“Building Alliances or Rallying the Base:Civil Religious Rhetoric and the Modern Presidency”

Flavio Hickel Jr. Ph.D. – June 2018

Abstract

Civil Religious rhetoric has been utilized throughout American history to legitimize political

interests by drawing upon broadly shared beliefs regarding the nation’s identity, meaning, and

purpose in the world. While scholars have traditionally assumed this rhetoric was employed to

unify the nation others contend that it has the potential to exacerbate conflict as policy debates

morph into battles over the national identity. This research project analyzes presidential

speeches from Franklin Roosevelt through Barack Obama (1939-2012) and finds that the type of

speech delivered, public approval of the president, and the partisan composition of Congress

influence the prevalence of civil religious rhetoric. It concludes that modern presidents have

more often relied upon civil religion to rally the partisan base than build alliances with the

opposition.

Introduction

These economic royalists complain that we seek to overthrow the institutions of America. What they really complain of is that we seek to take away their power. Our allegiance to American institutions requires the overthrow of this kind of power. In vain they seek to hide behind the Flag and the Constitution. In their blindness they forget what the Flag and the Constitution stand for. Now, as always, they stand for democracy, not tyranny; for freedom, not subjection; and against a dictatorship by mob rule and the over-privileged alike. (Franklin Roosevelt, Nomination Acceptance Speech, 1936)

But rarely in any time does an issue lay bare the sacred heart of America itself. Rarely are we met with a challenge, not to our growth or abundance, our welfare or our security, but rather to the values and the purposes and the meaning of our beloved nation. The issue of equal rights for American Negroes is such an issue. (Lyndon Johnson, “American Promise” Major Speech, 1965)

And that’s why our focus is the values, the principles, and the ideas that made America great. Let’s be clear on this point. We’re for limited government, because we understand, as the Founding Fathers did, that it is the best way of

1

ensuring personal liberty and empowering the individual so that every American of every race and region shares fully in the flowering of American prosperity and freedom. (Ronald Reagan, State of the Union Address, 1988)

Regardless of partisan evaluations, few would deny that Presidents Roosevelt, Johnson, and

Reagan all presided over and guided major transformations in American government and society.

As these excerpts illustrate, each leveraged their rhetorical skills to convince the public and

Congress that their preferred policies and broader political agenda were not only good for

America, but also consistent with the nation’s historic identity, meaning, and purpose in the

world. Rhetoric such as this may be characterized as Civil Religion - a set of broadly shared

beliefs, myths, and symbols, derived from the United States’ founding and history, which reflect

ideas about the sacred and transcendental nature of the American experience. (Bellah, 1967;

West, 1980)

Scholars have traditionally argued that civil religion would have a unifying effect upon the

populace as this shared American identity supersedes those characterized by partisanship,

ethnicityi, geography, or religion. (Bellah, 1967) There is no shortage of studies which

document presidents’ use of civil religion for these purposes (Adams, 1987; Pierard and Linder,

1988; Roof, 2009). Contentious political issues such as immigration policy (Beasley, 2004), the

cold war arms race (Ungar, 1991), and military interventions (Haberski, 2012) have in part been

legitimized through the use of civil religious rhetoric. Collectively, this “Traditional View” of

civil religion suggests that presidents should be more likely to employ such rhetoric when facing

public and/or congressional opposition - a scenario in which the unifying and legitimizing effects

of civil religion offers the potential to overcome political resistance.

However, civil religious rhetoric is not employed in an ideological vacuum. A frequent

critique of the Traditional view of civil religious scholarship has been its failure to adequately

2

address the extent to which elites can manipulate these beliefs in the course of political conflict.

(Cristi, 1997; Demerath and Williams, 1985; Williams and Alexander, 1994; Wuthnow, 1988)

The image of civil religion in America as a canopy of shared values, operating exclusively in terms of consensus and social cohesion turns attention away from the role that civil religion plays in defining (or obscuring) national self-understanding; stabilizing (or upsetting) social and national expectations, its sense of destiny and mission; maintaining (or undermining) social values and beliefs; strengthening (or weakening) social consensus; relieving (or exacerbating) social tensions (Cristi, 1997).

Accordingly, the coupling of civil religious themes with partisan goals sends a message that a

particular course of action is not only consistent with the American identity, but often necessary

to maintain that identityii. While civil religious rhetoric may be music to the ears of ideological

compatriots, it also suggests that the opposition is not only wrong, but un-American. As such,

civil religious rhetoric carries a political cost by diminishing a president’s capacity to find

common ground as policy debates morph into pitched battles over the national identity. This

“Critical View” of civil religion suggests that while the deployment of this rhetoric can serve as a

rallying cry for the political base, it may be ill-advised as a means of reaching out to the

opposition.

The purpose of this study is to examine the influence of public approval and partisan

composition of Congress on the prevalence of civil religious language in modern presidential

speeches. While it is beyond the scope of this project to examine the specific motivations of a

president and his speechwriters in each particular case, the results of this analysis allow us to

conclude whether this rhetoric has generally been employed in accordance with the expectations

of the Traditional or Critical view of civil religious scholarship.

Making use of a unique dataset that quantitatively analyzes 180 speeches spanning the

presidencies of Franklin Roosevelt through Barack Obama (1939-2012), I conclude that the

3

prevalence of civil religious language varies according to the type of speech delivered, public

approval of the president, and the partisan composition of Congress. Civil religion was more

prevalent while presidents were articulating a vision for their campaign and future presidency

(Inaugural Address and Nomination Acceptance Speeches) than when explaining policy

decisions made in the course of governing (Major Speeches and State of the Union Addresses).

Furthermore, presidents were more likely to rely on this rhetoric as their partisan majority in both

Houses of Congress increased; lending credence to the Critical view of civil religion. The

influence of public approval of the president on civil religious rhetoric was contingent on the

partisan composition as well. When neither party enjoys a partisan legislative advantage,

popular presidents rely more heavily on civil religion to increase the political pressure on

Congress. However, when the president enjoys an overwhelming partisan advantage in

Congress, public approval had a negative effect upon the prevalence of civil religion. Under

these conditions, presidential success is already reasonably assured and they are unwilling to

incur the costs of employing such rhetoric. Collectively, the evidence suggests that the

deployment of civil religious rhetoric is more likely to be motivated by a desire to rally the

partisan base than build alliances with the opposition.

This study not only advances our understanding of civil religion’s role in presidential

communication but also contributes to the study of the rhetorical presidency more broadly.

Shogan (2007) argues that moral rhetoric is a means by which presidents establish their

credibility and nurture existing authority but that its utility depends on the presence of supporting

circumstances. Knowing how presidents adjust their rhetorical strategies in response to the

political environment provides insight into the circumstances that they consider to be favorable

and the motivations behind this behavior. It also paves the way to a more sophisticated

4

appreciation for the limits of their influence, the conditions under which they may be successful,

and more accurate predictions for how they will behave in the future. (Eshbaugh-Soha, 2010)

I begin with a theoretical overview of civil religion before considering the literature on

presidential rhetoric and the theoretical expectations derived from applying the insights of the

latter with the former. I then describe the research design and present the results before offering

some concluding remarks.

Civil Religion

Robert Bellah initiated a flurry of academic interest in civil religion when he asserted the

existence of a “public religious dimension expressed in a set of beliefs, symbols, and rituals” that

“played a crucial role in the development of American institutions” (Bellah, 1967). He

conceptualized civil religion as an organic source of unity and integration for a diverse and

secularized society; one that provided cultural legitimacy for its political institutions with a

transcendental understanding of American history. The defining feature of civil religion is the

belief in the sanctity and inviolability of the American identity and its constitutive principles.

Artifacts (e.g. American Flag, Declaration of Independence), events (e.g. Nation’s Founding,

Memorial Day), and heroes (e.g. George Washington, Abraham Lincoln) from the past were

sanctified and mythologized to provide a means by which Americans understood their nation’s

identity, meaning, and purpose in the world. Furthermore, the existence of sacred beliefs and

practices necessarily implies that there are also consequences for their violation. It is the fear of

these consequences that has allowed civil religion to serve as an evaluative standard by which

the legitimacy of the political system and its institutions could be judged. (Bellah, 1992) While

this has often found expression in explicit references to God, a Covenant, or American

5

Exceptionalism, they are not necessary to evoke ideas about the inviolability of the American

identity and values.

In the years that followed Bellah’s initial publication, scholars documented the existence of

civil religious beliefs among the public (Christenson & Wimberley, 1978), its effects upon

electoral and policy preferences (Wimberley, 1980; Wimberley & Christenson, 1982; Chapp

2012), and that such beliefs were clearly distinguishable from “church religion” (Coleman, 1970;

Flere & Lavric, 2007; West, 1980; Wimberley et al, 1976). There have also been a number of

insightful qualitative analyses to illustrate the manner in which civil religion has been utilized to

provide legitimacy for immigration policy (Beasley, 2004), the cold war arms race (Ungar, 1991)

and military interventions throughout American history (Haberski, 2012).

Impressive as this body of research may be, it takes for granted that members of society

uniformly interpret civil religious entities in a manner that breeds consensus and cohesion. As a

consequence, this Traditional view of civil religion often fails to adequately address the potential

for division when elites manipulate the meanings of these beliefs in the course of political

conflict. (Cristi, 1997; Demerath & Williams, 1985; Williams & Alexander, 1994) Whillock

(1994) argues that “American civil religion is essentially a study of competing groups who are

seeking to define what America is and the morality that should prevail.” Similarly Wuthnow’s

work (1988) focuses on the alternative civil religious visions of America articulated by liberals

and conservatives and notes that they “have been the subject of disagreement and polarization

more than of consensus and mutual understanding.” This Critical view of civil religion is not

only more consistent with the literature on “multiple traditions” (Smith 1993), and conflict over

the American Creed (Huntington, 1981), but also encourages researchers to analyze the

polarizing potential of civil religious rhetoric.

6

As such, civil religion will be defined here as a set of broadly shared beliefs, myths, and

symbols, derived from the United States’ founding and history, which reflect ideas about the

sacred and transcendental nature of the nation’s identity, meaning, and purpose in the world.

While these civil religious entities are themselves non-ideological, their meanings can be

manipulated and deployed by political actors to legitimize or mask their own ideological agenda.

What is novel about my definition of civil religion is that it is related to but analytically

distinct from ideology, which I define as a set of coherently related shared beliefs and issue

positions that are held by a social class or group, which reflect a set of tangible or intangible

group or self interests, and which influence and legitimize political behaviors and relationships

of poweriii. Although these conceptualizations of ideology and civil religion share some

elements, they differ in that the former reflect a set of coherent issue positions and the latter has

no inherent connection to any particular policyiv. In other words, the association between civil

religious entities and policy/conduct is politically contested. Ideology influences and legitimizes

political behaviors and relationships of power by presenting them as “being in the ‘common

good’ or as generally accepted.” (Williams & Demerath, 1991) However, because ideologies

reflect social class or group interests, they are naturally divisive. In contrast, civil religious

entities are widely shared and culturally significant, and therefore serve as one means by which

these particular ideological interests may be presented as being in the common good.

This conceptualization of civil religion guided the coding of presidential speeches and

hypotheses generated for this project. However, before detailing the procedures for the former, I

will review the literature on presidential rhetoric in the modern context as it pertains to civil

religion so as to generate a set of plausible hypotheses.

7

The Civil Religious Rhetoric of the Modern Presidency

The overriding narrative in the evolution of the American Presidency is the incremental

accumulation of power and influence by which occupants of the executive office are expected to

assert their authority as the sole representative of the entire nation to advance a truly “national”

agenda (Greenstein, 1978). However, because of our Constitutional system of checks and

balances, the ability to exert leadership and/or avoid challenges to presidential authority is not

guaranteed; it often depends upon having the support of Congress. For example, a hostile

legislative branch can deny the president necessary funding for his initiatives, launch

investigations into his conduct, or pursue an alternative political agenda (among other actions

that could threaten a president’s leadership and authority). Historical accounts demonstrate that

presidents believe their legislative success is powerfully influenced by their standing with the

public (Edwards III, 2009) and analysis of Congressional activity corroborates this expectation

(Bond and Wood, 2003; Rivers and Rose, 1985). Furthermore, Gronke and Wilson (2003) have

found that the electoral prospects of members of Congress are influenced by whether they

support the president in accordance with the desires of their constituents. As such, the president

has an interest in maintaining a high level of support among both Congress and the public.

Civil religion is theorized to be one means by which a president can maintain and generate

this support. As the symbolic representative of the American identity (Bellah, 1967; Adams,

1987; Pierard and Linder, 1988) and most influential actor in our government (Cohen, 1995;

Hill, 1998; Kernell, 2007), the president is arguably well positioned to articulate a civil religious

discourse that can impact public opinion and the political process. Zarefsky (2004) notes that

“because of his prominent political position and his access to the means of communication, the

president, by defining a situation, might be able to shape the context in which events or proposals

8

are viewed by the public.” On the other hand, presidential scholars have also argued that the

effectiveness of these powers has been exaggerated (Edwards, 2003) and point to the constraints

imposed by changes in the media environment (Cohen, 2004), partisan/ideological

predispositions of the public and legislators (Lee, 2008), and the general lack of attentiveness

toward political matters by much of the public (Converse, 1964; Delli, Carpini, and Keeter,

1996).

Although the capacity for the president to influence public opinion and the political agenda is

debatable, few would deny that presidents have increasingly relied on official speeches in an

attempt to enhance their authority and leadership (Tulis, 1987) and that civil religion is an

important component in these efforts. (Beasley, 2001; Schonhardt-Baily et al., 2012) A wealth

of literature has demonstrated that the strategic decision to deliver a speech is influenced by

changes in public approval, the economy, and foreign events. (Brace and Hickley, 1993;

Eshbaugh-Soha, 2010; Ragsdale, 1984) However, the factors which may influence the decision

to deliver a speech are not always the same as those that influence its content. Shogan’s (2007)

research demonstrates that the presence of war and economic turmoil had no appreciable effect

on the moral content of State of the Union and Inaugural Addresses. Similarly, although the

presence of divided government does not influence the quantity of presidential speeches

delivered (Eshbaugh-Soha, 2010; Lewis, 1997; McGauvran and Eshbaugh-Soha, 2017), that

context does have an impact on the policies proposed by presidents (Cohen, 2011) and their

willingness to engage in partisan rhetoric (Rhodes, 2014).

What is largely missing in the literature is a comprehensive analysis of the contextual factors

that influence the relative frequency of civil religious rhetoric employed by modern presidents.

The potential influence of public approval and partisan composition of Congress on speech

9

content is of particular interest since the existing scholarship is inconclusive. Although Shogan’s

(2007) regression analysis failed to detect a significant relationship between divided government

and moral rhetoric, her work did not control for public approval of the presidentv or its

interaction with the partisan composition of Congress. Furthermore her case studies

simultaneously suggest that moral rhetoric is a good strategy when Congress threatens to assume

leadership on an issue or when public support is marginal but a bad strategy when the president

is in a position of weak political authority (Shogan, 2007).

The logic of the Traditional view of civil religion suggests that it is precisely at those

moments when the president faces heightened opposition from Congress and the public that the

unifying potential of this rhetoric provides the greatest benefits to leadership. Existing studies

demonstrate that the president’s agenda setting power is diminished under divided government

(Cummins, 2006) as are the prospects of his legislative agenda (Franklin and Fix, 2016). Given

the numerous veto points in which a majority opposition party in either house of Congress may

quash a president’s preferred policies, it is intuitive to assume that the latter will seek to build

bridges with the former to avoid a legislative defeat. While a moderation of policy positions

could accomplish this (Cohen, 2011), a president may alternatively seek to convince or pressure

the opposition by appealing to a shared national identity. The Traditional view of civil religion

theorizes that increasing the salience of an American identity relative to partisan identities

reduces the bias associated with the latter. In other words, this rhetoric calls for the populace to

put aside partisan animosities and ideological preferences and prioritize policies which are

consistent with the American identity and sacred values (as articulated by the president). In the

context of divided government, such actions serve to put pressure on a recalcitrant Congress and

provide them with the ability to “defend their votes as supporting a president [or American

10

interests] rather than on substantive policy grounds alone” (Edwards III, 2009). While it is

beyond the scope of this project to evaluate the effectiveness of such efforts, we should expect to

find that presidents increasingly rely on civil religion as their congressional majorities diminish

and public approval is on the decline.

Hypothesis #1: Consistent with the Traditional view of civil religion, the frequency of civil religious rhetoric will be positively associated with partisan congressional opposition to the president and declining public support for the president.

On the other hand, given the well documented difficulties associated with overcoming

partisan bias (Edwards, 2003; Lee, 2008) a healthy dose of skepticism regarding civil religion’s

capacity to deliver on these promises is warranted. The conceptual weakness of the Traditional

view of civil religion is the failure to appreciate the cost associated with this type of rhetoric. In

contrast, the Critical view of civil religion asserts that such rhetoric may alienate a president’s

opponents and exacerbate conflict as policy debates evolve into battles over the national identity.

Those who share a partisan affiliation with the president should respond with greater enthusiasm

for their shared policy goals since they are now framed as being consistent with the national

identity, meaning, and purpose in the world. This expectation is consistent with the work of

Canes-Wrone (2006) who demonstrates that presidents avoid making public appeals on policy

issues that the public disagrees with. Conversely, his partisan opponents may be less inclined to

assent given that their preferred policy options are now implicitly or explicitly framed as “un-

American.” In contrast to the Traditional view of civil religion, this suggests that employing

such rhetoric while the power of a president’s partisan opponents is ascendant or his public

support is in decline will reduce the prospects for success as political compromises become more

elusive. However, these properties render civil religion an ideal leadership tool under united

government or when congressional majorities are marginal. In these circumstances, legislative

11

success depends more on maintaining the support of your base and is less contingent upon

bargaining with the opposition. Furthermore, rhetorical attacks against the opposition serve to

rally the base against a common enemy while simultaneously delegitimizing the former’s

concerns and potentially their resolve. This expectation is consistent with Rhodes’s (2014)

discovery of a positive association between unified government and rhetorical attacks on the

opposition.

Hypothesis #2: Consistent with the Critical view of civil religion, the frequency of civil religious rhetoric will have a positive relationship with partisan congressional support for the president and increasing public support for the president.

Finally, it must be emphasized that the purpose of civil religious rhetoric (for the Critical or

Traditional view) is to enhance the influence and authority of the president. It stands to reason

that the relative utility of this rhetoric declines as the president’s prospects for success increase.

When the president enjoys extreme levels of public support and large partisan majorities in

Congress his prospects for success are relatively assured; there is therefore less benefit to be

gained from employing civil religious rhetoric. Because the Critical view of civil religion

conceptualizes a cost associated with articulating such rhetoric, we should expect to find that

presidents eschew these costs when they are already confident in their prospects for success.

This is consistent with Canes-Wrone’s (2006) analysis demonstrating that presidents avoid

making public appeals when they can expect legislative success without it.

Hypothesis #3: Consistent with the Critical view of civil religion, the frequency of civil religious rhetoric will have a negative relationship with very high levels of public approval of the president and very strong partisan majorities in Congress.

Civil Religious Rhetoric and Speech Type

This research is primarily concerned with exploring the effects of public opinion of the

president and the partisan composition of Congress on the prevalence of civil religious rhetoric.

12

However, it is also likely that the type of speech delivered will have an impact on the relative use

of this rhetoric. Although no formal comparison of the relative prevalence of civil religious

rhetoric between speech types exist, the literature on presidential communications clearly

demonstrates that important differences in the form and purpose of Inaugural Addresses

(Campbell and Jamieson, 1985), State of the Union Addresses (Schonhardt-Bailey et al, 2012),

Campaign Speeches (Chapp, 2012; Tetlock, 1981), and Major Speeches (Brace and Hinckley,

1993; Coe and Neumann, 2011).

Robert Bellah has asserted that the Inaugural Address is an “important ceremonial event” for

American civil religion, and this claim has been corroborated by content analyses which have

found ample evidence of civil religious language in these speeches (Beasley, 2001; Pierard and

Linder, 1988; Toolin, 1983). Campbell and Jamieson (1985) argue that the Inaugural Address is

of a “distinct rhetorical type” compared to other presidential communication because of their

tendency to:

(1) unify the audiences by reconstituting it as “the people” who witness and ratify the ceremony; (2) rehearse shared values drawn from the past; (3) enunciate the political principles that will guide the new administration; (4) demonstrate that the president appreciates the requirements and limitations of executive power; and (5) achieves these ends through means appropriate to epideictic discoursevi (Campbell and Jamieson, 1985).

Although not mentioned by name, these tendencies share much in common with the

conceptualization of civil religion articulated in this study and reflected in Beasley’s (2001)

documentation of the manner in which such rhetoric has been employed to generate unity and

ideological consensus in these speeches. Furthermore, Sigelman’s (1996) study of these

speeches demonstrates a general increase in the prevalence of unifying symbols and references to

traditional American values throughout U.S. history.

13

While considerably more scholarship has been devoted to understanding the role of civil

religion in Inaugural Addresses, the literature suggests that the prevalence of this rhetoric may be

influenced by the degree of policy specificity generally found in different types of speeches.

Although their work does not explicitly compare the two speech types, Schonhardt-Bailey et al,

(2012) argue that “unlike Inaugurals, State of the Union speeches tend to be more policy

prescriptive and so are less likely to embellish upon the broader, more principled rhetoric of civil

religion.” This suggests that civil religion is stylistically better suited to speeches that seek to

discuss policy in broad strokes rather than fine detail. Campbell and Jamieson (1985) note that

when policy proposals are employed in Inaugural Addresses they serve as “illustrations of the

political philosophy of the president” while those in the State of the Union Address reflect “a call

to immediate actionvii.” Furthermore, Coe and Neumann (2011) argue that because Inaugural

Addresses are a “ceremonial genre” they are theoretically less likely than Major Speeches to

discuss policy in a deliberative manner. According to Brace and Hinckley (1993) the latter is

strategically delivered to explain difficult decisions that have the potential to result in declining

public approval. While civil religious rhetoric may be a component of such efforts, it reasonable

to assume that more emphasis will be placed on substance over symbolism in this context. We

might therefore extrapolate from these insights that because State of the Union Addresses and

Major Speeches are less ceremonial and more likely to dwell on policy details, that they will

feature comparatively less civil religious language than Inaugural Addresses out of stylistic

concerns.

These insights can also be applied to presidential campaign speeches which are more likely

to exhibit simplistic, sweeping generalizations than the more nuanced, complex explanations of

policy offered while in office (Tetlock, 1981; Thoemmes and Conway III, 2007). While this

14

account does not posit a ceremonial function to campaign speeches in the way we ascribe that

status to an Inaugural Address, Chapp’s (2012) analysis demonstrates that appeals to civil

religion were highly prevalent when candidates spoke of religionviii. Furthermore, it may be

argued that because Nomination Acceptances Speeches aim to strike a delicate balance between

articulating a policy platform and presenting a broad vision for their candidacy and tenure in

office - a vision which is often expressed as being consistent with the national identity, meaning,

and purpose in the world - they more closely approximate the form and function of an Inaugural

Address than other campaign speeches. Several authors have gone so far as to assert that the

Nomination Acceptance Speech functions as a civil religious jeremiad where candidates assert

their consistency with the past and admonish their opponents for their misunderstanding of

America’s sacred mission. (Ritter, 1980; Benoit, 1999) As such, we should expect it to occupy

a middle ground in terms of the prevalence of civil religion between the Inaugural Address and

the State of the Union Address/Major Speeches. These expectations are reflected in the

following hypothesis:

Hypothesis #4: Frequency of civil religious rhetoric will vary according to the type of speech delivered. Inaugural Addresses will have a higher prevalence of civil religious rhetoric than Nomination Acceptance Speeches, which in turn will have a higher prevalence of civil religious rhetoric than State of the Union Addresses and Major Speeches.

Research Design

This research project relies on an original dataset comprised of 180 presidential speechesix

delivered between 1939 and 2012x. This includes all “Obligatory” speeches (State of the Union

Address, Inaugural Address, and Nomination Acceptance Speech) deliveredxi during the time

period along with a random sample of “Discretionaryxii” speeches. The latter are confined to

“Major Addresses” which are defined as radio/televised speeches delivered during prime time

15

broadcasting hours and which preempt regular programming on local broadcast networks

(Ragsdale, 2014). After a full list of these Major Speeches were obtainedxiii, I randomly sampled

a number of them from each president equal to the number of years they were in office from each

termxiv. This sampling method provided a control for idiosyncratic tendencies in discretionary

speech making among the presidents and variability in the political context which may have had

an effect on the number of speeches given during their tenure. Out of 302 Major Speeches

delivered during this time period, 72 (23%) were incorporated into my sample for subsequent

content analysis. Table 1 provides an overview of the sample with the number of speeches

delivered by each president along with the percentage of each speech type delivered. Because

Nomination Acceptance Speeches and Inaugural Addresses are only delivered once per term,

they are obviously a small percentage of the total sample compared to Major Speeches and State

of the Union Addresses.

(Table 1 about Here)

Two undergraduates from the Rutgers University Political Science Department were hired to

code the speeches in the summer of 2014. After studying “Civil Religion in America” (Bellah,

1967), they were trained by examining several presidential speeches not included in the sample

used in this studyxv. The coders were instructed to identify complete sentences that included civil

religious rhetoric;xvi which was broadly defined as any reference to the American identity,

meaning, or purpose in the world within a transcendental context. Specifically, they focused on

discussions of what it means to be an American and its hallowed values, along with references to

historical figures and important moments in American history which reflect the sacred nature of

the American experience or God’s blessingsxvii. This coding criterion is consistent with Bellah’s

contention that civil religion has “its own prophets and its own martyrs, its own sacred events

16

and sacred places, its own solemn rituals and symbols.” (Bellah, 1967) The following list of

keywordsxviii was provided to the coders to aid in the identification of civil religious passages:

“Sacred, God, Divine, Providence, Covenant, Consecrate, Destiny, Identity, Purpose, Spirit,

Values, Ideals, Truths, Principles, Forefathers, Ancestors, Constitution, Declaration of

Independence, Historical Figures (e.g. Lincoln, Washington, Roosevelt, Reagan), Historical

Events (e.g. Revolution, Constitutional Convention, Great Depression) xix. However, the coders

were not restricted to identifying passages that contained these keywords; nor was the mere

existence of these keywords sufficient for a sentence to be classified as civil religious.

After refereeing their workxx, I calculated the number of words contained in sentences that

included civil religious referencesxxi. I then calculated a continuous, interval level variable that

reports the percentage of those words compared to the total words in each speech (CRWP). This

variable represents the prevalence of civil religious rhetoric and serves as a proxy measure of the

relative utility that each president believes it will have in that particular context. This served as

my dependent variable for all subsequent analyses performed in this study.

The distribution of CRWP (Figure 1) illustrates that it accounted for an average of 15% of

each speech in the sample and ranged from 0% to nearly 70%. Although a very small

percentage (6%) of speeches contained no civil religious language, the vast majority (70%)

features 15% CRWP or lessxxii. This underscores the fact that civil religion is but one component

in a president’s rhetorical repertoire and that there are occasions when it used sparingly or not at

all. This analysis also reveals four potential outliers that ranged from 57%-69% CRWP. This

included the Inaugural Addresses of President Eisenhower (1953) and Obama (2009) along with

President Reagan’s Major Speech to the nation on Independence Day (1986) and President

George W. Bush’s Major Speech to the nation on the first anniversary of 9/11 (2002). After

17

thoroughly reviewing the data, I decided to retain these outliers in my sample because no errors

were made in the coding and there was nothing contextually abnormal about their occurrence

that would warrant their exclusionxxiii. Moreover, that a speech could contain this high

prevalence of civil religious rhetoric is corroborated by the work of Schonhardt-Bailey et al

(2012) who found that roughly 50% of President Reagan’s speech text in their study was

classified as civil religious through a similar coding scheme.

Figure 1 also demonstrates that CRWP is a bounded variable (0-1) with a non-normalxxiv,

positively skewed distributionxxv. Several data transformations (log, square root, and arcsine)

were attempted to correct for these issues but all failed to improve the normality of the dataxxvi.

Although the nature of the data must be kept in mind when evaluating the results of this study,

the robustness of OLS against non-normality with a sufficiently large sample size provides us

with a reasonable level of confidence in the forthcoming analysis.

(Figure 1 about Here)

To investigate the influence of a president’s partisan relationship with Congress, I created

two variables to represent the size of a president’s partisan majority in both chambers

(HouseMargin and SenateMargin). I first calculated the percentage of members that share a

partisan affiliation with the president (or presidential Nominee in the case of Nomination

Acceptance Speeches). I then subtracted .50 from that number so that it reflects the size of the

partisan majorityxxvii. I also created an interaction term (House*Senate) to estimate the effects of

a scenario where the partisan majority increases in both Houses of Congress. This specification

provides more precision and explanatory power than would be achieved with a categorical

classification of the governing context as united or divided. To explore the influence of public

approval of the president, I relied upon the most recent Gallup Poll presidential job approval

18

questions that preceded a speechxxviii. I created a variable (ApprovalMargin) to represent the size

of a president’s approval margin by subtracting .50 from the Gallup Poll approval percentagexxix.

As such, this variable reflects the amount of public support a president enjoys beyond 50%. I

also created a set of interaction terms to investigate the conditional effects of approval on CRWP

depending on the governing context (Approval*House, Approval*Senate,

Approval*House*Senate). A series of binary variables to represent State of the Union Addresses

(SOTU), Inaugural Addresses (INA), Nomination Acceptance Speeches (NOM) and Major

Speeches (MAJ) were also created to evaluate Hypothesis 4.

Furthermore, this study employed a series of independent variables to control for the effects

of other factors which may have an influence on these hypothesized relationships. The existing

literature on presidential rhetoric demonstrates that the presence of military conflicts and the

performance of the economy have an effect upon the delivery of a Major Speech. (Brace and

Hickley, 1993; Eshbaugh-Soha, 2010; Ragsdale, 1984)xxx Although Shogan (2007) did not

detect a similar effect upon the content of presidential communication, it is with an abundance of

caution in mind that I include these variables in my analysis. A binary variable (WAR) was

created to represent the presence of active military conflicts which were likely to garner

widespread public interest and have a noticeable effect on presidential speechmakingxxxi.

Economic performance (GDP) is represented by the percentage change in the seasonally adjusted

gross domestic product from the previous data point (Annual change from 1929-1946 and

Quarterly change from 1947-2013).xxxii

To ensure the accuracy of my analysis, additional controls were introduced to account for

other factors which could potentially influence the prevalence of civil religious rhetoric. To

control for unmeasured idiosyncrasies of each particular president and partisan influence over

19

rhetoric, a nominal variable (PRES) and a binary variable representing Republican Party

affiliation (REP) were created. Because it is possible that the employment of civil religious

rhetoric could be influenced by election year politics (Chapp, 2012) a binary variable was

created to represent whether a speech was delivered during a presidential/midterm election year

or not. Finally, the hypothesized relationships could be further complicated when we consider

the historical trend toward greater elite polarization and development of the “Partisan

Presidency” (Skinner, 2009) or “Polarized Presidency” xxxiii (Cameron, 2004). Rhodes (2014)

argues that presidents Carter through Obama have responded to increasing partisan polarization

among members of Congress and political activists by adopting a “bipartisan leadership

posture.” In this context, he contends that “recent presidents have used bipartisan themes both to

obscure their own ideological positions and to create a positive contrast with a highly partisan

Congress” (Rhodes, 2014). Although civil religious themes are often coupled with

partisan/ideological policy goals, they are themselves non-ideological and bipartisan. Therefore,

it is possible that these broader changes in polarization will have an impact on the frequency of

civil religious language in presidential speeches. Although I offer no formal expectations on

these effects, a binary variable was created (POLAR) that included presidents who held office

during this time of heightened polarization: Carter, Reagan, George H.W. Bush, Clinton, George

W. Bush, and Obama.

(Table 2 about Here)

Table 2 presents summary statistics of the variables included in this analysis. We can see

that the sample is fairly well balanced in terms of the partisan affiliation of the president, the

presence of war, an election year context, and whether a speech was delivered during the age of

bi-partisan posturing (POLAR). On average the president enjoyed elevated public approval

20

ratings (57%) and annual GDP growth rates (3%). In the Senate, the president’s party averaged a

2% majority over the opposition while in the House of Representatives party control was evenly

divided. As Figure 2 illustrates, we can also see that the employment of civil religion has been a

constant feature of presidential rhetoric with no discernible, overall growth or decline over the

time period. However, the prevalence of civil religious rhetoric does fluctuate a great deal from

year to year. The analysis that follows aims to shed light on the factors which influence this

variability.

(Figure 2 about Here)

Results

The following analysis examines the influence of governing context, public approval, and

speech type on the prevalence of civil religious rhetoric. The results of the linear regression

model (Table 3a) clearly demonstrate that the type of speech delivered has an independent

influence upon CRWP. Consistent with Hypothesis 4, we can see that compared to Inaugural

Addresses (reference group in the model) the other speech types feature significantly less

civil religious language. Table 3b presents the estimated marginal means of CRWP for each

speech type and reports on the significance of those differences. While civil religious

statements account for over 40% of Inaugural Addresses, the average percentage for other

speech types was between 10% and 16%. As hypothesized, Nomination Acceptance

Speeches had significantly higher CRWP than Major Speeches (16% compared to 10%).

Although civil religion was more prevalent in these speeches than in State of the Union

Addresses (12%), the difference was only marginally significant (.078). No significant

differences exist between State of the Union Addresses and Major Speeches. Collectively,

these findings suggest that presidents believe that using civil religious language provides

21

greater benefits and/or is more appropriate while laying out a vision for their campaign and

future presidency (Inaugural Address and Nomination Acceptance Speeches) than when

explaining policy decisions made in the course of governing (Major Speeches and State of

the Union Addresses).

(Tables 3a and 3b about Here)

The regression analysis (Table 3a) also provides support for the Critical view of civil

religion. Because of the inclusion of interaction terms, the results for Approval, House, and

Senate Margins represent the conditional effects of each while the remaining terms equal 0.

When House and Senate Margins equal 0 (e.g. when there are equal percentages of a

president’s partisan supporters and opponents in both chambers) Approval Margin has a

significant, positive relationship to CRWP. Although this particular relationship was not

hypothesized - and seldom occurs - it suggests that a popular president will employ civil

religion to increase political pressure on Congress when their support is marginal. Neither

House nor Senate Margin had a significant effect upon CRWP while the Approval Margin

equals 0 (e.g. when the president enjoys 50% public approval). However, consistent with

Hypothesis 2, the interaction between House and Senate Margin did have a significant

positive effect upon CRWP under these same conditions. In other words, when the president

enjoys 50% public approval, growing partisan majorities in both chambers results in a higher

prevalence of civil religious rhetoric. Substantively, this suggests that a president is relying

on civil religion to help solidify his political base.

The three way interaction between Approval, House, and Senate margin was also

statistically significant but requires more elaborate interpretation. Consistent with

Hypothesis 3, the negative coefficient indicates that as Public Approval, House, and Senate

22

Margins increase, the prevalence of civil religion declines. However, to properly evaluate

this relationship we must take a closer look at the conditional effects of the House and Senate

Margin interaction term on CRWP at various levels of Public Approval (Table 3c). These

estimates were generated by “Process”: a statistical procedure developed by Hayes (2013)

which calculates the conditional effects of each component of the interaction on the

dependent variable at various values of the other terms. Although the size of the Interaction

effect does decline as Public Approval increases, we can see that it is both positive and

significant at all reported levels of Public Approval. Substantively, this suggests that (1) civil

religion is more prevalent under conditions of united government regardless of public

approval (Hypothesis 2) but that (2) as public approval increases, the size of this effect

declines (Hypothesis 3).

(Table 3c about Here)

To better understand these results, Table 3d presents the conditional effects of Public

Approval on CRWP at various levels of both House and Senate Margins. The results

indicate that public approval only exerts a statistically significant influence on CRWP under

two conditions. The first, as reported earlier, occurs when there is no partisan advantage for

either party in both the House and Senate. Under these conditions the prevalence of civil

religion increases with public approval. The second occurs when the president enjoys a 20%

(or greater) partisan majority in both the House and Senate. Under these conditions, public

approval has a significant negative effect upon CRWP (Hypothesis 3). (Supplemental

analysis that visually depicts these conditional effects is available in the Appendix)

It is worth briefly remarking on the lack of significance for the control variables employed in

this model. The party affiliation of the president, election-year politics, economic growth, and

23

the presence of war all failed to register a significant effect on CRWP. Similarly, the age of Bi-

Partisan Posturing has no significant effect upon the frequency of civil religious rhetoricxxxiv.

Collectively, the lack of significance for these variables adds credence to the conclusion that the

prevalence of civil religious rhetoric is primarily influenced by speech type, governing context,

and public approval.

Discussion

The results of this analysis clearly demonstrate that the relative frequency of civil religious

rhetoric incorporated into modern presidential speeches is influenced by the type of speech

delivered. Consistent with Hypothesis 4, civil religious rhetoric was most prominent in

Inaugural Addresses, followed by Nomination Acceptance Speeches, and then Major Speeches

and State of the Union Addresses. The prominence of civil religion in Inaugural Addresses

compared to the other speech types was not surprising considering the extensive existing

literature. However, what is novel are the results indicating the relative prominence of civil

religion in a Nomination Acceptance Speeches compared to a Major Speeches (and marginally

significant differences compared to State of the Union Addresses), and the lack of a significant

difference between State of the Union Addresses and Major Speeches. These findings suggest

that presidents believe civil religion is more useful and/or appropriate when articulating their

general plans for the future than explaining governing decisions already made.

This study also failed to find any evidence in support of the Traditional view of civil religion

that the prevalence of this rhetoric would increase with the growth of partisan congressional

opposition and/or declining public approval (Hypothesis 1). There is therefore little evidence to

suggest that modern presidents have employed civil religion as a means of building alliances

with the opposition. On the other hand, the results of this analysis provide support for the

24

alternative Critical view of civil religion (Hypothesis 2 and 3). When facing a Congress

controlled by the partisan opposition, modern presidents have relied less on civil religion.

Presumably, this occurs because presidents recognize the dangers of antagonizing partisan

adversaries under conditions where their cooperation is vital to effectively govern. Conversely,

the prevalence of civil religion has a positive relationship to the size of a president’s partisan

majority in Congress. This suggests that civil religion is viewed as a means of rallying the

partisan base and weakening the resolve of the opposition.

The results also demonstrate that the influence of public approval of the president on civil

religious rhetoric was contingent on the partisan composition of Congress. When neither party

enjoyed a partisan legislative advantage the prevalence of civil religion increased with public

approval; suggesting that such rhetoric was viewed as a means of applying political pressure on

Congress. However, when the president enjoyed an overwhelming partisan advantage in

Congress, public approval had a negative effect upon the use of civil religion. Under these

conditions, there is no need for a president to incur the costs of civil religious rhetoric when the

achievement of presidential goals is likely to occur without it.

Conclusions

Fifty years ago Robert Bellah’s “Civil Religion in America” (1967) encouraged us to think

about how discussion of our national identity, meaning, and purpose in the world influenced our

politics. While Bellah and others holding the Traditional view theorized about the unifying

potential of civil religious rhetoric, others adopting the Critical view argued for its polarizing

potential. My research seeks to understand whether the rhetorical practices of modern presidents

better conform to one view or the other. The results demonstrate that both the type of speech

delivered and the partisan composition of Congress have an effect upon the prevalence of civil

25

religious rhetoric. Substantively, it is clear that modern presidents have been more likely to rely

on civil religion while preparing to govern (Inaugural Addresses and Nomination Acceptance

Speeches) than in the course of governing (State of the Union Addresses and Major Speeches).

The results also provide strong evidence for the contention that modern presidents are more

likely to employ civil religion when their party controls the legislative branch than when they do

not. This is consistent with the Critical view of civil religion that such rhetoric is used to rally

the partisan base rather than build alliances with the opposition. Furthermore, the finding that

presidents have been less likely to employ civil religion under extreme levels of public support

and partisan majorities reinforces the assertion of the Critical view that there is a political cost to

such rhetoric.

This study infers a general motivation behind the use of civil religion based upon whether the

context in which it is most prevalent is consistent with the expectations of the Traditional or

Critical view of civil religious scholarship. Each provides contrasting expectations for when

civil religious rhetoric would be most prevalent and a theoretical motivation for why a president

would make these discursive decisions. However, it is critical to emphasize that we cannot

assert a direct connection between the context and motivation with the data examined here.

While there is strong theoretical justification for these inferences, the results of this analysis

should be considered to be suggestive of a motivation. Future scholars are encouraged to

conduct comparative case study analyses that could more directly explore the motivations behind

particular speeches.

Similarly, if modern presidents do believe that civil religious rhetoric is more likely to

polarize than unify their audience, it remains unclear whether this belief is valid. The subfield

of political psychology has yet to explore the effects of civil religious framing on public opinion

26

and political polarization. The results of such an analysis would allow us to evaluate whether

presidents have effectively employed civil religious rhetoric and offer invaluable insights to

legislators, interest groups, and social movement actors seeking to influence our politics.

Works Cited

1. Adams, David S. 1987. Ronald Reagan’s “Revival”: Voluntarism as a Theme in Reagan’s

Civil Religion. Sociological Analysis 48(1): 17-29.

2. Beasley, Vanessa B. 2001. The Rhetoric of Ideological Consensus in the United States:

American Principles and American Prose in Presidential Inaugurals. Communication

Monographs, 68(2): 169-183.

3. Beasley, Vanessa B. 2004. We the People: American National Identity in Presidential

Rhetoric. College Station, TX: Texas A&M University Press.

4. Bellah, Robert N. 1967. Civil Religion in America. Daedalus, 134(4).

5. Bellah, Robert N. 1992. The Broken Covenant: American Civil Religion in Time of Trial, 2nd

ed. Chicago, IL: University of Chicago Press.

6. Benoit, William L. 1999. Acclaiming, Attacking, and Defending in Presidential Nominating

Acceptance Addresses, 1960-1996. Quarterly Journal of Speech. 85(3): 247-267.

7. Bond, Jon R., Fleisher, Richard, and Wood, B. Dan. 2003. The Marginal and Time-Varying

Effect of Public Approval on Presidential Success in Congress. Journal of Politics. 65(3): 92-

100.

8. Brace, Paul and Hinckley, Barbara. 1993. Presidential Activities from Truman through

Reagan: Timing and Impact. Journal of Politics, 55(2): 382-398.

9. Cameron, Charles M. 2002. Studying the Polarized Presidency. Presidential Studies

Quarterly, 32(4): 647-663.

27

10. Campbell, Karlyn Kohrs and Jamieson, Kathleen Hall. 1985. Inaugurating the Presidency.

Presidential Studies Quarterly, 15(2): 394-411.

11. Chapp, Christopher B. 2012. Religious Rhetoric and American Politics: The Endurance of

Civil Religion in Electoral Campaigns. Ithaca, NY: Cornell University Press.

12. Christenson, James A. and Wiberley, Ronald C. 1978. Who is Civil Religious? Sociological

Analysis, 39(1): 77-83.

13. Coe, Kevin and Neumann, Rico. 2011. The Major Addresses of Modern Presidents:

Parameters of a Data Set. Presidential Studies Quarterly, 41(4): 727-751.

14. Cohen, Jeffrey E. 1995. Presidential Rhetoric and the Public Agenda. American Journal of

Political Science, 39(1): 87-107.

15. Cohen, Jeffery E., 2004. If the News is So Bad, Why are Presidential Polls So High?

Presidents, the News Media, and the Mass Public in an Era of New Media. Presidential

Studies Quarterly, 34(3).

16. Cohen, Jeffery E., 2011. Presidents, Polarization and Divided Government. Presidential

Studies Quarterly, 41(3).

17. Coleman, John A. 1970. Civil Religion. Sociology of Religion, 31(2), 67-77.

18. Converse, Phillip E. 2006. “The Nature of Belief Systems in Mass Publics (1964)”, Critical

Review, 18:1-3, 1-74.

19. Cristi, Marcela. 1997. On the Nature of Civil and Political Religion: A Reexamination of the

Civil Religion Thesis. Ontario, Canada: University of Waterloo.

20. Cummins, Jeff. 2006. The President’s Domestic Agenda, Divided Government, and the

Relationship to the Public Agenda. The American Review of Politics. 27 (Winter): 269-294.

28

21. Delli Carpini, Michael X. and Keeter, Scott. What Americans Know About Politics and Why

It Matters (Yale University, 1996).

22. Demerath III, N.J., and Williams, Rhys H. 1985. Civil Religion in an Uncivil Society. Annals

of the American Academy of Political and Social Science. 480(Religion in American Today):

154-166.

23. Edwards III, George C. On Deaf Ears: The Limits of the Bully Pulpit (Yale University, 2003)

24. Edwards III, George C. 2009. Presidential Approval as a Source of Influence in Congress.

In George C. Edwards III and William G. Howell (Eds.) “The Oxford Handbook of the

American Presidency” (338-361). Oxford University Press.

25. Eshbaugh-Soha, Matthew. 2010. The Politics of Presidential Speeches. Congress & the

Presidency, 37(1): 1-21.

26. Flere, Sergej and Lavric, Miran. 2007. Operationalizing the Civil Religion Concept at a

Cross-Cultural Level. Journal for the Scientific Study of Religion, 46(4): 595-604.

27. Franklin, Daniel Paul and Fix, Michael P. 2016. The Best of Times and the Worst of Times:

Polarization and Presidential Success in Congress. 43(3): 377-394.

28. Gerring, John. 1997. Ideology: A Definitional Analysis. Political Research Quarterly. 50:4

29. Greenstein, Fred, “Change and Continuity in the Modern Presidency,” in The New American

Political System, ed. Anthony King (Washington, DC: American Enterprise Institute, 1978)

30. Gronke, Paul, Koch, Jeffrey, and Wilson, J. Matthew. 2003. Follow the Leader? Presidential

Approval, Presidential Support, and Representatives’ Electoral Fortunes. Journal of Politics.

65(3): 785-808.

31. Haberski Jr., Raymond. 2012. God and War: American Civil Religion Since 1945. New

Brunswick, NJ: Rutgers University Press.

29

32. Hayes, A. F., 2013. Introduction to Mediation, Moderation, and Conditional Process

Analysis: A Regression-Based Approach. New York, NY, Guilford Press.

33. Hill, Kim Quaile, 1998 “The Policy Agendas of the President and the Mass Public: A

Research Validation and Extension”, American Journal of Political Science, 42:4.

34. Huntington, Samuel P. 1981. American Politics: The Promise of Disharmony. Cambridge,

MA: Belknap Press of Harvard University Press.

35. Johnson, Lyndon B. 1965, March. We Shall Overcome. Speech presented as a Special

Message to Congress, Washington, D.C.

36. Kernell, Samuel. Going Public: New Strategies of Presidential Leadership, 4th edition,

(Washington, DC: CQ Press, 2007)

37. Lee, Frances E. 2008. Dividers, Not Uniters: Presidential Leadership and Senate

Partisanship, 1981-2004. The Journal of Politics, 70:4, 914-928.

38. Lewis, David. 1997. The Two Rhetorical Presidencies: An Analysis of Televised

Presidential Speeches, 1947-1991. American Politics Quarterly. 25(3): 380-395.

39. Pierard, Richard V. and Linder, Robert L. 1988. Civil Religion and the Presidency. Grand

Rapids, Michigan: Academie Books.

40. Reagan, Ronald W. 1988. State of the Union Address. Speech presented before the United

States Congress.

41. Ragsdale, Lyn. 1984. “The Politics of Presidential Speechmaking, 1949-1980”; American

Political Science Review; 78:4 (971-984).

42. Ragsdale, Lyn. 2014. “Vital Statistics on the Presidency”, 4th Edition, Congressional

Quarterly, Washington DC.

30

43. Rhodes, Jesse H. 2014. Party Polarization and the Ascendance of Bipartisan Posturing as a

Dominant Strategy in Presidential Rhetoric. Presidential Studies Quarterly, 44(1).

44. Ritter, Kurt W. 1980. American Political Rhetoric and the Jeremiad Tradition: Presidential

Nomination Acceptance Addresses 1960-1976. Central States Speech Journal. 31(3): 153-

171.

45. Rivers, Douglas, and Rose, Nancy L. 1985. Passing the President’s Program: Public Opinion

and Presidential Influence in Congress. American Journal of Political Science. 29(2): 183-

196.

46. Roof, Wade Clark. 2009. American Presidential Rhetoric from Ronal Reagan to George W.

Bush: Another Look at Civil Religion. Social Compass, 56(2): 286-301.

47. Roosevelt, Franklin D. 1936, June. Acceptance Speech. Speech presented at the Democratic

National Convention, Philadelphia, PA.

48. Schonhardt-Bailey, Cheryl; Yager, Edward; and Lahlou, Saadi. 2012. Yes, Ronald Reagan’s

Rhetoric Was Unique – But Statistically, How Unique? Presidential Studies Quarterly, 42(3):

482-513.

49. Shogan, Colleen J. The Moral Rhetoric of American Presidents. (Texas A&M University

Press, 2007).

50. Sigelman, Lee. 1996. Presidential Inaugurals: The Modernization of a Genre. Political

Communication. 13: 81-92.

51. Skinner, Richard M. 2009. George W. Bush and the Partisan Presidency. Political Science

Quarterly. 123(4): 605-622.

52. Smith, Rogers M. 1993. Beyond Tocqueville, Myrdal, and Hartz: The Multiple Traditions in

America. American Political Science Review, 87(3).

31

53. Teten, Ryan L. 2003. Evolution of the Modern Rhetorical Presidency: Presidential

Presentation and Development of the State of the Union Address. Presidential Studies

Quarterly. 33(2): 333-346.

54. Tetlock, Philip E. 1981. Pre to Postelection Shifts in Presidential Rhetoric: Impression

Management or Cognitive Adjustment? Journal of Personality and Social Psychology, 41(2):

207-212.

55. Thoemmes, Felix J. and Conway III, Lucian Gideon. 2007. Integrative Complexity of 41

U.S. Presidents. Political Psychology, 28(2): 193-226.

56. Toolin, Cynthia. 1983. American Civil Religion from 1789 to 1981: A Content Analysis of

Presidential Inaugural. Review of Religious Research, 25(1): 39-48.

57. Transue, John E. 2007. Identity Salience, Identity Acceptance, and Racial Policy Attitudes:

American National Identity as a United Force. American Journal of Political Science, 51(1):

78-91.

58. Tulis, Jeffrey K. 1987. The Rhetorical Presidency. Princeton University Press, Princeton,

NJ.

59. Ungar, Sheldon. 1991. Civil Religion and the Arms Race. Canadian Review of Sociology

and Anthropology, 28(4).

60. West, Ellis M. 1980. A Proposed Neutral Definition of Civil Religion. Journal of Church and

State. 22: 23-40.

61. Whillock, Rita Kirk. 1994. Dream Believers: The Unifying Visions and Competing Values of

Adherents to American Civil Religion. Presidential Studies Quarterly. 24(2): 375-388.

32

62. Williams, Rhys H. and Alexander, Susan M. 1994. Religious Rhetoric in American

Populism: Civil Religion as Movement Ideology. Journal for the Scientific Study of

Religion, 33(1): 1-15.

63. Williams, Rhys H. and Demerath III, N.J. 1991. Religion and Political Process in an

American City. American Sociological Review, 56(4).

64. Wimberley, Ronald C; Clelland, Donald A; Hood, Thomas C; and Lipsey, C.M. 1976. The

Civil Religious Dimension: Is it There? Social Forces: 54(4).

65. Wimberly, Ronald C. 1980. Civil Religion and the Choice for President Nixon in ’72. Social

Forces. 59(1): 44-61.

66. Wimberly, Ronald C. and Christenson, James A. 1982. Civil Religion, Social Indicators, and

Public Policy. Social Indicators Research, 10(2): 211-223.

67. Wuthnow, Robert. 1988. Divided We Fall: America’s Two Civil Religions. The Christian

Century, 395-399.

68. Zarefsky, David. 2004. Presidential Rhetoric and the Power of Definition. Presidential

Studies Quarterly, 34(3): 607-619.

33

Appendix

(Figure 3 about Here)

Figure 3 provides a visual depiction of the conditional effects of Public Approval on

CRWP at various levels of Senate Margin while holding the House Margin constant. The

upper graph depicts a scenario in which the president enjoys a 10% partisan majority in the

House with each line representing a Senate Margin of -10% (solid line), 0% (dash-dotted

line) or 10% (dashed line). Although the effect of public approval on CRWP is negative

under conditions of united government (dashed line), the preceding analysis demonstrates

that this is only significant under extreme levels of partisan support for the president

(Hypothesis 3). Consistent with the expectations of Hypothesis 1, public approval appears to

have a positive effect on CRWP under conditions of partially divided government (solid line

indicating that the House is controlled by the president’s party and the Senate is controlled by

the opposition), however, this relationship was not statistically significant.

The middle graph depicts a scenario in which the House is evenly split between the

president’s party and the opposition. Although it appears that public approval has a positive

effect upon CRWP in each condition, the preceding analysis demonstrated that the

relationship was only significant when the House and Senate margins both equal 0.

The bottom graph depicts a scenario in which the president’s opponents enjoy a 10%

partisan majority in the House with each line representing either a Senate Margin of -10%

(solid line), 0% (dash-dotted line) or 10% (dashed line). We can see that the effect of

approval on CRWP is negative under conditions of divided government (solid line).

However, this relationship was not statistically significant. Similarly, while it appears that

34

approval has a positive effect upon CRWP under conditions of partially divided government

(Hypothesis 1), this relationship was not statistically significant.

35

Tables and Figures

Table 1: Number of Speeches Delivered by Type and President

President Major Speech Nomination Inaugural State/Union TotalFranklin Roosevelt 5 6.9% 2 9.5% 2 11.1% 6 8.7% 15 8.3%Harry Truman 7 9.7% 1 4.8% 1 5.6% 5 7.2% 14 7.8%Dwight Eisenhower 8 11.1% 2 9.5% 2 11.1% 7 10.1% 19 10.6%John Kennedy 3 4.2% 1 4.8% 1 5.6% 3 4.3% 8 4.4%Lyndon Johnson 5 6.9% 1 4.8% 1 5.6% 6 8.7% 13 7.2%Richard Nixon 6 8.3% 2 9.5% 2 11.1% 4 5.8% 14 7.8%Gerald Ford 2 2.8% 0 0% 0 0% 3 4.3% 5 2.8%Jimmy Carter 4 5.6% 2 9.5% 1 5.6% 3 4.3% 10 5.6%Ronald Reagan 8 11.1% 2 9.5% 2 11.1% 8 11.6% 20 11.1%George H.W. Bush 4 5.6% 2 9.5% 1 5.6% 4 5.8% 11 6.1%Bill Clinton 8 11.1% 2 9.5% 2 11.1% 8 11.6% 20 11.1%George W. Bush 8 11.1% 2 9.5% 2 11.1% 8 11.6% 20 11.1%Barack Obama 4 5.6% 2 9.5% 1 5.6% 4 5.8% 11 6.1%Total 72 40% 21 11.7% 18 10% 69 38.3% 180 100%Notes: Figures represent the number of speeches delivered by each president and the percentage of that type of speech delivered by each president.

36

37

Table 2: Summary Statistics of Variables used in Study

Continuous Variable Mean

SD Min. Max.

Civil Religious Word Percentage .149 .138 0 .689Approval Margin .070 .136 -.240 .390Gross Domestic Product (Annual % Change) .029 .050 -.019 .283House Margin .003 .092 -.170 .270Senate Margin .023 .084 -.150 .260

Binary Variables Freq. % - -Republican 89 49.4 - -POLAR 92 50.8 - -Election 89 49.4 - -War 84 46.7 - -Notes: N Cases = 180 for all variables. Republican, POLAR, Election, and War are binary variables with 1 = the presence of these traits. Freq. and % represent cases where these variables = 1.

38

39

Table 3a: Regression Predicting CRWP

Variable B/(SE)(Constant) .350 (.039)***Rep. .015 (.022)Pres. -.002 (.006)Election .034 (.021)Polar. .056 (.049)War .022 (.020)GDP .000 (.201)Approval Margin .195 (.086)*House Margin -.309 (.196)Senate Margin .298 (.214)Approval*House .235 (1.385)Approval*Senate -.086 (1.500)House*Senate 4.732 (1.435)**Approval*House*Senate

-21.600 (9.550)*SOTU -.305 (.029)***NOM -.257 (.034)***MAJ -.318 (.029)***Adj. R2 .430N 180Inaugural Speeches are omitted to as a reference group. Significance levels are presented as: *<.05, **<.010, ***<.001.

40

Table 3b: Estimated Marginal Means and Differences by Speech Type for CRWP

Speech Type Mean (SE) N NOM (B/SE)

INA (B/SE) SOTU (B/SE)MAJ .106 (.013) 72 MAJ -.061 (.027)* -.318

(.029)***-.013 (.018)

NOM .167 (.024) 21INA .424 (.026) 18 NOM - -.257

(.035)*** .048 (.027)

SOTU .119 (.013) 69Total N 18

0INA - - .305 (.029)***

Notes: Significance levels are presented as: *<.05, **<.010, ***<.001. Covariates were evaluated at their means: Rep = 0.49; Pres = 7.16; Election = 0.19; POLAR = 0.51; App. = 0.07; War = 0.47; GDP = 0.02; House = 0.00; Senate = 0.02; House*Senate = .00; App*House = 0.01; App*Senate = 0.00; App*House*Senate = 0.00

41

42

Table 3c: Conditional Effects of House*Senate Margins on CRWP

Approval Margin B/(SE)-20% 9.052 (3.045)**-10% 6.892 (2.180)**-0% 4.732 (1.435)**10% 2.572 (1.092)*20% 0.411 (1.465)

Significance levels are presented as: *<.05, **<.010, ***<.001.

43

Table 3d: Conditional Effects of Public Approval Margin on CRWP

House Margin Senate Margin B/(SE)-20% -20% -.698 (.417)-10% -10% -.035 (.127)-10% 0% .171 (.137)-10% 10% .378 (.298)0% -10% .203 (.190)0% 0% .195 (.086)*0% 10% .186 (.153)10% -10% .433 (.342)10% 0% .218 (.185)10% 10% -.006 (.099)20% 20% -.639 (.324)*

Significance levels are presented as: *<.05, **<.010, ***<.001.

44

End Notes

45

i More recently, Transue (2007) found that elevating the salience of a shared national identity reduces policy differences among members of different racial groups.ii Although an explicit coupling of this sort does not always occur, the partisan orientation of the President is seldom forgotten. (Edwards III, 2003) It is therefore unlikely that an audience will fail to draw a connection between the President’s rhetoric and his partisan agenda even when the latter is not explicitly mentioned. iii Gerring (1997) makes a strong argument against attempts to construct or employ “all-purpose definitions [of ideology] that can be utilized for all times, places, and purposes.” Rather, we must recognize that there are a variety of arguably valid attributes commonly associated with ideology and scholars should employ a definition that is “context-specific” to their particular research question. In sum, different definitions will be useful for different purposes and the responsible researcher will carefully consider why certain attributes are employed or discarded.iv Gerring (1997) reinforces this perspective by arguing that the primary distinction between ideologies and belief, philosophical, and cultural systems is that they specify a concrete political program and/or issue positions. v Her work does control for the presence of an electoral mandate as measured by a president’s margin of victory in the Electoral College. This term had a statistically significant negative effect on the prevalence of moral rhetoric but fails to capture the president’s standing with the public. viThis term was coined by Aristotle and refers to a “form of rhetoric that praises or blames on ceremonial occasions, addresses an audience that evaluates the rhetor’s skill, recalls the past and speculates about the future while focusing on the present, employs a noble dignified, literary style, and amplifies or rehearses admitted facts” (Campbell and Jamieson, 1985). vii Teten (2003) similarly argues that the Inaugural Address is a “ceremonial speech in which partisan positions and issue propositions are absent, with reverence and general reflection of the past and its unification with the present in their stead.” In contrast, State of the Union Speeches are “essentially the presidential platform from which policy is proposed and evaluated.” viii Chapp (2012) does not provide concrete percentages on these civil religious appeals as they relate to total campaign communications, but does report that religious rhetoric (in general) accounts for less than 1% of all words spoken in campaign speeches and that these civil religious appeals account for 55.8% of that subset. ix The full text of these speeches was obtained from the Public Papers of the President and/or The American Presidency Projectx Given the necessity to control for the influence of public opinion in the subsequent analysis, the data set begins in 1939 when Gallup began to more consistently gather public opinion data. Because speech coding occurred prior to the conclusion of President Obama’s second term, the data set ends in 2012. xi I have excluded speeches which were written but not delivered in person during this time period to avoid any potential confounds associated with a difference between the two delivery formats (Tulis, 1987). xii While the former are mandated by law or custom, the latter are delivered voluntarily in response to situations that arise. xiii Lyn Ragsdale’s “Vital Statistics on the Presidency” (1998; 2014) provides a comprehensive listing of all Major Speeches. Her listing was itself originally obtained from successive volumes of the Public Papers of the President.xiv i.e. I sampled three speeches from John F. Kennedy, four speeches from Ronald Reagan’s first term and four more speeches from his second term. xv Calvin Coolidge’s Inaugural Address (1925) and Herbert Hoover’s State of the Union Address (1932)xvi Given financial constraints, the coders were unfortunately unable to distinguish between civil religious rhetoric that was employed to legitimize ideological positions versus that which was used for other purposes. However, as noted in an earlier footnote, the fact that a speech is delivered by a known partisan (the President) renders a distinction between implicit vs. explicit connections to ideology/policy less important for these purposes. xvii While discussions of how God has guided the U.S. or has intervened on our behalf were included, the end of speech refrain “May God Bless America” was excluded. This was done both to provide a conservative accounting of civil religious rhetoric and because asking for God’s blessing (religion) is different from arguing that God has actually blessed America (Civil Religion). This is consistent with Schonhardt-Bailey et al (2012) who argue that phrases such as “God bless America” or “God bless you” are “vacuous” and less substantive usages of “God”. xviii This list is consistent with Schonhardt-Bailey et al.’s (2012) computer-assisted content analysis that sought terms that represent the “common denominator” and “fundamental basics for civil religious rhetoric”: “mission, sacrifice, destiny, chosen, freedom, divine/providence/spirit/God, American as an international example.” xix To help illustrate the distinction between a historical figure/event that was coded as civil religious and those that were not, consider the following two examples. President Johnson’s 1964 State of the Union Address made the following reference to President Kennedy and it was not coded as civil religious: “Let us carry forward the plans and programs of John Fitzgerald Kennedy – not because of our sorrow or sympathy, but because they are right. In his memory today, I especially ask all members of my own political faith, in this election year, to put your country ahead of your party, and to always debate principles; never debate personalities.” However, in President Johnson’s 1967 State of the Union Address, the following reference to Abraham Lincoln was coded as civil religious: “At the heart of this attack on crime must be the conviction that a free America – as Abraham Lincoln once said – must ‘let reverence for the laws…become the political

religion of the Nation.’”xx The inter-coder reliability of their work was a 90% Agreement and .4 Krippendorf’s Alpha. Although the Krippendorf’s Alpha score did not meet the minimum threshold for that statistic (.67), it should be noted that the presence of civil religion in these speeches is something of a “rare-event” (mean of 15%) and under such conditions disagreements are magnified and overstate the level of error in the coding.xxi This approach (coding of sentences that contain civil religious references) is consistent with Shogan’s (2007) coding methodology for the moral rhetoric of presidents.xxii More specifically, 20% of speeches contained less than %5 CRWP, an additional 26% contained between 5% and 10% CRWP, and an additional 24% contained between 10% and 15% CRWP.xxiii An analysis that excluded these four outliers revealed no substantive differences in the results. xxiv A Shapiro-Wilk test of normality failed to reject the null hypothesis of non-normality with a test statistic of .802. xxv Skewness (1.839; Std. Error .181) and Kurtosis (3.423; Std. Error .360) test statistics fell outside of the normal range.xxvi Alternative models (Poisson and Tobit) were considered, but ultimately rejected because they are not appropriate for the type of data analyzed here.xxvii i.e. 0 represents a condition where there is an equal amount of members of both political parties. 0.05 would represent a condition where the president’s party controls 5% more seats than his opposition. -0.10 would represent a condition where the president’s party controls 10% less seats than his opposition. xxviii Polls were selected as close to the speech date as possible but no more than 6 weeks prior. The vast majority of these statistics were gathered from the Gallup Poll archives at the Roper Center IPoll Database. (http://www.ropercenter.uconn.edu/CFIDE/roper/presidential/webroot/presidential_rating.cfm) In those cases where Job Approval data was not available (i.e. first month in office) Election results were used as a proxy. In other instances where polls were not available, data was gather from Favorability questions or “Who would you vote for questions”. xxix Alternative variables such as a Presidential Job Approval to Disapproval Ratio and a Percentage Monthly Change in Job Approval Rating were evaluated but found to be less effective for the models presented in this study. xxx Unfortunately, the literature on civil religion is less clear on how these conditions may influence the use of this rhetoric. While there are commendable studies on the content of civil religious rhetoric during military conflicts (Haberski, 2012; Ungar, 1991) and political campaigns (Chapp, 2012), these studies do not evaluate whether these conditions systematically influence the use of civil religion. xxxi This variable excluded conflicts in which there were there was a relatively small number of ground troops involved and/or where the mission was primarily peace-keeping (i.e. Lebanon, 1982; Somalia, 1993; Bosnia, 1994; and Kosovo, 1999). The conflicts included were: World War II, Korean War, Vietnam War, Invasion of Grenada, Invasion of Panama, Gulf War, Invasion of Haiti, and the Wars in Afghanistan and Iraq. Alternative variables such as Active Duty Military Personnel (excluding National Guard and Reserves) for each year and Percentage change in Active Duty Military Personnel from the last year were found to be less effective in the models presented in this study. xxxii Data obtained from the Bureau of Economic Analysis (US Department of Commerce) (www.bea.gov/iTable/index nipa.cfm). These figures reflect the percentage change from the previous data point; seasonally adjusted annual rates from 1929-1946 and quarterly rates from 1947-2013. Alternative economic indicators such as the Unemployment Rate, Consumer Price Index, and Income Shares for the top 10% (proxy for inequality) were considered but found to be less effective in the models presented in this study.xxxiii Cameron (2004) defines this condition as a situation “when politics is polarized and control of government is divided by party.”xxxiv It should be noted that this particular relationship may warrant further analysis since it is measured rather crudely in this study. One could imagine more nuanced data on the varying levels of polarization within each particular Congress (i.e. Poole-Rosenthal Data Set) providing a more refined test on this potential relationship.