A Portrait of Civil Religion Identity
The results of the National Civil Religion Identity Study indicate that many Americans deeply identify with American civil religion but also that civil religion identification is not uniformly accepted across the electorate. Even though civil religion is theoretically a nondenominational religion, in practice its adherents are primarily Christians, suggesting limits to the ability of civil religion to serve as a transcendent cultural glue.
Many American agree with the basic tenets of American civil religion and see themselves as members of this spiritual community (figure 6.1). Between one-third and one-half of the individuals surveyed saw the basic American institutional framework as infused with religious elements. About 35 percent of Americans somewhat agreed or strongly agreed that the United States has entered into a special covenant with God, and 32 percent saw the U.S. Constitution as a holy document. Likewise, given that presidents regularly (and often famously) make strong religious appeals, we might expect a sizable percentage of the public to see this institution as having a religious element.5 The data enforce this prediction— about 57 percent agreed that “The office of the presidency is a sacred position.”
FIGURE 6.1 Civil religion identity in the American public
Weighted agreement with the six civil religion identity (CRI) indicators. Bars represent the percentage of respondents who “somewhat” or “strongly agree ” with the statement on a 5-point scale.
A sizable percentage of respondents also agreed with the survey items that explicitly tap into the group-membership aspects of civil religion identity by making references to citizenship, using the we voice, and making references to shared responsibility. Fully 76 percent of the public agreed or strongly agreed that “As Americans, we are blessed with special opportunities.” Likewise, almost half of Americans saw citizenship as a sacred responsibility, and 39 percent thought that America, collectively, holds a “special higher power.”
Together these results indicate that a large portion of the American public identifies themselves as members of an American civil religion and agrees with its basic tenets. Critically, these survey items all appear to be measuring the same underlying religious dimension; that is, individuals who agreed with one question tended to agree with all the questions, and the opposite is also true.6 This is important because it suggests that the group-membership aspects of civil religion are closely intertwined with beliefs about the privileged place of American institutions. If you agree that America has entered into a special compact with God, you also probably have a spiritualized sense of attachment to the United States and understand your own citizenship in quasi-religious terms.7
To further explore the composition of this identity, I created a scale from these six survey question, and compared this CRI scale with other important political and demographic characteristics. Even though civil religion identity is not entirely uniformly represented across the political spectrum, the data indicate that no single nonreligious demographic group has a monopoly on civil religion adherents. For example, among different racial groups, whites were only slightly more likely to strongly identify with civil religion identity, Hispanics slightly less, and African Americans are roughly evenly split between high- and low-CRI individu- als.8 Overall, the differences in the racial composition of civil religion identifiers are far from staggering. Likewise, a slightly higher percentage of men identified with American civil religion, although the difference between men and women is not statistically significant. Older Americans generally scored higher on the CRI scale, although the difference between young and old is relatively modest.9 Finally, although Republicans did score significantly higher on the CRI scale than did self-identified Independents and Democrats, civil religion identity is still present in all partisan groups. About 20 percent of Democrats and 35 percent of Inde?pendents still averaged an “agree” response on all six items on the CRI scale.
The most striking and normatively consequential differences in civil religion identity adherence are those between different religious traditions. Civil religion rhetoric theoretically aspires to transcend faith traditions and engender a sense of inclusivity in the political community. However, rhetorically civil religion might not always perform as promised. For example, as we see in chapter 3, references to God tend to privilege an image of God as parent-provider, which runs counter to the understanding of some traditions. This raises the question of whether America is accorded a unique quasi-religious status among those who find little to agree with in the substance of these beliefs.
The results of the National Civil Religion Identity survey indicate that Americans do not equally identify with civil religion across all religious traditions.10 Catholics, Protestants, and Mormons strongly identify with American civil religion, whereas Muslims, Jews, agnostics, and atheists find little to agree or identify with on the CRI scale.11
This finding presents something of a paradox. Membership in the American civil religion is characterized as being open to everyone—being a common denominator of all faith traditions. Civil religion theoretically exists as a solution to American religious diversity, providing common moral and ethical denominators that can serve as a basis for political unity. And, throughout history, the rhetorical construction of the identity has been trending toward greater inclusiveness. Prothero (2007) concurs, noting that the rhetorical expressions of American nonsectarianism has evolved from just including Christian denominations to referencing the Judeo-Christian tradition and to, much more recently, referencing the “Abrahamic” tradition.
Although trending toward greater inclusivity, however, civil religion has always been an exclusive group, and this exclusivity may have observable consequences in the American public.1 2 Regardless of whether civil religion rhetoric is trending toward universal inclusivity, Americans who are non-Christian certainly do not feel included—a finding that may have its origins in religious political rhetoric. This presents a challenge to theorists who claim that civil religion identity can effectively embrace the full religious diversity of the United States. Although the aim of civil religion may be to unite, in practice this identity reveals deep divisions along religious lines.13
Two important conclusions can be drawn from the previous section. First, in the American electorate, civil religion identity is unevenly distributed across religious traditions. It is striking that members of many faith traditions uniformly reject an identity that is commonly invoked in presidential political rhetoric. Insofar as symbolic representation involves the active “alignment of wills” between rulers and the ruled, it seems that large portions of the American electorate are not being adequately represented along these lines.
Second, despite these divisions, a large portion of the American electorate—particularly Christians—strongly identify with American civil religion. This has important representational consequences when we consider high-CRI individuals as a voting bloc, potentially united to make political demands. On one hand, if high-CRI voters have a substantive issue agenda in mind, then leaders who rhetorically embrace the language of American civil religion have a mandate to enact specific policies. This issue-based hypothesis often dominates electoral interpretations when we ask which issues the “religious vote” cares most about. On the other hand, it is not entirely clear what the substantive issue agenda for high-CRI voters would actually look like. After all, the language of civil religion has been used to rally support behind a number of political causes across American history. For those who identify with the civil religion tradition, the true basis of political representation involves electing a leader who appears to be a prototypical group member—a candidate with whom citizens feel a common sense of identity and shared fate. According to this hypothesis, high-CRI voters may desire a specific symbolic representational style, not any particular issue agenda.
If this latter hypothesis is correct, this has important consequences for how we understand the role of religion American elections. Specifically, if high-CRI voters are more concerned with a candidate’s religious image than with particular issues, leaders in government should use caution in aggressively promoting a culture war-style policy platform or in claiming any sort of religious policy mandate. Consider the 2004 presidential campaign, in which Bush frequently and deftly employed the language of American civil religion. When Bush achieved a lopsided victory over Kerry among (church-attending) religious voters, many suggested that this electoral division signaled a political agenda the president should pursue. Bush advisor Karl Rove, being interviewed by Fox News’s Jim Angle shortly after the election, made this case explicitly. After Rove trumpeted the significant victory among religious voters, Angle followed up by asking what policy consequences this electoral cleavage would have:
ANGLE: Will the president press the issue now and push for a constitutional amendment defining marriage?
ROVE: Well, look, the president has this old fashioned notion that when you run on something in a campaign, you attempt to do it in office. He ran in 1999 and 2000 on a certain series of issues. And once in office, pursued each and every one of those issues. I think the American people can have confidence that he treats the things that he said in the campaign as significant promises and pledges, which he will now attempt to fulfill.
ANGLE: So the answer is yes?
Rove here makes the assumption that religious voters prioritize issues such as same-sex marriage, a view consistent with theories of a culture war in American politics. For Rove, this policy mandate is justified on the grounds that Bush performed exceedingly well among frequent church attenders. Nevertheless, insofar as Bush’s rhetoric was actually activating civil religion as the relevant evaluative dimension, there is good reason to challenge Rove’s assumption. It stands to reason that what civil religion voters saw in Bush was a leader who, like they did, viewed America as having a special place in a divine plan and a sacred set of responsibilities. Despite Bush’s personal opposition to same-sex marriage, it is not clear that this issue was salient for high-CRI voters.14
Data from the National Civil Religion Identity Survey supports the hypothesis that high-CRI voters prefer a leader who symbolizes moral clarity and a sense of American spiritual values but who does not necessarily have a substantive issue platform.15 To address this, let us begin by examining the rival issues hypothesis to determine whether civil religion voters are particularly concerned with traditional moral issues, such as abortion and same-sex marriage (figure 6.2). As the graph illustrates, voters who score high on the CRI index were only modestly more likely to see abortion and same-sex marriage as important issues when evaluating candidates. This indicates that the extent to which individuals embrace membership in American civil religion does not greatly influence the salience they attach to issues typically thought to be the fodder of the culture war.
FIGURE 6.2 Issue and image salience by level of civil religion identity
Above: Figure 6.3 reports the Importance of a candidate’s stance on abortion, same-sex marriage, moral character, family values, and religion in assessing candidates. There is a relatively small difference between high- and low-CRI respondents on abortion and same-sex marriage issue salience, but there is a large significant difference between high- and low-CRI respondents on each of the image variables (p < .01). For a multivariate test of these relationships, see the online appendix, http://facstaff.uww.edu/chappc/. CRI, civil religion identity
In contrast, high-CRI voters do think about candidates differently when image is taken into consideration (figure 6.2). Civil religion identifiers were far more likely than their low-CRI counterparts to acknowledge that a candidate’s religion and moral character are important and to say that it is important for a candidate have strong family values. It is clear that civil religion identifiers were not evaluating candidates on the issues any more than low-CRI voters were but, rather, desired candidates who appeared to be the rightful inheritors of the civil religion tradition and who led in a manner that communicated a sense of moral purpose. Civil religion identifiers were nearly 20 points more likely to say that a candidate’s religion is important, 30 points more likely to say that moral character is important, and 39 points more likely to say that family values are important. In contrast, high-CRI voters were only modestly more concerned with the dominant religious issues of the day—abortion (13 points) and same-sex marriage (7 points).16 Clearly, voters who endorse and identify with the core principles of civil religion are more likely to desire a certain type of leader, although this does not necessarily translate into a substantive issue platform. This is consistent with the work of Larry Bartels (2008), who finds that frequent churchgoers attach relatively little importance to cultural issues, compared with economic issues such as job creation and economic spending, and that the salience of cultural issues does little to differentiate frequent versus infrequent church attenders. But even though cultural issue salience do little to distinguish voters along religious lines, the evidence presented here indicates that high-CRI individuals are distinguished by their symbolic concern that candidates be of a certain moral pedigree and hold certain values.17
Additional evidence from the National Civil Religion Identity Study suggests that adherents of this identity are most concerned with a particular symbolic representative style. As figure 6.3 illustrates, high-CRI voters were 36 percentage points more likely to desire a candidate that makes them proud of their country and were twice as likely as low-CRI voters to say that it is important to elect a candidate like themselves. This supports the contention that high-CRI voters do not evaluate candidates along the
FIGURE 6.3 “Pride in country” and “a candidate like me” by level of civil religion identity
Importance of “the candidate makes me proud of my country” and “the candidate is someone like me” in assessing candidates. There is a large significant difference between high- and low-CRI voters on each of these variables (p < .01). These differences persist when the analysis controls for national identity and partisan identification. For a multivariate test of these relationships, see the online appendix, http://facstaff.uww.edu/chappc/. CRI, civil religion identity
lines of issue positions; rather, they want candidates who fit a specific group prototype consistent with a spiritualized sense of national identity. Moreover, these relationships hold up even when the analysis controlled for other factors such as national identity.18 Civil religion identity is closely intertwined with concern for both pride in country and common identity.
One potential objection to this line of reasoning questions the utility of the concept of civil religion identity. Perhaps the CRI index is simply measuring secular national identity, or perhaps the CRI index is a redundancy, too similar to other measures of religiosity to be conceptually important. There are both empirical and theoretical responses to these objections. Empirically, data from the National Civil Religion Identity Survey provide a strong refutation to both these concerns. Using multiple regression, I examined the relationship between the CRI index and concerns about a candidate’s moral character while statistically holding constant other factors, such as party identification, religious fundamen- talism,19 and national identity. The data indicate that CRI is the best available predictor of the importance people attach to a candidate’s moral character. All else being equal, neither national identity, religious fundamentalism, religious commitment, nor party identification have a statistically significant relationship with evaluations of a candidate’s moral character. And, although social conservatives were moderately more likely to say that moral character is important, civil religion identity is still the strongest predictor in the model.20
This finding makes theoretical sense. Civil religion identity is distinct from secular measures of national identity, as well as from measures of religious identity such as fundamentalism that specifically imply a specific set of religious beliefs (e.g., biblical inerrancy). Civil religion identity is a highly spiritualized sense of national identity that requires not that its followers adopt specific positions on religious matters but, rather, that they believe that the United States itself is intertwined with the sacred. In fact, the survey data indicate that high-CRI individuals and religious fundamentalists ultimately evaluate candidates on different grounds. For religious fundamentalists and others with a more traditional religious orientation, clearly issues (e.g., abortion and same-sex marriage), not representational style, drive political evaluations. Whereas high-CRI voters attached relatively little importance to these issues compared to low- CRI voters, voters identifying with religious fundamentalism were highly likely to see these issues as politically salient.21
Thus, the importance that religious voters attach to these traditional cultural issues depends on exactly what is meant by religious. The evidence indicates that individuals who strongly identify as religious fundamentalists evaluated candidates based on the candidates’ positions on the traditional cultural issues; this is consistent with accounts of limited but salient religious divisions in American politics (Layman and Green 2005). However, fundamentalists does not describe most voters.22 In contrast, a sizable portion of the electorate—particularly the Christian electorate—strongly identifies as members of American civil religion. These voters are not fundamentally concerned with cultural issues but, rather, representational style. They prioritize a specific vision of moral leadership intertwined with pride in country.
This analysis is consistent with the notion that that an issue-laden culture war is not driving American elections, except perhaps for a small group of religious fundamentalists. This finding has important representational implications, recasting our understanding of religious voting in American elections. It is a mistake to suppose that religious cleavages are always issue-driven, as many analysts and politicians did following the 2004 election. Indeed, a concern with moral values may have had very little to do with substantive issues for many voters. In actuality, many voters—those who strongly identified with American civil religion but not necessarily religious fundamentalism—were probably attracted to Bush (over Kerry) because of the type of symbolic leadership his rhetoric conveyed, not because of any particular issue stance. In other words, Kerry’s problem with frequent church attenders was not that he was pro-choice, although some voters might have rationalized their vote choice in these terms (Lodge, McGraw, and Stroh 1989; Rahn, Krosnick, and Breuning 1994). Rather, Kerry’s problem was that he did not seem like he could authentically steer the nation in a spiritually righteous direction.
Civil Religion Identity, Political Rhetoric, and American Political Culture
Political representation involves not only the evaluative standards that individuals bring to bear on political candidates but also how voters respond to rhetorical cues. Does religious rhetoric shift the basis of candi?date evaluation, and what role do religious cues play in defining American political culture? Given an ostensible separation of church and state, it is important to know whether religious rhetoric is driving a wedge between voters or whether it is working to weave diverse faith traditions together. In other words, how are rhetorical choices influencing American political culture?
I investigated this question using a question-order experiment embedded in the National Civil Religion Identity Study, as well an experiment with undergraduates that exposed participants to various types of religious rhetoric. Both studies employed a similar strategy, designed to mimic the rhetorical cues present in a real campaign, albeit under controlled conditions. In this way, I was able to examine how nuanced rhetorical choices affect the basis of vote choice. I primed participants with various religious cues before asking them a series of questions about how they evaluate political candidates. The results suggest that civil religion rhetoric is a powerful force in American elections but that the effectiveness of the appeal depends a good deal on the religious identity of the individual voter. That is, even though civil religion aims to be nonsectarian, it is clearly not a panacea for greater inclusivity.