The relationship between church and state has long been a contentious issue in the United States. Previous studies have argued that despite the constitutional separation, religious ideology pervades social and political institutions (Corbin, 2012; Friedenburg, 2002; Smidt, 2007; Steinfels, 2007). While the First Amendment to the United States’ Constitution protects freedom of religion and denotes a separation of church and state, the historical influence of religion in laws, social policy, and political representation have left nonreligious individuals feeling excluded (Gresock, 2001). There is also a social stigma placed on the nonreligious as immoral, unpatriotic, and less trustworthy (Corbin, 2012; Cragun et al., 2012; Edgell et al., 2006; Gervais & Norenzayan, 2012; Zuckerman, 2009). As a result of such social exclusion, this study examines whether atheists and nonreligious Americans, compared to those who identify as religious, have lower levels of confidence in institutions in the United States.

This study is particularly timely as the contemporary religious landscape in the United States is undergoing rapid changes, with an increasing proportion of religiously unaffiliated (Pew Research Center, 2015a; Sherkat, 2014; Zuckerman, 2014). Additionally, the United States is seeing a sharp decline in confidence in social and political institutions (Abrajano & Alvarez, 2010; Chanley, 2002; Rudolph & Evans, 2005). Since confidence in major social institutions is a crucial measure of a successful democracy (André, 2014; Inglehart, 1999; Newton & Norris, 1999), the results of this investigation can provide a better understanding of the importance of interreligious respect and awareness that also incorporates the nonreligious.

Utilizing data from the four most recent waves (1995, 1999, 2006, and 2011) of the World Values Survey’s data on the United States, we examine the relationship between religiosity and confidence in the armed forces; major companies; churches; the government; the police; and political parties. We also explore the extent to which these relationships are explained by year of the survey, church attendance, political orientation, race, gender, age, income, marital status, level of education, and general social trust. We predict that, compared to those who are religious, individuals who identify as nonreligious or as atheist will have lower levels of confidence in all six institutions.

In the next section, we examine key literature on institutional confidence and religion. This includes a discussion of the changing trends in both religion and civic engagement in the United States. This is followed by a description of our methods, including a detailed account of our data, hypotheses, and variables. In the fourth section, we present our findings and discuss their implications. Finally, we conclude with a summary of our study and propose suggestions for future research.

Institutional Confidence in the United States

Since the 1960s, the United States has experienced a drastic decline in confidence in political and social institutions (Abrajano & Alvarez, 2010; Chanley, 2002; Dalton, 2005; Hetherington, 1998; Hetherington & Rudolph, 2008; Rudolph & Evans, 2005). In 1958, around 7 out of 10 Americans trusted the government “to do what is right”, a figure that declined to 3 in 10 in the 1990s (Chanley, 2002). After a temporary upswing in political trust after September 11 (Chanley, 2002), trust in the government is now at an all-time low at 19 percent of the population (Gao, 2015). Furthermore, confidence and trust in other U.S. institutions also seems to be eroding. For example, in 2015, unfavorable opinions of the U.S. Supreme Court reached a record high of 43 percent of the American population (Pew Research Center, 2015c). Data from Gallup (2016) found that the number of respondents answering none to very little confidence has risen across different institutions. For example, this lack of confidence has increased from 13 to 23 percent for churches or organized religion, from 22 to 53 percent for congress, and from 20 to 40 percent for the president.

The Vietnam War, the Civil Rights Movement, and Watergate are recognized as potential causes for the initial drop in trust and confidence (Abrajano & Alvarez, 2010; Markus, 1979), but the reason for the continued decline is under debate. Scholars offer a range of explanations for the persistence of this decline, such as a general decrease in civic engagement (Blind, 2006; Miller, 1980; Putnam, 2000), political polarization (Jones, 2015), economic inequality (Blind, 2006; Citrin & Green, 1986), dissatisfaction with public policy (Chanley, 2002; Miller, 1974; Miller & Listhaug, 1998; Mitchell & Scott, 1987), and an increasing populace of skeptics (Cook & Gronke, 2005).

Trust and confidence in major social institutions is a critical element of a healthy and thriving democracy (André, 2014; Inglehart, 1999; Newton & Norris, 1999). Confidence and trust are part of social capital, which refers to the connections between individuals (Bourdieu, 2011). Putnam (2000, pg. 147) explains that such connections and networks generate “norms of reciprocity and trustworthiness”. Engaging with Putnam’s work, Lowndes & Wilson (2001, pg. 630) state that, “People learn to trust one another through face-to-face interaction in associations and informal social networks; norms of trust and reciprocity ‘spill over’ into society at large; a capacity is created for collective action in pursuit of shared goals; citizens expect, and representatives provide, competent and responsive government”.

For the purpose of this study, we examine confidence in political and social institutions. Some studies (e.g. Grönlund & Setälä, 2012; Weaver, 2003) treat trust and confidence in institutions as synonyms of the same construct. Others argue that the two terms measure two distinct concepts that are nonetheless closely related. Generally, a lack of trust relates to lower levels of confidence in institutions (Cook & Gronke, 2005; Siegrist, 2010). Offe (1999) asserts that trust is a more appropriate term when the object is individual actors, while confidence should be used when examining efficiency and legitimacy of institutions. The author argues that institutions are “factual constraints of action, the durability and validity of which we can view with confidence. Trust, in contrast, can only be extended to actors and the ways in which they perform and enact their roles within institutions” (Offe, 1999, pg. 45). Additionally, previous literature (e.g. Cook & Gronke, 2005) has argued that given the range of institutions that fall under state governance, an examination of institutional confidence is multifaceted in nature. Therefore, it is important to explore multiple institutions and dimensions. This is consequently the aim of this study, where we examine confidence in the government, the armed forces, the police, churches, major companies, and political parties.

Religion and Institutional Confidence

Despite a recent decline in religious affiliation (Pew Research Center, 2015a), the United States remains the most religious country in the West (Zuckerman, 2014). Sherkat (2014, pg. 1) describes that, in the United States, religion “influences whom we marry, how we raise our children, our educational and occupational opportunities and choices, and our moral and political commitments”. According to Pew Research Center (2016), 70.6 percent of Americans identify as Christian, which is comprised of 25.4 percent Evangelical Protestants and 20.8 percent Catholics. Non-Christian faiths make up 5.9 percent of the population with 1.9 percent Jewish and 0.9 percent Muslim. Unaffiliated religious “nones” make up 22.8 percent of the population, with 15.8 percent identifying as “nothing in particular”, 4 percent as agnostic, and 3.1 percent as atheist. Moreover, we are currently seeing significant transitions in the religious landscape in the United States. Particularly noteworthy is the increase in the religiously unaffiliated. The percentage of Americans who identify with no religion has increased from 15 percent in 2007 to 22.8 percent in 2015 (Pew Research Center, 2015a). This trend of disaffiliation highlights the relevance and timeliness of studying how secular Americans view social and political institutions.

Findings from previous research on religion and social and political trust and confidence are mixed. Some scholars maintain that religious participation can, in certain cases, increase social trust (Mencken & Fitz, 2013; Welch et al., 2004). Mencken and Fitz (2013) state that religious participation can build trust in the community, which encourages volunteering and helping others. Welch et al. (2004) show mixed results regarding the relationship between religious affiliation and trust, but found that Pentecostals who are frequent participants have higher levels of trust. Other studies (e.g. Alesina & Ferrara, 2002; McCleary & Barro, 2006) found no relationship between religion and trust. Conversely, some studies (e.g. Berggren & Bjørnskov, 2011; Putnam & Campbell, 2010) present a negative relationship between religiosity and general social trust in the United States. Berggren and Bjørnskov’s study (2011) also shows the same trend at the cross-national level. They conclude that religiosity is associated with high levels of trust within their own social group, but that, given the notion that out-groups are morally inferior, the trust toward individuals of other religious affiliations is lower.

In relation to institutional trust and confidence, studies on various contexts (e.g. Abramson, 1983; Alvarez et al., 2008; de Vroome et al., 2013; Hero & Tolbert, 2004; Kasselstrand & Kandlik Eltanani, 2013; Levitt, 2014; Tyler, 2005) have presented lower levels of institutional confidence and trust among minority groups more broadly. Discussing the effects of religious affiliation on trust in the state, Kasselstrand and Kandlik Eltanani (2013) argue that the history of government exclusion of religious minorities can explain their lower level of political trust in the Nordic countries. Following this argument, we maintain that the historical relationship between church and state in the United States is a crucial aspect of understanding institutional confidence among secular Americans today.

Despite a constitutional separation of church and state, religion is thoroughly entrenched in the sociopolitical landscape in the United States. Historically, Christianity shaped the early republic (Davis, 1994; Friedenburg, 2002). For instance, there are religious undertones in the Declaration of Independence as it mentions a “Creator”: “We hold these truths to be self-evident, that all men are created equal, that they are endowed by their Creator with certain unalienable Rights, that among these are Life, Liberty and the pursuit of Happiness” (The Declaration of Independence, 1776, Para. 2). The Boisi Center for Religion and American Public Life (2007) uses the inclusion of “One Nation Under God” in the Pledge of Allegiance as well as the fact that churches are exempt from paying income tax in the United States as more contemporary examples of the remaining bonds between church and state. Along these lines, Corbin (2012) explains that, while unenforceable, several state constitutions require that public positions are filled by believers in God. The underrepresentation of secular Americans in politics is further reflected in the fact that only one elected official (0.2 percent) in the 114th Congress is nonreligious (Pew Research Center, 2015b).

Zuckerman et al. (2016) explain that only one in five of the unaffiliated agree that America is the greatest country in the world, whereas 40 percent of Evangelical Christians show support for this statement. Cook and Gronke (2005, pg. 785) state that “low trust in government and low confidence in institutions reflect skepticism, an unwillingness to presume that political authorities should be given the benefit of the doubt”. Thus, the levels of political trust may continue to decrease as secularity increases since the nonreligious are, overall, not willing to put their trust into an institution without proof of proficiency. However, Kasselstrand and Kandlik Eltanani (2013) found that, in the Nordic countries, the religiously unaffiliated have higher levels of trust in political institutions than individuals who identify with Protestant free churches, suggesting a potential limitation to using a skeptical disposition as the core explanation for a lack of institutional confidence among secular Americans. Instead, we argue that a lack of institutional confidence among secular Americans is, at least in part, a result of their minority status in the United States, particularly given the evidence presented below that they are one of the most stigmatized minority groups in the country. This is in stark contrast to the Nordic context where “religion has very little political sway; non-religious people are not maligned or mistrusted; and the church is not the pervasive center of most people’s social lives” (Zuckerman, 2012b, pg. 19).

Stigmatization of Secular Americans

When compared to other marginalized minority groups, atheists and secular individuals remain some of the most distrusted groups in the United States (Corbin, 2012; Edgell et al., 2006; Franks & Scherr, 2014; Gervais et al., 2011; Gervais & Norenzayan, 2012; Zuckerman, 2009). Franks and Scherr (2014) maintain that while other minority groups have seen a substantial increase in acceptance over the last few decades, the acceptance of atheists has only increased marginally. The authors further state that atheists “suffer the greatest disadvantage as political candidates because they received the lowest levels of voting intentions and elicited fear, disgust, and strong levels of distrust” (Franks & Scherr, 2014, pg. 687). In terms of perceived marginalization and discrimination among atheists, Hammer and colleagues (2012) mention that a typical source of stress for atheists is the lack of the social and organizational resources that religious individuals access. They further note that many atheists feel like there is “unreciprocated tolerance” in the United States where “religious and other forms of tolerance are generic values in American culture, but these same cultural ideals may not be applied to atheists” (Hammer et al., 2012, pg. 55).

Studies have shown that people with an established religion, even if it is a minority religion, believe that the lack of religion makes an individual morally inferior, less patriotic, and less trustworthy. In fact, belonging to a minority religion is considered superior to having no religion at all (Cobin, 2012; Gervais et al., 2011). Data from Pew Research Center (2014) that explore how warmly Americans view different religions show that Buddhists and Hindus receive neutral ratings, being viewed more warmly than atheists and Muslims, but less so than the Judeo-Christian religions. The particularly low acceptance of atheists was further noted by Edgell et al. (2006, pg. 230) who state that “It is striking that the rejection of atheists is so much more common than rejection of other stigmatized groups. For example, while rejection of Muslims may have spiked in post-9/11 America, rejection of atheists was higher”. Franks and Scherr (2014) further note that, compared to other secular individuals, feelings of discrimination may be higher among atheists. However, the authors clarify that the primary source of the prejudice appears to be a lack of a belief in god rather than the atheist label itself. This suggests that the religiously unaffiliated may face lower levels of prejudice and discrimination than explicit atheists and nonbelievers in God. Also exploring anti-atheist prejudice, Cragun et al. (2012, pg. 108) describe atheists as having “a more pronounced out-group status” than nonbelievers and the unaffiliated.


Hypotheses and Data

The aim of this study is to examine the relationship between secular identities and confidence in political and social institutions. We theorize that secularity is a key factor that influences institutional confidence. More specifically, we believe that alienation and discrimination of the nonreligious (e.g. Cragun et al., 2012; Edgell et al., 2006; Franks & Scherr, 2014; Zuckerman, 2009) may influence confidence in U.S. institutions. Following Edgell et al.’s (2006) findings that secular Americans face particularly high levels of stigma and exclusion, we hypothesize that those who identify as atheist and nonreligious have lower levels of confidence in political and social institutions in the United States compared to those who identify as religious. However, with self-identified atheists being more likely to feel discriminated against and with the evidence that secular prejudice is often grounded in a lack of belief in God (Franks & Scherr, 2014) rather than in institutional disaffiliation, we believe that atheists will have lower levels of confidence in social institutions than those who identify as nonreligious.

For our analysis, we used data from the World Values Survey. This randomly sampled data is available for multiple countries throughout the world and has been collected in six waves between 1981 and 2014. The four most recent waves (1994–1998, 1999–2004, 2004–2009, and 2010–2014) were used for this this study as they contained the variables of interest. We focused on the data from the United States only, which had a sample size of 6223 individuals over the four waves (with data collected in 1995, 1999, 2006, and 2011). The data were analyzed using contingency tables and binary logistic regressions. Contingency tables allowed us to look at the association between our two key variables of interest over time. Binary logistic regression was used to get a more complex, nuanced picture of the relationship between our independent and dependent variable while controlling for a range of relevant factors.


As a response to the decline in confidence in political and social institutions in the United States (Abrajano & Alvarez, 2010; Chanley, 2002; Dalton, 2005; Rudolph & Evans, 2005), the six dependent variables provide measures of confidence in various institutions: confidence in the armed forces; confidence in major companies; confidence in churches; confidence in the government; confidence in the police; and confidence in political parties. In the World Values Survey, respondents were asked to rate their confidence in these institutions using the following attributes: a great deal of confidence; quite a lot of confidence; not very much confidence; or none at all. All of the dependent variables were recoded into dummy variables for the analysis with 0 meaning none at all or not very much confidence and 1 corresponding with quite a lot or a great deal of confidence.1

As presented in Table 1, the highest level of confidence is seen in the armed forces (with more than 8 in 10 respondents having at least quite a lot of confidence). This is followed by the police and churches. The respondents have the lowest levels of confidence in political parties (with confidence ranging from 12.8% to 21.2% over the four waves of the survey). Four of the six institutions (the armed forces, churches, major companies, and political parties) have experienced a significant decline in confidence over the last two decades. The most noteworthy decline can be seen for churches. In 1995, 76 percent of the respondents had confidence in churches. In the most recent wave, the same figure was 58.6 percent, following the trend of religious disaffiliation in the United States (Pew Research Center, 2015a).

Table 1

Percentage of respondents who have quite a lot or a great deal of confidence in U.S. institutions.

1995 1999 2006 2011 Sig.

Armed Forces 86.4% 81.7% 82.3% 83.2% ***
Churches 76.0% 74.6% 66.3% 58.6% ***
Major Companies 53.8% 54.1% 26.7% 33.6% ***
Government 30.6% 37.8% 38.2% 33.3% ***
Police 71.2% 71.4% 70.4% 69.4%
Political Parties 21.2% 22.6% 15.3% 12.8% ***

*p < 0.1 **p < 0.05 ***p < 0.01.

The independent variable in our study is secularity. This variable is based on the survey question that asks: “Independently of whether you attend religious services or not, would you say you are: (1) a religious person; (2) not a religious person; or (3) an atheist”. A majority of the respondents in the overall sample (74.8 percent) stated that they are a religious person. Not a religious person was the second largest group (22.4 percent), with atheists as the smallest group (2.8 percent). However, in later waves of the survey, the share of atheists and nonreligious respondents was higher than for earlier waves. For the findings presented in Table 2 that explore differences in institutional confidence between secular and religious individuals over time, this variable was coded as 1 for secular, which includes both respondents who answered not a religious person and atheist and 0 for a religious person. The reason for combining the two secular categories for this part of the analysis was due to the very small samples of atheists when trends were examined across different years of the survey.

Table 2

Percentage of religious and secular respondents who have quite a lot or a great deal of confidence in U.S. institutions.

Armed Forces Churches Government

Religious Secular Sig. Religious Secular Sig. Religious Secular Sig.

1995 88.3% 81.0% *** 83.5% 46.0% *** 30.9% 30.5%
1999 83.5% 72.9% *** 82.1% 41.7% *** 38.6% 33.3%
2006 85.4% 75.9% *** 78.6% 34.1% *** 41.2% 30.6% ***
2011 86.4% 76.6% *** 72.8% 28.6% *** 36.1% 27.4% ***
Political Parties Police Major Companies

Religious Secular Sig. Religious Secular Sig. Religious Secular Sig.

1995 22.0% 19.3% 72.5% 66.4% ** 55.1% 50.4%
1999 23.4% 18.9% 72.8% 63.9% ** 52.9% 60.4% *
2006 16.9% 11.9% ** 71.6% 67.2% 28.8% 22.1% **
2011 14.5% 9.2% *** 73.0% 62.4% *** 36.6% 27.3% ***

*p < 0.1 **p < 0.05 ***p < 0.01.

For logistic regression analysis, the larger combined sample allowed for a differentiation of atheists and nonreligious respondents. This distinction makes it possible to examine nuances in institutional trust across two different categories of secularity. This approach is beneficial given that previous research has found that while secular Americans share some key characteristics, they are far from a homogeneous group (Baker & Smith, 2009; 2015; Zuckerman, 2014). For example, Baker and Smith (2009) found that both atheists and unchurched believers are more strongly opposed to religion in the public sphere than are agnostics. For the regression analysis, the independent variable therefore consists of two dummy variables separating those who identify as atheist, nonreligious, and religious. One dummy variable is called nonreligious and was coded as 1 for individuals who are not a religious person. The second dummy variable is called atheist and was coded as 1 for those who identify as atheist. This makes the category a religious person the reference category.

We added a total of ten control variables to our regression models. Church attendance was included as previous studies have suggested that there is a relationship between religious participation and social trust (Berggren & Bjørnskov, 2011; Mencken & Fitz, 2013; Putnam & Campbell, 2010; Welch et al., 2004). Since nonreligious individuals may have a more skeptical outlook than those who are religious (Devos et al., 2002; Zuckerman, 2012), we controlled for general trust. While not necessarily one and the same, distrust and skepticism tend to go hand in hand (Cook & Gronke, 2005). The models also controlled for year of the survey following the decline in institutional confidence in the United States over time (Abrajano & Alvarez, 2010; Chanley, 2002; Rudolph & Evans, 2005).

Previous studies (e.g. Baker & Smith, 2015; Keysar, 2007; Zuckerman et al., 2016) have explored key demographic characteristics of secular Americans. They found that men, millennials, European Americans, and Asian Americans are overrepresented among the secular. On average, they also hold higher levels of education. In terms of political affiliation, the largest proportion of the secular identify as Independent. As such, our models include a range of demographic variables in order to study the relationship between secularity and institutional confidence while accounting for such distinct characteristics.

More specifically, we control for political orientation, as it is closely associated with religious identity (Zuckerman, 2012a) and institutional confidence (Zmerli, 2006). For example, secular Americans are generally “more liberal and progressive than their religious peers, being less likely to support the death penalty, the War in Iraq, the governmental use of torture and more likely to support Democratic candidates, women’s equality and gay rights” (Zuckerman, 2012a, pg. 10). Due to younger individuals being more likely to be nonreligious (Keysar, 2007; Zuckerman et al., 2016), we controlled for age. Race is closely associated with both religious affiliation and institutional power structures (e.g. Edwards, 2008). The remaining control variables: income, marital status, gender, and education have been shown in the literature to correlate with a person’s religious identity (e.g. Kasselstrand, 2015; Merino, 2012; Zuckerman et al., 2016; Wolfinger & Wilcox, 2008).

The variable age is measured in years. Income constitutes a scale of ten steps. Political orientation is measured on a left-to-right scale between 1 and 10. Yearof Survey was coded as the year that the data for that wave was collected in the United States. Trust is coded as 1 for respondents who generally feel that they can trust people most of the time and 0 for respondents who generally feel that they cannot trust people most of the time. Marital status is coded as 1 for respondents who are married and 0 for respondents who are not married. Gender was coded as 1 for respondents who are female and as 0 for respondents who are male. The variable education was coded as 1 for university degree and 0 for less education. Attendance is coded as 1 for those who attend church at least monthly and 0 for less often than monthly. Finally, race was coded as 1 for white and 0 for not white.

Findings and Discussion

Secularity and Institutional Confidence Over Time

Table 2 displays the percentage of respondents who have quite a lot or a great deal of confidence in the specified institutions during a particular wave of the WVS. The table also shows differences in confidence between secular and religious respondents. The results suggest that secular individuals are, compared to those who identify as religious, less likely to have confidence in all institutions in our study, for every wave of the survey, with the single exception of major companies in 1999. Moreover, all differences are statistically significant, except for confidence in government and political parties in 1995 and 1999, the police in 2006, and major companies in 1995.

Predictably, the largest gap in confidence between religious and secular respondents is in churches with a difference ranging from 37.5 to 44.5 percentage points across the four waves of the survey. For the secular institutions, the largest difference can be seen for confidence in the armed forces, with an average difference of 9.3 percentage points over the four waves. However, both groups show high levels of confidence in the armed forces (an average of 85.9 and 76.6 percent respectively), but the difference between the two groups is noteworthy. After churches and the armed forces, the largest difference in confidence between religious and secular Americans are for the police (with a difference of 7.5 percentage points) and the government (6.2 percentage points). While there has not been an overall decline in confidence in the government since 1995, the difference between religious and secular respondents demands further attention. In 1995, the percentage of religious and secular individuals who had confidence in the government was nearly equal (30.9 and 30.5 percent respectively), but has since then grown substantially, with the largest difference (10.6 percentage points) observed during the 2006 wave. This difference remained statistically significant but weakened slightly in 2011 (8.7 percentage points). Yet, among secular respondents, confidence in government was, at this time, at the lowest level across all four waves (27.4 percent).

Another key finding displayed in Table 2 is that secular respondents overall seem to follow the same trend as religious respondents in terms of the decline in confidence. As seen in Table 1, over the last two decades, there has been a significant decline in confidence in all institutions in the study except in the police and the government. For the four institutions where a significant decline was observed (the armed forces, political parties, churches, and major companies), both secular and religious respondents seem to contribute to this decline in confidence. The fact that, between the first and the fourth wave of the survey, confidence in churches declined by 10.7 percentage points among religious participants is striking. However, among secular respondents, the decline was even larger at 17.4 percentage points.

As secular respondents overall have lower levels of confidence in the institutions in the study, and with a majority of such differences being statistically significant, these findings offer support for our hypothesis that secular Americans are less confident in social institutions compared to those who are religious. However, while the findings presented in Table 2 provide some indication of lower confidence in institutions among the secular, more comprehensive analysis that takes other factors into account is required to better understand the relationship between religious identity and institutional confidence. This is presented below with the results from binary logistic regression analysis, where this relationship is examined while controlling for key demographic characteristics of the respondents.

Binary Logistic Regression Models

Twelve binary logistic regression models are presented in Table 3. The table displays odds ratios and significance levels for each model. Two models were created for each dependent variable in order to show the change in confidence between atheists, the nonreligious, and the religious before and after controlling for church attendance and general trust. As odds ratios below 1 indicate a negative relationship between the independent or control variable and the dependent variable, the variable year confirms that for three institutions (major companies, churches, and political parties), there has been a decline in institutional confidence over the last two decades just as presented in the literature and in Tables 1 and 2.

Table 3

Binary Logistic Regression: Quite a lot or a great deal of confidence (Odds Ratios).

Armed Forces Major Companies Churches

Nonreligious 0.692*** 0.746*** 0.845** 0.867** 0.183*** 0.290***
Atheist 0.268*** 0.291*** 0.791 0.807 0.056*** 0.110***
Year 1.004 1.005 0.943*** 0.943*** 0.958*** 0.966***
Education 0.976 0.943 1.138* 1.105 0.923 0.770***
Female 0.881 0.857* 0.819*** 0.805*** 1.220*** 1.109
Married 1.243*** 1.231** 1.024 1.013 1.122* 0.937
White 1.935*** 1.931*** 1.090 1.050 0.835** 0.972
Age 1.012*** 1.012*** 1.002 1.001 1.005** 1.003
Politics 1.232*** 1.226*** 1.135*** 1.134*** 1.190*** 1.146***
Income 0.992 0.984 1.080*** 1.072*** 1.005 0.997
Trust 1.260*** 1.351*** 1.381***
Attendance 1.210** 1.079 5.489***
Cox & Snell 0.058 0.061 0.070 0.074 0.191 0.271
N 5430 5373 5332 5277 5430 5375
Government Police Political Parties
Nonreligious 0.725*** 0.737*** 0.773*** 0.832** 0.712*** 0.755***
Atheist 0.697** 0.718* 0.508*** 0.553*** 0.686 0.753
Year 1.005 1.006 1.006 1.007 0.962*** 0.963***
Education 1.049 0.987 1.214** 1.135 0.830** 0.790**
Female 1.075 1.051 1.180*** 1.128* 1.054 1.036
Married 1.029 1.032 1.238*** 1.236*** 0.929 0.923
White 0.734*** 0.694*** 2.390*** 2.286*** 0.740*** 0.725***
Age 0.995*** 0.993*** 1.007*** 1.006*** 0.998 0.996
Politics 1.012 1.011 1.122*** 1.119*** 1.068*** 1.061***
Income 0.999 0.989 1.035** 1.019 1.032* 1.028
Trust 1.652*** 1.797*** 1.519***
Attendance 1.048 1.223*** 1.145*
Cox & Snell 0.011 0.023 0.065 0.079 0.021 0.026
N 5404 5349 5402 5346 5381 5325

*p < 0.1 **p < 0.05 ***p < 0.01.

Findings from regression analysis further show that those who are nonreligious are significantly less likely than those who identify as religious to have confidence in all six institutions. This holds true in both models for each dependent variable, with or without controlling for trust and church attendance. Atheists, compared to those who are religious, are also significantly less likely to have confidence in four of the six institutions (armed forces, government, churches, and police). However, it is important to note that had there been a larger number of atheists in the sample, the difference between atheists and the religious may have been significant for major companies and political parties as well. The reasoning behind this is that for all models, the odds ratios are smaller for atheists compared to the nonreligious, suggesting that atheists appear less likely than the nonreligious to hold confidence in the six institutions. This goes hand in hand with the notion that atheists hold a more salient out-group status than those who identify as nonreligious (Cragun et al., 2012). Yet, the nonreligious are still less likely than those who are religious to have confidence in these institutions, suggesting that the difference in confidence between the secular and the religious cannot solely be explained by lower levels of confidence among self-identified atheists.

Given the link between secularity and skepticism (Devos et al., 2002; Zuckerman, 2012a) and religious participation and social trust (Berggren & Bjørnskov, 2011; Mencken & Fitz, 2013; Putnam & Campbell, 2010; Welch et al., 2004), two factors that are particularly important to discuss are the effects of general trust and church attendance. When these two factors are introduced as control variables, the negative relationship between secularity and confidence weakens, a finding that holds true for both atheists and nonreligious individuals. This can be seen by comparing the odds ratios between the two models for each dependent variable. This means that general trust and church attendance may, in part, explain the lower levels of confidence among the secular. However, as secular individuals are more likely to lack confidence in institutions even after taking religious participation and general trust into account, this suggests that this absence of institutional confidence cannot exclusively be explained by a lack of general trust or the alleged absence of social ties when not being part of a religious community. It is also important to mention that for all six dependent variables, the models improved when the variables general trust and church attendance were added. However, with low Cox & Snell R-squared statistics across all models, but in particular for government, political parties, and churches, much of the variation in institutional confidence can be explained by factors that are not included in the models.

In the full models, the strongest difference between religious and secular respondents can be observed in terms of confidence in churches (odds ratios of 0.110 for atheists and 0.290 for the nonreligious). For atheists, this is followed by the armed forces (OR = 0.291), and the police (OR = 0.553). For all institutions except for churches, the odds ratios that measure the difference in confidence between the nonreligious and the religious range from 0.737 and 0.867. It is important to note that the disparities in institutional confidence found in Table 2 remain after accounting for age, gender, marital status, income, educational level, race, political orientation, year of the survey, general trust, and church attendance. This offers key evidence for the distinctive opinions and experiences of secular Americans.

Church attendance is a significant predictor of confidence in the armed forces (OR = 1.210), the police (OR = 1.223), and churches (OR = 5.489). Furthermore, all six models show a significant relationship between general trust and institutional confidence (with odds ratios ranging from 1.260 to 1.797). Other key factors that may shape the attitudes of secular Americans are political orientation, gender, race, income, age, marital status, and education. A position to the right on a left-to-right scale of political orientation significantly predicts higher levels of confidence in all institutions except the government. Having a college degree is associated with lower confidence in political parties and churches. Women are less confident in the armed forces and in major companies but more confident in the police. Older individuals are more likely than those who are younger to be confident in the armed forces and in the police, but less likely to have confidence in the government. Similarly, compared to racial and ethnic minorities, white Americans are more likely to be confident in the armed forces and in the police, but less confident in the government and in political parties. Finally, married individuals are more likely to be confident in the armed forces and in the police. Again, the relationship between religious identity and institutional confidence remains even after controlling for such factors, and the lack of institutional confidence among nonreligious Americans appears to span across a range of different institutions unlike what is observed among other minority identities (e.g. race and gender) in this study, which may be an indicator of the feelings of exclusion, prejudice, and discrimination of secular Americans as a a minority group at a time when most other minorities have seen a substantial gains in social acceptance (Franks & Scherr, 2014).


With this study, our intention was to answer the following research question: Is there a relationship between secularity and confidence in various social and political institutions? Overall, the empirical evidence presented above does show a relationship between secularity and institutional confidence. More specifically, the results from both the contingency table and the binary logistic regressions suggest that secular Americans are less likely than those who are religious to have confidence in the armed forces, major companies, churches, the government, the police, and political parties.

Suggestions for future research include expanding the study of this topic to additional countries with different religious and sociopolitical landscapes. This could provide a more comprehensive understanding of the role of culture, prejudice, and stigma in shaping social trust and social and political attitudes. In addition, in order to obtain a more detailed portrait of how individuals perceive the relationship between religious identity and institutional confidence, qualitative studies on the topic would provide a more nuanced understanding of the role that nonreligiosity plays in shaping confidence in social institutions. Finally, a limitation to this study lies in the conceptualization and measurement of confidence, trust, and skepticism. Additional survey data that involve a wide range of indicators that operationalize such concepts would be beneficial.

A healthy democracy depends upon social capital and civic engagement in order to thrive. With historically low levels of confidence in institutions in the United States (Gao, 2015), our study adds to the literature on predictors of such confidence by focusing on the effect of a secular identities. This investigation is particularly pertinent given the recent increase in the proportion of Americans who are religiously unaffiliated (Pew Research Center, 2015a). Findings reveal potential consequences of the challenges that secular individuals face in the United States today. As religion influences laws, policies, and political representation and as secular Americans face prejudice and stigma (Corbin, 2012; Cragun et al., 2012; Edgell et al., 2006; Gervais & Norenzayan, 2012; Zuckerman, 2009), lower levels of institutional confidence may be an outcome of such exclusion from American social life. In conclusion, the findings highlight the need for further attention to the inequalities, discrimination, and prejudice that come with the minority status of being secular in the contemporary United States.