The Psychological Underpinnings of Civil Religion
Representatives are not just agents who enact specific policies but are symbolic carriers of the American creed, expected to uphold certain standards of moral righteousness. This representative dynamic begins, in many ways, with the rhetorical context established in political campaigns. This is not to suggest that voters have quasi-religious expectations of leaders who invoke the language of a civil religion in every political culture. In American politics, however, the expectations about representatives are steeped in a normative context that places religious imperatives on candidates’ self-presentation.32
The argument, then, is that in American politics voters—and particularly civil religion identifiers—have a set of quasi-religious expectations about political leadership. These expectations exist because of norms that are deeply ingrained in American political life. Although explicitly crossing the wall of separation between church and state is prohibited, civil religion provides a vehicle through which religious evaluative standards become acceptable or even desirable. When candidates invoke the civil religion, they magnify the effects of this evaluative standard on their own candidacy.
Definitively testing a thesis that involves the transmission of cultural norms is inherently challenging. Data from the National Civil Religion Identity Survey, however, provide one interesting way to approach the problem. The survey contains a short battery of questions adapted from Mark Snyder’s (1974) self-monitoring scale (see also Gangestad and Snyder 2002). Self-monitoring refers to the tendency of individuals to differ in the extent to which they monitor and regulate their self-presentation in public appearances. High self-monitors are highly responsive to social cues and norms about appropriateness. The concept of self-monitoring can help shed light on the extent to which civil religion rhetoric works by cueing expectations about religious political norms. If civil religion works by creating the expectation that religious evaluative standards are politically and culturally appropriate, then high self-monitors who identify as members of the American civil religion should be especially likely to weigh religious considerations more heavily when exposed to civil religion cues. In other words, because high self-monitors tend to place a greater weight on public appearances, they should be particularly sensitive to the language of civil religion.
To test this, I looked at two evaluative standards that bear several conceptual similarities. I examined the use of family values and the same- sex marriage issue, which is often framed by political conservatives as a family values issue, as evaluative criteria. In this way, I could examine whether civil religion rhetoric primed the expectation that a religious image was a desirable quality in a candidate and whether this extended to religious issues as well. To begin, figure 6.6, panel A, displays the relationship between CRI score and the importance of the same-sex marriage issue in the control condition. The three lines correspond to different levels of self-monitoring. As expected, the graph shows very little difference between high and low self-monitors. This essentially replicates what we see in figure 6.2—high-CRI voters were not any more concerned with cultural issues than low-CRI voters were. Interestingly, the negative slope for high self-monitors indicates that, if anything, civil religion
FIGURE 6.6 Self-monitoring and the normative social pressure applied by civil religion rhetoric
Scatterplots of CRI scores (x axis) and the importance of either same-sex marriage (panels A and C) or family values (panels B and D) scores (y-axis), with regression lines added for levels of selfmonitoring. Separate plots are displayed for control (no religious cue; panels A and B) and civil religion (panels C and D) conditions. The interaction between CRI scores and self-monitoring is significant only in panel D, suggesting that, when high-CRI individuals encounter civil religion rhetoric, they take this as a cue that it is normatively appropriate to evaluate the candidate on quasi-religious grounds. The dotted lines are linear regression lines for low self-monitors (diamonds). Dashed lines are linear regression lines for average self-monitors (squares). Solid lines are linear regression lines for high self-monitors (triangles). For multivariate tests of these relationships, see the online appendix, http://facstaff.uww.edu/chappc/. CRI, civil religion identity
identity voters worried about public appearances were actually less likely to report that same-sex marriage is an important issue. In the absence of religious cues, the expectation of high-CRI voters is that the same-sex marriage issue should not matter in candidate evaluation.
Figure 6.6, panel C, displays the same set of relationships in the civil religion condition. For high self-monitors, the slope of the self-monitoring line is now positive. This suggests that even a cue as subtle as the question- order experiment can create the expectation that candidates ought to be evaluated based on religious issues. Nevertheless, the difference between high and low self-monitors in panel C is far from staggering (and not significant), indicating that, although civil religion cues might create a modest imperative to evaluate candidates on cultural issues, this social expectation is quite slight.
The story was different when individuals were asked whether a candidate’s family values were an important part of their evaluative process. When no religious cue was present (figure 6.6, panel B), all levels of selfmonitors follow a similar (and by now familiar) pattern—higher levels of civil religion identity translate into a greater perceived importance of a candidate’s family values. Moreover, when first exposed to a civil religion cue (panel D), a significant gap emerges between high and low selfmonitors. That is, high civil religion identifiers who were most concerned with maintaining a public presentation consistent with cultural norms nearly universally endorsed the notion that family values was an important component of candidate evaluation. In contrast, high self-monitors who did not endorse the tenets of civil religion (low-CRI voters) were more likely to reject family values as an important voting rationale. Priming American civil religion identity tends to push those most sensitive to social pressures in different directions.
This indicates that part of the reason for the success of civil religion rhetoric is because it engenders the expectation among many voters that representation based in a religious candidate image is culturally and socially desirable. Thus, when Bush regularly invoked the language of civil religion in the 2004 campaign, he reinforced a social norm that religion— or at least quasi-religious values—have a place in politics. High-CRI voters making a decision in the 2004 electoral context would have had the distinct impression that religious voting was entirely appropriate. Note, however, that this evaluative standard did not apply to religious issue positions but only to vague character appraisals such as the candidate’s family values.
Although the U.S. Constitution erects a wall of separation between church and state, voters regularly bring religious standards to bear on political figures. The dynamics of this complex representational process depends on the religious standards in question, the voters’ own religious identity, and the nature of candidate rhetoric. Although many have interpreted religious voting patterns as placing a substantive issue-based mandate on the officeholder, most religious voters are actually more concerned with a symbolic representational style. Indeed, the most common style of religious rhetoric—American civil religion—tends to direct many voters’ attention to the religious image, not any particular religious policy platform. It does so by promoting the view that, in American politics, religious evaluative criteria are social acceptable and normatively desirable.
On one hand, these findings are an effective counterexample to the view that a rhetorical cultural war is driving cultural cleavages in the United States. Although self-identified religious fundamentalists may be overwhelmingly concerned with abortion and same-sex marriage, fundamentalist does not describe most of the American electorate. Thus, consistent with Geoffrey Layman and John Green (2005), we find that an issue-driven culture war is real but is being “waged by limited religious troops on narrow policy fronts” (83). On the other hand, the analysis also reveals deep tensions with respect to the way religious political rhetoric is normally expressed. Candidates’ characterizations of America as a blessed nation resonate with many, but they leave many others unhappy with these candidates and less likely to weigh religious evaluative criteria. Thus, although civil religion is properly considered a broad superordinate identity, it is far from an identity with universal appeal.