Persuasive Appeal of Civil Religion

In chapter 5, I present evidence that presidential candidates who use civil religion rhetoric tend to activate religious evaluative criteria as a basis for vote choice. Although the data in chapter 5 have the advantage of analyzing actual voters immersed in actual campaigns, the data lack the precision of an experimental analysis. The two studies used in this chapter compensate for this weakness, exposing individuals to religious cues under controlled conditions and obtaining more accurate measures of civil religion identity. The results from these studies are consistent with (and extend) the results from chapter 5, indicating that civil religion rhetoric directs attention to religious evaluative criteria. For high-CRI respondents, civil religion rhetoric effectively induces positive emotions and provides a point of common religious identification between candidates and voters.

Data from the undergraduate civil religion experiment provide a strong test of how candidate rhetoric interacts with individuals’ religious predispositions. Figure 6.5 displays the likelihood of respondents

Vote choice by rhetorical cues and civil religion identity

FIGURE 6.5 Vote choice by rhetorical cues and civil religion identity

Percentage of Christians likely to vote for a hypothetical candidate (scoring 5 or higher on a 7-point scale) by rhetorical condition (no cues, civil religion rhetoric, or subgroup religion rhetoric) and level of civil religion identity. Consistent with chapter 5, candidates were evaluated less favorably by high-CRI voters when no religious cues were present (p < .05). For a multivariate test of this relationship, see the online appendix, CRI, civil religion identity

voting for the hypothetical candidate in each of the three rhetorical conditions. As the graph illustrates, high-CRI respondents were significantly more likely than low-CRI respondents to endorse a candidate whose message contained some sort of religious cue. Interestingly, high-CRI voters also appeared to punish candidates who failed to use civil religion rhetoric, whereas low-CRI voters viewed these candidates more favorably. This parallels the negative slope shown in figure 5.2 (panel A) in the previous chapter, where candidates who failed to cue American civil religion were punished by more religiously committed voters.27

Evidence from the National Civil Religion Identity Study also indicates that civil religion identity has a unique persuasive power. And, consistent with the previous discussion, its persuasive force is more closely connected to a symbolic assessment of leadership than to a substantive issue stance. Table 6.1 shows the results from two multivariate regressions, illustrating how question-order priming interacts with religious identity to predict the salience of religious voting criteria. I consider here two broad measures of political evaluation (as in figure 6.2). The first, religious image, combines the importance that individuals attach to a candidate’s moral character, religion, and family values and is designed to approximate symbolic expectations about representative performance. The second, religious issues, combines the importance that individuals place on a candidate’s stance on abortion and same-sex marriage, and is intended to approximate a substantive representation dynamic.

Just as figure 6.2 shows that high-CRI voters are more concerned with religious images than with issues, table 6.1 indicates that religious cues play an important role in activating these image-laden evaluative criteria. High-CRI individuals who were exposed to subtle civil religion cues were significantly more likely to report that image-laden concerns are important in evaluating candidates. And, consistent with the findings from chapter 5, not all religious orientations reacted in the same way. In fact, all else being equal, self-described religious fundamentalists were actually less likely to say that these factors are important in the civil religion condition. Moreover, the regression predicting religious issue salience does not obtain similar results. Civil religion priming did little to encourage voters to desire a substantive issue agenda—the effects of civil religion rhetoric were entirely symbolic.

TABLE 6.1 Religious cues and the religious basis of candidate evaluation

Religious image

Religious issues


.430** (.043)

.391** (.073)



Party identification

.01 (.006)

.009 (.010)


.06** (.014)

.054** (.024)

Civil religion identity

.068 (.295)

-.063 (.112)



.103** (.019)

.074** (.032)

Question order

Civil religion condition

.022 (.026)

.038 (.044)

Subgroup religion condition

.024 (.025)

.058 (.042)


CRI * Civil religion

.17* (.098)

.112 (.169)

CRI * Subgroup religion

.054 (.092)

-.089 (.159)

Fundamentalism * Civil religion

-.05** (.025)

-.036 (.043)

Fundamentalism * Subgroup religion R-square

-.031 (.027) 0.462

.072 (.047) 0.194

Notes: All entries are unstandardized regression coefficients. All components of interaction terms are grand mean centered. Dependent variables are the Religious image and Religious issues scales. Standard errors are in parentheses.

* significant at p <.1, ** significant at p <.05.

< Prev   CONTENTS   Source   Next >