., Reference Armaly, Buckley and Enders2022). Just as God instructed Joshua to march around Jericho seven times until priestly trumpeting brought the walls down, Christians gathered in D.C., blowing the ram's horn blown at Rosh Hashanah, to banish the “darkness of election fraud” and ensure that “the walls of corruption crumble” (Green, Reference Green2021).
Force looms larger as an option if the likelihood of achieving intensely held goals dwindles. Conservative Christianity is shrinking because it is losing traction with the younger generation. A 2024 report from the Public Religion Research Institute found that Generation Z adults (ages 18–25) are less likely than previous generations to belong to an institutionalized religion like Christianity.Footnote 9 One pastor (001F) who considered the use of force legitimate described to us how he has been unable to combat declining church attendance:
Let's go back to say, 1980. The churches were full of worshipping people… . The churches, right now, are dwindling. I've got two churches. One of my churches will hold 500 people; Sunday, there were thirteen there. Thirteen. I've got another church that will hold 150–200, and there were 43 there. And there's nothing that I have ever been able to do.
White grievance
White grievance is more political than interpersonal. Eighty-eight percent of pastors in the sample agree that “immigrants contribute a great deal to American society.” Support among pastors who are open to force is slightly lower, at 77%.
White grievance is driven by policies designed to counter the effects of systemic racism, but there is substantial disagreement: 74% of the force pastors compared to 47% of the entire sample strongly disagree with the statement “Racial discrimination is the main reason why many Black/African-Americans can't get ahead these days.” Those strongly disagreeing tend to draw a distinction between past racism, which they acknowledge, and systemic racism as an ongoing evil in American society, which they deny. Individuals have complex understandings of race relations, and public opinion data show variation in the types of racial resentment that plague the United States today (Cramer, Reference Cramer2019). For example, one pastor (004F) conceded the racist history of North Carolina, but considered that era done after the 1960s:
You know, we are living in an area where slavery was a real problem and there was a lot of prejudice against African-Americans, but that's long since been dealt with, and people have a high respect for Martin Luther King and for the Civil Rights Movement.
Recognition of the persistent effect of racism is one thing, condoning segregation on skin color quite another. One pastor (004NF) who strongly disagrees that racial discrimination is the main reason for why Black Americans can't get ahead paid a personal price for denouncing explicit racism:
When I was an undergraduate … I took about three or four students home from college, and one of those who was a friend of mine was Black. I hadn't been home fifteen minutes, and my father called me: “What are you doing bringing home that Black young man?” And I said, “Well, Daddy, he wanted a ride and I took him home.” And he said, “Well, you're not going to drive him in that car anymore. You need to call him and say you won't take him back.” And I said, “Daddy, no, I am not.” So I took him back to school. I had money enough in the bank to buy that car, and so I went home early [the following week], got the money out of the bank and went into [my father's] office and said: “Master, here is your money. I just want to say whoever wants to ride in my car, I am paying for it.” And he said, “Well, I am done paying for your education.”
Racial segregation does not show up in the data we have on pastors' personal friendships with people of color. Respondents with high White grievance are no less likely to report that they have immigrant or ethnically different friends. Twenty-four percent of pastors who score above the median on White grievance have at least three friends with different ethnicity/race or immigrant status compared to 20% for those who score below the median on White grievance. Ordinary least squares regression reveals no statistically significant relationship between social heterophily and White grievance (online Appendix C.7).
We cannot exclude that social desirability might contribute to that nonfinding, but social desirability is unlikely to explain why force pastors are no more likely than nonforce pastors to eschew interracial collaboration in church affairs. In a conjoint experiment, we asked respondents to choose between two churches as their twinning partner over the next 5 years. The twinning program required a schedule of regular, funded visits, so pastors should anticipate intensive inter-church exchange. A conjoint design allows one to assess how a person makes decisions over bundled outcomes (Hainmueller et al., Reference Hainmueller, Hopkins and Yamamoto2014). The pastors were presented with four pairs of church profiles that varied by religious denomination (Baptist, Presbyterian, Methodist, Catholic); state (Georgia, New York); location (rural, urban); partisan composition (Republican, Purple church, Democratic); and ethnic composition (predominantly White, ethnically diverse, predominantly Black). If the use of force is motivated by a preference for racial segregation, we would expect force pastors to avoid twinning with Black or ethnically mixed churches. Figure 2 reveals that ethnic composition is not a systematic discriminator for force pastors when choosing a twinning partner (online Appendix C.7). Indeed, force pastors lean toward prioritizing ethnically diverse churches, though the effect does not reach conventional significance (p = 0.13).
Figure 2. Choosing twinning partner church. Note: The coefficients are estimated average marginal component effects. Error bars reflect 95% confidence intervals; constraints for gender, rural, age, education; clustered by respondent.
The only church feature that discriminates significantly is partisan composition. Force pastors systematically under-select churches that are mainly composed of Democratic voters or that are Purple, i.e., politically pluralistic. Nonforce pastors also tend to prefer Republican churches over Democratic or Purple churches, though the partisan effect is less strong. This is consistent with the argument that affective polarization in contemporary America is driven by partisan sorting (Mason, Reference Mason2018).
While there is little evidence that personal racism drives White grievance, there is suggestive evidence of systemic racism. Fear and hostility against Black Lives Matter is pervasive among pastors in the sample (66%) and yet higher, at 88%, among those who think force may be necessary. The difference is statistically significant in a difference-of-means test, and robust in model specifications. Several pastors we spoke to single out Black Lives Matter as a reason for their willingness to consider force, as a pastor (004F) from a church in rural North Carolina explains:
When the riots were happening in Raleigh [in June 2020],Footnote 10 this frightened me. There were these rumors of people taking this fight from the cities to the suburbs and from the suburbs to the rural communities. There were a number of people who looked me in the eye and said: “Pastor, I will have my gun ready.” The frightening thing for me, the concerning thing for me, I knew they were not kidding… . if push came to shove and rioters did come to our county, there would be pushback on a very real and, I am afraid, violent, level.
These findings are consistent with recent research that draws a link between negative Black Lives Matter attitudes, systemic racism, and violence (Drakulich and Denver, Reference Drakulich and Denver2022).
Electoral distrust
Force appears to be a response to democratic disillusion. The statement we put to pastors following the 2020 presidential election, “It is hard to trust the results of elections nowadays,” is similar to the one Bartels posed to Republican voters prior to that election. Perhaps surprisingly, the proportion reporting that elections are hard to trust is similar across the two surveys: 33% in our sample of pastors strongly agree (among Republican voters it is 34%); 39% of pastors somewhat agree (among voters 40%). Among pastors who lean toward legitimizing force, those percentages are, respectively, 58 and 35%. Electoral distrust fuels anger that may spill over into violence. One interviewee (002F) expressed it as follows:
Going back to the presidential election, yes, there is a very large number of people who feel like “it simply wasn't honest.” They may not know how it was not honest, they may not be willing to act on their beliefs. But they are angry … And we will practice a little disobedience, if need be.
Election distrust mediates the effect of White grievance on force (43%), and the direct effect of White grievance is weak (and nonsignificant). The mediation effect of electoral distrust for Christian nationalism is much weaker (8% and not significant), while the direct effect of Christian nationalism on force is large at 30%. There is a qualitative difference, then, between the two chief motivations for using force: White grievance spills over into a willingness to condone force to the extent that it feeds on electoral distrust; Christian nationalism is an independent motivating force that can powerfully induce pastors to legitimize the use of force—irrespective of their views on electoral integrity (Table 6).
Table 6. Electoral distrust mediating Christian nationalism and White grievance on force
Note: This table shows regression standardized coefficients from a mediation analysis. The coefficients are obtained with the Zhao et al. (Reference Zhao, Lynch and Chen2010) Monte Carlo simulation (500 iterations). Control variables are set at their means.
Conclusion
Research on political polarization and research on religion and politics tend to exist on different islands. A search for the concept Christianity rarely yields a mention in work on political polarization. Bartels' (Reference Bartels2020) fine study of the use of force among Republicans does not mention Christianity or religion, and Kalmoe and Mason's (Reference Kalmoe and Mason2022, 26, 161, 163) Radical American Partisanship mentions Christian nationalism just three times. The literature on the politics of religion is beginning to examine the ways that political elites and belief structures contribute to polarization, with a focus on interparty polarization. Recent directions include experimental comparisons of distinct yet polarizing effects of civil religion and Christian nationalism (Vegter et al., Reference Vegter, Lewis and Bolin2023) and Djupe et al., The Full Armor of God (2023), which examines the mobilization of Christian nationalism in America using longitudinal and panel data. This paper situates itself in this new vein of research. It sheds a stark light on a key recruitment ground for mobilizing agents of Christian nationalism, and in doing so we nuance the idea that the United States is rift in homogenous polar camps. Conservative Protestant pastors, the religious scaffold of hard right Republicanism in North Carolina, are divided. This is evident in the tensions within the major evangelical and mainline organizations, as well as in the responses to our survey and interviews.
White grievance and Christian nationalism produce deep divisions among Republican pastors. The most powerful factor leading a pastor to condone the use of force is Christian nationalism—the ideology that traditionalist Christianity should be infused in American public policy. The second most powerful factor is White grievance—anger, resentment, and perceived victimization in response to the challenge to White privilege and power. Both factors, and particularly the latter, are intensified by distrust of the electoral process.
Our survey uses a purposive sample rather than a representative one, and we encourage future researchers to pose these questions to new samples of Christian leaders to probe the character and magnitude of political division among Republican-leaning pastors. We also strongly encourage future researchers to test the hypotheses among pastors of color (Perry et al., Reference Perry, Cobb, Whitehead and Grubbs2022).
Over the past 5 years, every major Protestant organization—the Southern Baptists, United Methodists, Presbyterians—has been riven by conflict and secession. Polarization within the wider society has not produced unity within the contending camps. When one conservative Methodist (005NF) complained to us that “They've taken a scorched earth policy,” he was referring to his fellow Methodist pastors, not the Democratic Party.
Supplementary material
The supplementary material for this article can be found at https://doi.org/10.1017/S1755048324000191.
Acknowledgments
The authors wish to thank the religious association leaders in North Carolina for their support in disseminating our survey. We also thank the editors and two anonymous reviewers for their constructive feedback. Earlier versions were presented at the European University Institute, Florence, and the European Political Science Association in Glasgow.
Financial support
This research was funded by the Advanced European Research Council grant TRANSNATIONAL (#885026) administered by the Robert Schuman Centre, EUI, Florence, and the Kenan and Burton Craige Research Foundations at the University of North Carolina at Chapel Hill.
Competing interests
None.
Stephanie N. Shady is a lecturer of political science at Washington University in St. Louis. Her work on religion, national identity, and politics has been published in West European Politics and the Journal of European Public Policy.
Liesbet Hooghe is a W.R. Kenan distinguished professor of political science at the University of North Carolina at Chapel Hill and a research professor at the European University Institute, Florence, Italy. She has published on political polarization, party competition, and multilevel governance in western societies.
Gary Marks is a Burton Craige distinguished professor of political science at the University of North Carolina at Chapel Hill and a research professor at the European University Institute, Florence, Italy. He has published on American exceptionalism, multilevel governance, and comparative politics.
Open access
