Introduction
Much like American politicians, larger political entities (e.g., parties, administrations, branches of government, institutions) often carry reputations about their general posture or orientation toward religion. For example, the idea of “the God gap” between Republicans and Democrats is a reflection of genuine differences in religious composition between parties, but it also reflects differences in perceptions about which party is more “pro-religion” or “pro-Christianity” (Claassen, Reference Claassen2015; Castle et al., Reference Castle, Layman, Campbell and Green2017). Polling data since 2003 have shown Americans generally perceive the Republican Party as more “friendly” toward religion compared to the Democratic Party. Yet entities within parties can carry their own reputations. In 2009, Americans rated the Obama Administration much higher in terms of friendliness toward religion compared to the party he represented (Pew Research Center, 2009). As scholarship on political campaigns and rhetoric recognize, these religious reputations are often strategically cultivated, since it generally benefits politicians, parties, and administrations to be viewed as pro-religion (Coe and Chapp, Reference Coe and Chapp2017; Bramlett and Burge, Reference Bramlett and Burge2021; Campbell et al., Reference Campbell, Layman and Green2021; Perry, Reference Perry2023).
Other entities, however, are less helped by such a reputations, particularly if they are expected to be religiously impartial. The Supreme Court, for example, is ideally supposed to be non-partisan and unbiased, rendering decisions on whether laws or policies are “Constitutional,” not on whether they advance a preferred religious doctrine or constituency. Consistent with this idea, 83% of Americans believe Supreme Court justices should not bring their own religious views into how they decide major cases. Yet the composition of the Supreme Court changed considerably during the Trump Presidency, and subsequently rendered decisions in favor of conservative Christians, involving freedom to refuse serving gay couples, public prayer at high school football games, and most famously, overturning Roe v. Wade (Totenberg, Reference Totenberg2022). Unsurprisingly, in 2022, 44% of Americans believed the justices have relied “too much” on their own religious beliefs in recent decisions and the percentage of Americans who believe the Supreme Court is “friendly” toward religion nearly doubled since 2019 from 18% to 35%, while views of their religious neutrality declined 18% (Pew Research Center, 2022).
This research note examines how “identity congruence,” or the correspondence between Americans' partisan and ideological identities with the perceived identity of politicians or political groups, shapes how Americans perceive the religious orientations of specific political entities. Drawing on group identity theories of religion and politics, Perry and Davis (Reference Perry and Davis2024) find that identity congruence powerfully influences the perceived religiosity of individual politicians. Specifically, they find the leading predictor of how Americans rated the religiosity of prominent politicians like Donald Trump, Joe Biden, or Kamala Harris was Americans' own partisan or ideological identity rather than their own religious characteristics or even knowledge of a given politician's religious identity.Footnote 1 The authors also find religious perceptions were shaped by Christian nationalism (Broeren and Djupe, Reference Broeren and Djupe2024; Djupe et al., Reference Djupe, Lewis and Sokhey2023; Vegter et al., Reference Vegter, Lewis and Bolin2023), which can not only be an aspect of identity congruence on its own (Perry, Reference Perry2023), but other studies show also works in combination with partisan and ideological identities, suggesting the influence of each on Americans' perceptions would be contingent (Djupe et al., Reference Djupe, Lewis and Sokhey2023; Perry and Davis, Reference Perry and Davis2024).
In the current study, I expect patterns observed by Perry and Davis (Reference Perry and Davis2024) to extend to entities like political parties and institutions. Pew Research Center (2009, 2022) has long observed that Americans' perceptions of these entities' religious orientations differ strongly across partisan and religious identities in expected directions. Yet the reputations of individual politicians as well as parties and institutions are most often benefited (or at least not penalized) by being viewed as more religious or pro-religion by in-group members (Castle et al., Reference Castle, Layman, Campbell and Green2017; Coe and Chapp, Reference Coe and Chapp2017; Campbell et al., Reference Campbell, Layman and Green2021). In contrast, I expect that partisan and ideological interests would lead Americans to perceive the religious orientation of the Supreme Court differently. Though in-group bias might incline Democrats and Republicans, liberals and conservatives to view their corresponding parties as “friendly toward religion” with positive valence, those same in-group considerations might incline Democrats and liberals, for example, to view the Supreme Court as “friendly toward religion” with negative valence (given its legitimacy depends on fairness rather than partisan or religious fidelity), while inclining Republicans and conservatives to view the Supreme Court as “neutral” toward religion.
The potential influence of Christian nationalism, however, is an additional puzzle that warrants more examination, specifically its role as a factor in identity congruence. On the one hand, research suggests Christian nationalism sacralizes in-groups vis-à-vis out-groups (e.g., Perry et al., Reference Perry, Schleifer, Whitehead and Frantz2024a, Reference Perry, Shortle, McDaniel and Grubbs2024b), and thus, it might be expected that greater Christian nationalism would amplify positive in-group evaluations, inclining Americans—Democrat or Republican, liberal or conservative—to view their own political in-groups as friendly to religion, given that pro-religion friendliness is still seen as a positive characteristic across parties (Campbell et al., Reference Campbell, Layman and Green2021; Perry and Davis, Reference Perry and Davis2024). Yet, Djupe et al. (Reference Djupe, Lewis and Sokhey2023) argue Christian nationalism is essentially “a Republican project” resulting in cross-pressures for Democrats that might temper their own in-group preferences. In other words, perhaps Christian nationalism among Democrats will mitigate their tendency to evaluate the religious orientations of the Republican Party or current Supreme Court with negative valence, like we might expect for Democrats in general.
Integrating and testing these ideas, I draw on data from a large nationally representative survey of American adults to predict Americans' perceptions of friendliness, neutrality, or unfriendliness toward religion focusing on the Democratic Party, the Republican Party, and the Supreme Court. My analyses largely support expectations that both Christian nationalism and congruence on political identities, separately and in combination, predict how Americans perceive the posture of all three political entities toward religion. Yet I also find the entity itself matters. My findings thus extend our understanding of how Americans' group orientations shape their evaluations of key political entities in complex ways.
Methods
Data
Data for this study are from Wave 114 of the Pew Research Center American Trends Panel survey (hereafter ATP 114), which are publicly available at www.pewresearch.org. The ATP 114 was fielded from September 13–18, 2022 by Ipsos, a global marketing and polling firm. ATP waves are nationally representative, random-sample surveys of U.S. adults who participate via self-administered online surveys in English or Spanish. For the ATP 114, 11,687 panelists were sampled and a total of 10,588 responded for a response rate of 91%. The ATP surveys are all weighted in a multiple step process that accounts for multiple stages of sampling and nonresponse. The weights were used for the multivariate analyses in order to bring the sample into conformity with demographic and non-demographic benchmarks. Only half (N = 5,300) of ATP114 respondents were randomly selected to answer certain questions about religion and politics used in this study. After eliminating cases with missing data, the final N for full models is between 4,679 and 4,685.
Dependent variables
Outcomes for this study are three questions in which Pew asked respondents: “For each of the of the following groups, please indicate whether it is generally: FRIENDLY toward religion, NEUTRAL toward religion, or UNFRIENDLY toward religion.” (1) The Democratic Party, (2) The Republican Party, and (3) The Supreme Court.”Footnote 2 Because neutrality toward religion rather than friendliness, for at least one entity (The Supreme Court), might be viewed as the ideal evaluation, I treat these responses as categorical rather than ordinal. Thus, I use multinomial logistic regression for all three groups with “neutral toward religion” as the reference category.
Key independent variables
My key predictor variables are indicators of partisan and ideological identity as well as an indicator of Christian nationalism. Party identity is measured with four nominal categories: Republican (reference), Democrat, Independent, and Other Party. Conservative ideological identity is entered as an ordinal variable from 1 = Very liberal to 5 = Very conservative.Footnote 3
Christian nationalism has been measured in a variety of ways, ranging from single-item indicators (Perry and Whitehead, Reference Perry and Whitehead2019; Perry et al., Reference Perry, Schleifer, Whitehead and Frantz2024a) to multi-item scales (Whitehead and Perry, Reference Whitehead and Perry2020; Gorski and Perry, Reference Gorski and Perry2022; McDaniel et al., Reference McDaniel, Nooruddin and Shortle2022; Djupe et al., Reference Djupe, Lewis and Sokhey2023). I use a straightforward binary measure from the ATP114 that simply asks Americans: “Do you think the United States should be a Christian nation?” Respondents could answer yes or no. Though this measure does not allow for shades of agreement, if “Christian nationalism” represents the belief that the United States should be a Christian nation, this indicator is clear and direct.Footnote 4 Thus, rather than attempt to combine this indicator with other measures to capture a broader construct, I simply focus on those who affirm or reject the notion that the US should be a Christian nation.
Control variables
Though I anticipate the primary confounds will be among key predictor variables, I also control for a variety of demographic and religious characteristics theorized to influence Americans' religious evaluations. These include age (in 4 categories), gender (male = 0, female = 1),Footnote 5 racial identity, (White = reference, Black, Hispanic, and Other Race), education (1 = less than high school to 6 = post-graduate degree), household income (1 = less than $30,000 to 9 = $100,000 or more), and Southern residence.
Religious characteristics included measures for religious identity and practice. For religious identity, I organize respondents into White Evangelical Protestant (reference), White Non-Evangelical Protestant, Black Protestant, Catholic, Other Christian, Non-Christian Religion, Atheist, Agnostic, and Nothing in Particular. For religious practice, the ATP114 includes a question about religious service attendance, which is measured from 1 = never to 6 = more than once a week. Online Appendix Table A1 presents descriptive statistics for all variables in multivariate models.
Plan of analysis
Online Appendix Tables A2–A4 present multinomial logistic regression models predicting Americans' perceptions of friendliness or unfriendliness versus neutrality (reference) toward religion among the Democratic Party (Table A2), the Republican Party (Table A3), and the Supreme Court (Table A4). In each table, I first present a baseline model with full controls (Model 1), followed by interaction terms between Christian nationalism and party identity (Model 2) and ideological identity (Model 3). I use robust standard errors due to the weighting required for Pew data. For simplicity, I present my results below by plotting and interpreting average marginal effects.
Results
Figure 1 presents forest plots for all three political entities with my key predictor variables. Effects with error bars (95% confidence intervals) that do not cross the vertical line (the line of null effect) are statistically significant. Focusing on the Democratic Party in the first panel, Americans who affirm the US should be a Christian nation are more likely to view the Democratic Party as unfriendly toward religion and less likely to view it as neutral, consistent with findings from Perry and Davis (Reference Perry and Davis2024) regarding Democratic politicians and with Djupe et al.'s (Reference Djupe, Lewis and Sokhey2023) argument that Christian nationalism represents a “Republican project.” The effects of party identity and ideological identity are as expected. Compared to Republicans, Democrats are more likely to view the Democratic party as neutral or friendly toward religion and less likely to view it as unfriendly. Regarding ideological identity, the more Americans identify with conservatism as opposed to liberalism, they are more likely to view the Democratic Party as unfriendly toward religion and less likely to view it as neutral or friendly.
Average marginal effects of predictors on Americans' perceptions that the Democratic Party, Republican Party, and Supreme Court are friendly, neutral, or unfriendly toward religion.
Source: American Trends Panel Survey, Wave 114.
Note: Controls held at their means. Error bars are 95% confidence intervals.
The effect of Christian nationalism is somewhat different for Americans' perceptions about the Republican Party in the second panel. Among Americans who affirm the US should be a Christian nation, they are less likely to view the Republican Party as friendly toward religion, but no different in viewing it as neutral or unfriendly. The effect of conservative ideological identity gives a clue why. As Americans identify more with conservatism, they are more likely to view the Republican Party as neutral toward religion, less likely to view it as friendly, and no more likely to view it as unfriendly. This suggests that among Americans who identify with the Republican Party ideologically, they see the Republican Party as simply fair, rather than partial toward religion. The effects of party identity are more straightforwardly group-oriented. Compared to Republicans, Democrats are more likely to view the Republican Party as unfriendly toward religion and less likely to view it as neutral.
Lastly, focusing on the Supreme Court in the last panel, those Americans who affirm that the US should be a Christian nation are more likely to view the Supreme Court as neutral toward religion and less likely to view it as friendly. This pattern is nearly identical for conservative ideological identity. Taken together, these findings suggest religious impartiality and fairness is the more complimentary view for the conservative-dominated Supreme Court. Indeed, the partisan pattern is instructive here. Compared to Republicans, Democrats and Independents are less likely to view the Supreme Court as neutral toward religion and more likely to view it as friendly, suggesting the negative evaluation of the current Court's orientation to religion would be one of pro-religion bias, not unfriendliness.
Given that Christian nationalism's associations with various social attitudes is often highly contingent on key political and social identities (Gorski and Perry, Reference Gorski and Perry2022; Djupe et al., Reference Djupe, Lewis and Sokhey2023; Perry and Davis, Reference Perry and Davis2024), how do these factors combine to shape Americans' religious evaluations of political entities? Figure 2 presents predicted marginal probabilities of believing the Democratic Party, Republican Party, and Supreme Court are “friendly” toward religion by party identity and Christian nationalism.
Predicted marginal probabilities of perceiving the Democratic Party, Republican Party, and Supreme Court as friendly toward religion by Christian nationalism and partisan identity.
Source: Pew American Trends Panel, Wave 114.
Note: Controls held at their means. Error bars are 95% confidence intervals.
Regarding the Democratic Party, Democrats who affirm Christian nationalism are slightly more likely than those who do not to perceive the Democratic Party as friendly toward religion. This pattern does not hold for other parties, suggesting Christian nationalism in combination with being a political in-group member combines to increase perceptions of the Democratic party's friendliness toward religion. This theory is supported in that I observe the identical pattern among Republicans regarding the Republican Party in the middle panel.
Yet I observe the opposite pattern for Democrats and Independents regarding the Republican Party. Though Republicans who affirm the US should be a Christian nation are significantly more likely than those who do not to perceive the Republican Party as friendly toward religion, for Democrats and Independents, those who affirm Christian nationalism are less likely than those who do not do not to perceive the Republican Party as friendly toward religion. Again, these patterns suggest Christian nationalism combines with partisan in-group identity to shape Americans' perceptions of the orientations of political parties toward religion.
Regarding the Supreme Court, however, I observe a different pattern. Among Republicans, whether they affirm Christian nationalism or not, they are relatively unlikely to view the Supreme Court as friendly toward religion. However, among Democrats and Independents, those who reject Christian nationalism are more likely to view the Supreme Court as friendly toward religion, while those who affirm Christian nationalism are less likely to view the Supreme Court as friendly. Thus, in this case, it seems that while Christian nationalism trends to incline party members to view the partisan in-group as friendly toward religion (and the partisan out-group as unfriendly), evaluations of the Supreme Court's “friendliness” toward religion follow a different logic.
Plotting the interactions between ideological identity and Christian nationalist belief in Figure 3 reveal additional interesting patterns. Regarding the Democratic Party, Americans who affirm or reject Christian nationalism do not differ from one another in terms of viewing the Democratic Party as friendly to religion across ideological identity. However, regarding the Republican Party, the two groups of Americans most likely to view Republicans as friendly to religion are (1) those who identify as “very liberal” or “liberal” and reject Christian nationalism and (2) those who identify as “conservative” or “very conservative” and affirm Christian nationalism. These patterns likely reflect two valences given to “friendliness” toward religion. For the anti-Christian nationalist liberals, pro-religion friendliness is likely a negatively-valenced characteristic they eagerly assign Republicans. Alternatively, Christian nationalist conservatives likely view friendliness to religion as something positive they associate with Republicans.
Predicted marginal probabilities of perceiving the Democratic Party, Republican Party, and Supreme Court as friendly toward religion by Christian nationalism and ideological identity.
Source: Pew American Trends Panel, Wave 114.
Note: Controls held at their means. Error bars are 95% confidence intervals.
Regarding the Supreme Court, there is again a different pattern from the two parties. Among Christian nationalists, ideological identity makes virtually no difference. Whether Christian nationalists are “very liberal” or “very conservative,” they are unlikely to say the Supreme Court is friendly toward religion. However, among those who reject Christian nationalism, those most likely to say the Supreme Court is friendly to religion are those who are “very liberal” or “liberal,” suggesting that pro-religion friendliness is meant negatively among anti-Christian nationalist liberals.
Another way to show how evaluations of the Supreme Court differ from the two parties is to show how party identity, ideological identity, and Christian nationalism correspond to shape evaluations of the Supreme Court being “neutral” toward religion in Figure 4 (based on the same models in Appendix Table A4). Because the Supreme Court's general legitimacy is far more tied to being non-partisan and unbiased toward religion (Pew Research Center, 2022), I would expect Christian nationalists, largely in support of the current Supreme Court and its decisions, would be more likely to emphasize its neutrality toward religion rather than its friendliness. In contrast, I would expect anti-Christian nationalists who fear the Supreme Court as packed with religious conservatives would be less likely to perceive the Supreme Court as neutral.
Predicted marginal probability of perceiving the Supreme Court as neutral toward religion by Christian nationalism, partisan identity, and ideological identity.
Source: Pew American Trends Panel, Wave 114.
Note: Controls held at their means. Error bars are 95% confidence intervals.
Indeed, this is exactly what I find in Figure 4. Democrats and Independents who reject Christian nationalism (thus combining party identity with anti-Christian nationalism) are the least likely to see the Supreme Court as neutral toward religion. However, among Americans across party identities who affirm Christian nationalism, their likelihood of seeing the Supreme Court as neutral toward religion stays relatively high. Turning to ideological identity, those Americans who are “very liberal” or “liberal” and reject Christian nationalism are the least likely to view the Supreme Court as neutral toward religion (because they are the most likely to view the Supreme Court as friendly toward religion in Figure 3), while Americans who affirm Christian nationalism are consistently likely to view the Supreme Court as neutral toward religion regardless of their ideological identity.Footnote 6
Discussion and conclusions
The current study examined how Americans perceive the orientations of various political entities toward religion. Confirming previous research looking at individual politicians (Perry and Davis, Reference Perry and Davis2024), I find congruence on partisan and ideological identity predicts that Democrats and those who identify with liberalism view Democrats as friendly toward religion. However, among Republicans, Christian nationalism and being more conservative inclined Americans to view Republicans as neutral toward religion. This suggested, what would later be demonstrated with reference to the Supreme Court, that in-group preferences may lead to religious neutrality being the more positively valenced evaluation rather than being “friendly” to religion.
Though this latter finding was unexpected for Republicans, I anticipated that pattern regarding the Supreme Court and my analyses affirmed my expectations. Christian nationalism and conservative ideological identity inclined Americans to view the Supreme Court as neutral and away from viewing it as friendly toward religion. In contrast, Democrats and Independents, particularly if they rejected Christian nationalism, were more likely than Republicans to view the Supreme Court as friendly toward religion and less likely to view it as neutral. This affirmed that perceptions of “friendly” and “neutral” can be positively or negatively valenced depending on the entity in question and whether its legitimacy as an institution depends on its neutrality.
Also as anticipated, I found perceptions of entities' orientations toward religion were contingent on the entity and combinations of Christian nationalism and political identity. And the patterns affirmed both the idea of Christian nationalism as fundamentally about sacralizing in-groups vis-à-vis out-groups (Perry et al., Reference Perry, Schleifer, Whitehead and Frantz2024a, Reference Perry, Shortle, McDaniel and Grubbs2024b), as well as a source of cross pressure for Democrats (Djupe et al., Reference Djupe, Lewis and Sokhey2023). On the former point, Democrats who affirmed Christian nationalism were more likely to view the Democratic Party as friendly to religion, but not Republicans. In that circumstance, being “friendly” toward religion is positively valenced toward the in-group. Yet regarding the Supreme Court, Democrats and liberals who affirm Christian nationalism were less likely to view the Supreme Court as friendly toward religion, despite the fact that among Democrats and liberals, the Supreme Court's pro-religion friendliness was likely being understood in a negative, biased way (see Figure 1). Thus, perhaps Christian nationalism was tempering the extent to which Democrats and liberals evaluated the current conservative Supreme Court negatively. In contrast, Democrats and liberals who rejected Christian nationalism were more likely to say the Supreme Court was friendly to religion. Future studies should further tease out how Christian nationalism interacts with political identities to shape Americans' religious evaluations of political entities and individuals.
Recognizing some data limitations, the data are obviously cross-sectional and experimental designs (e.g., Castle et al., Reference Castle, Layman, Campbell and Green2017; Campbell et al., Reference Campbell, Layman and Green2021) might be useful in discerning whether religious rhetoric in combination with information about the specific entity (partisan or non-partisan institution) might work in combination with identity-related factors to shape Americans' perceptions. It would also be important to test these patterns at different time points, for example, under different Presidential administrations or if the overall composition of the Supreme Court changes.
Limitations notwithstanding, these findings extend our understanding of how Americans' group-based and ideological characteristics potentially shape their religious evaluations in the political arena. While recent research (Perry and Davis, Reference Perry and Davis2024) led me to expect positively valenced religious evaluations on the basis of identity congruence, examining political entities whose institutional legitimacy depends on it being religiously unbiased (Pew Research Center, 2022) reveals perceptions of pro-religion friendliness can be negative, but equally shaped by in-group/out-group dynamics.
Supplementary material
The supplementary material for this article can be found at https://doi.org/10.1017/S1755048324000245.
Data
Data for this study are freely available at www.pewresearch.org.
Financial support
None.
Competing interests
None.
Samuel L. Perry is the Sam K. Viersen Presidential Professor of Sociology at the University of Oklahoma. His research examines the ways demographic transitions, political realignments, and technological innovations are shaping American religion, race, and politics. Along with over 120 peer-reviewed journal articles, he is the author or co-author of five books, including Taking America Back for God (with Andrew Whitehead) and the Flag and the Cross (with Philip Gorski). His most recent book is called Religion for Realists: Why We All Need the Scientific Study of Religion (2024, Oxford).
Open access
