Institutional Trust, Spirituality, and Religious Practice in the United States

Abstract

This study explores the relationship between institutional trust and elements of religion in contemporary American society. Using a nationally representative survey, we utilize latent class and regression analyses to assess the contours of these relationships. Our findings reveal a diverse institutional trust profile, with pronounced distrust in democratic institutions, especially Congress, although the group with low institutional trust is the smallest one within the society. While trust in religious institutions, and to an extent broader institutional trust, predict religiosity, this is less so with the case of spirituality. Institutional trust similarly shows modest positive relationships to public religious practice, while a clear negative relationship to private religious practice. Interpreting these findings through the lens of secularization and privatization of religion, we hope to provide a strong empirical contribution to the literature regarding intersections of institutional trust and the evolving religious and spiritual orientations of today’s American landscape.

1. Introduction

The concept of trust within society stands as one of the core elements for social cohesion and the untroubled functioning of societies (Fukuyama 1995). One of the most prominent components of societal trust is institutional trust, or the amount of confidence that individuals have that institutions and their representatives are functioning in the way that the individual feels that they ought to (Fukuyama 1995; Rothstein and Stolle 2008). However, institutional trust is ultimately argued to exist when one’s belief in the functioning of said institution is not compromised by the actions of a single member of that institution whose conduct does not live up to the expectations of the individual, and thus they still trust in the institution (Bachmann 2021). This trust is thus informed by but also transcends personal interactions, permeating and shaping societal structures and organizations, such as political bodies, financial institutions, and critically for this paper, religious institutions, or the institution of organized religion broadly conceived (e.g., Fukuyama 1995; Granovetter 1985; Putnam 2000; Uslaner 2002; Zak and Knack 2001).

Religion is arguably one of the most influential institutions, even in many rapidly secularizing contemporary societies, because it is often regarded as a (sometimes literal) sanctuary of trust, or conversely, as a progenitor of institutional skepticism, both in religion itself but much more broadly across institutional categories (Casanova 1994; Norenzayan and Shariff 2008). The specific connection between religion, its beliefs and practices, and institutional trust generally remains an unsaturated and complex topic of exploration. The belief systems embedded within religious frameworks can act to either affirm trust in social institutions or introduce skepticism towards them (e.g., Grzymala-Busse 2015; Pargament 1997; Sosis and Ruffle 2003). The importance of understanding these dynamics is of interest to social scientists and practitioners of religion, but also beyond as they impact upon a broad range of core institutions and cultural systems in contemporary societies.

The importance of trust and of the institutional outgrowths of interpersonal and societal trust are perhaps highlighted most forcefully in more diverse societies, where trust across wider interpersonal and group differences require it as a pre-requisite for healthy interaction, both amongst groups and with the institutions with which they interact (Putnam 2007). As one of the most diverse societies on earth along many measures, the United States represents one such society. Among the most highly developed countries in the world, the United States also represents a highly developed national context where religion is comparatively deeply entwined in the contemporary socio-cultural landscape (Huntington 2004). This socio-cultural salience of this religiosity is both visible and comparatively invisible, coming in both public and private forms.

Public religiosity encompasses those visible, often collective expressions of faith, which prominently include attending religious services and participating in both organized and unorganized public displays of religious adherence or belief. Conversely, private religiosity reflects the personal or private beliefs with their associated practices and rituals that mostly happen behind closed doors, and often are much more individualized in their nature. Social scientific studies of religion arguably focus primarily on these public practices of religion, especially religious service attendance (e.g., Lim and Putnam 2010).

While theories of trends and contemporary dynamics of institutional trust and of religious engagement have some prominence in the existing scholarly literature, there is a relative lack of comprehensive empirical studies considering them together. This article seeks to offer a detailed exploration of the relationship between religious engagements and institutional trust and contribute to the broader social science literature on both religion and institutional trust in the process. One element of theoretical focus for this study is anchored in secularization theory, examining its implications beyond religious institutions. Thus, this study introduces some exploration into how religious decline might relate also to trust in secular institutions, a relatively underexplored area in secularization discourse. Specifically, through the research of this paper we seek to contribute to understandings of the relationships between institutional trust, people’s views of their own religiosity and/or spirituality, and the effect of these trust dynamics on public and private religiosity and religious practice in the diverse and pluralistic society that is the contemporary United States.

2. Trust, Religion, and Society

Institutional and interpersonal trust are both understood individually and in combination as important to the fabric of social life in contemporary societies. For instance, Robert Putnam’s (e.g., Putnam 1993, 2000) seminal scholarly contributions on issues of social relations and debates around concepts like social capital highlight the importance and broader societal benefits derived from individuals’ social connections. These are connections largely based on interpersonal and institutional trust. Their benefits largely arise from a sense of trust, community responsibility, and obligation to one’s neighbors, which in turn boosts civic and social engagement in groups and institutions (Putnam 2001). One such prominent group and institution is that of religious congregations and organizations. Such civic, social, and religious affiliations and connections not only strengthen internal ties with other participants but also foster connections in wider societal contexts with unlike people, contributing to a reinforcing cycle of social participation rooted in interpersonal and institutional trust (Lim and Putnam 2010; Putnam and Campbell 2012; Wuthnow 2011).

Subsequent empirical research has validated and expanded on evidence for these notions of social capital and the importance of trust. For instance, Paxton (1999) corroborated Putnam’s assertion of dwindling social trust in the United States during the latter part of the 20th Century, while Abascal and Baldassarri (2015) found that heightened socio-economic disparity in predominantly white neighborhoods in the U.S. correlates with diminished interpersonal trust. The importance of such interpersonal and neighborhood trust has been linked to a variety of broader implications, from people’s voting patterns to crime rates (Messner et al. 2004), which represent concrete examples of broader phenomena such as declining trust in democratic or public safety institutions, like the United States Congress (Jones 2015) or local government and municipal authorities like the police (Silva et al. 2022).

Similarly, interpersonal interactions with individual representatives of institutions can influence and reproduce feelings of trust and the level of that trust. Giddens’ (1990) concept of “face-work” highlights the role that individual representatives play in symbolizing and embodying otherwise more abstract institutions and social systems, with positive interpersonal interactions reinforcing people’s trust in the competence and good motives behind such institutions, and vice-versa. Bachmann (2021) also underscores that this type of trust not only originates from repeated interactions with specific institutions but also from more abstract political, cultural, and otherwise ideological narratives that may not be grounded in direct experiences with those institutions. One such example is that the cultural orientations and collective attitudes often lead individuals to trust their country’s domestic institutions more so than institutions of and in foreign countries (Bachmann 2021). Therefore, while it can be informed by such interactions, trust in institutions extends beyond the actions of individuals or individual representatives of institutions to rather encompass confidence in broader organizations or institutions, and to overarching systems within which those organizations and institutions are embedded (Bachmann 2021). Regardless of the source, it is the aggregate level of trust in society that ultimately allows the creation and functionality of large and complex social institutions (Fukuyama 1995). For this reason, we argue that while trust varies across institutions, understanding aggregate levels of institutional trust, and trust in non-religious institutions along with religious ones, may also have acute or residual effects on trust in religious institutions, and is an important component to understanding the relationship between religiosity and spirituality beyond simply religiously focused institutions.

Secularization, Privatization, and Trust in Religious Institutions

When it comes specifically to the interaction of religion and religious peoples with social institutions, or indeed also the interaction of people with religious institutions, one prominent dynamic worthy of consideration is that of secularization. For clarification, ‘secularism’ in the context of this paper refers to a worldview emphasizing non-religious values and precepts. ‘Secular’ relates to aspects devoid of religious basis, formal organization, and telos. ‘Secularity’ denotes separation from religious or spiritual influences, while ‘secularization’ describes the process of societal movement away from religious beliefs and practices, although the scholarly conceptualization of a secularization theory includes many complexities and debates within it (Gorski and Altınordu 2008), some of which are addressed in this paper.

Many contemporary societies have seen an increase in secularization in recent decades (e.g., Voas and Doebler 2011), including the United States (Voas and Chaves 2016), a trend that, depending on the social circumstances, can also experience an accelerating impulse (Stolz 2020). Regarding the United States context, specifically, it has been observed that secularization has proceeded at a different speed and to a different level than in Europe (Norris and Inglehart 2007), and indeed, recent studies have reaffirmed the relative exceptionalism of European secularization (Díaz de Rada and Gil-Gimeno 2023). Norris and Inglehart (2007) also argue, however, that one should be careful to not conflate the salience of religious language or religious themes in the public or political discourse, such is the case sometimes in the United States, with a growing underlying societal religiosity. In the American case, as in other cases, such as Europe, although there is noticeable and consistent generational reproduction and inter-generational transmission of belief (Gemar 2023b), the lack of full reproduction leads to successive generational cohorts with relatively lower religiosity, thus ultimately empirically confirming secularization processes in the American context along this metric (Voas and Chaves 2016).

But this relative decline in religious engagement includes a complex interplay between belief, practice, institutional trust, and religious authority. A concept of secularization can be understood and observed in a number of ways, such as decreasing attendance at religious services, diminishing influence of religious institutions on individual behavior and public policy, or simply a more generalized decline in the importance placed upon religion in the lives of individuals (Bruce 2011). For instance, religious engagement has not only shifted in terms of extent of religious identification and engagement, but also in terms of the substance and orientation of religious belief and engagement. This includes one prominent shift being a type of ‘privatization’ of religion (Charles 2010, p. 231), or the decoupling of religious belief from identifications with and participation in, more traditional religious institutions (Charles 2010; see also Norris and Inglehart 2004). Therefore, while secularization represents a broader societal shift, the privatization of religion represents a more personal adaptation to these societal shifts, reflecting the movement of religious practices from public to private spheres. In this way, privatization is not a shift towards the personal worldview of secularism.

For instance, in scenarios of privatization, it may be common for individuals to still claim a religious identification, a sign of either strong trust in that religious institution, a signal of cultural membership, or both. For instance, McDowell (2018) finds that Christian youth, while dropping out of church because of disillusionment with the institution of organized religion, nonetheless still believe in God, and may even identify readily as Christian. It has been argued that these types of dynamics may qualify the empirical realities of religious decline in some ways, and rather couches it more as a loss of religious authority exerted by religious institutions (Chaves 1994). However, this can also accompany connected mechanisms of institutional trust, such as interpersonal trust. Indeed, Charles (2010) finds that although the country of Georgia has comparatively low levels of religiosity and religious practice, interpersonal trust is a significant factor in predicting trust in religious institutions.

Kasselstrand et al. (2017) reveal similar dynamics of institutional trust assessed in this paper, finding a clear negative relationship between secularity and confidence in major institutions in the United States. This is an especially important finding within the context of secularization and our empirical focus in this paper on aggregate institutional trust. Kasselstrand et al.’s (2017) findings specifically assert that atheists and nonreligious individuals display lower institutional trust. Our research likewise contributes to this ongoing conversation by further investigating how varying degrees of religiosity and spirituality may be connected or impact upon one’s institutional trust, both in its aggregate and nuanced formation. While diminishing trust in religious institutions among nonreligious, or even spiritual (as opposed to religious), individuals might be expected, our focus extends also to understanding how this relates to trust in non-religious institutions. This is vital for comprehending the broader societal impacts of secularization.

The dynamics of intersecting aspects of secularization, privatization, and the erosion of religious authority can also be viewed through the literature on spirituality in the landscape of religious belief and practice, and the contested category of the ‘spiritual but not religious’ (e.g., Ammerman 2013; Carey 2018; Chaves 2011; Ellingson 2001; Fuller 2001; Saucier and Skrzypinska 2006; Parsons 2018; Popp-Baier 2010; Seto 2021). This concept is often used to capture the erosion of the influence of traditional religious institutions, often accompanied by an erosion of trust in those institutions, by which people still hold to many traditional (and untraditional) views of God and the supernatural broadly. These views often remain an important aspect of individual lives and identities, and, as McDowell (2018) finds, can also be subsumed under a traditional identity, even if trust in the traditional institutions which represent that identity wane.

The dynamics described above can also be observed in the phenomenon of people claiming multiple religious identities (e.g., Gemar 2022, 2023a), or the tendency towards a ‘buffet-style’ or ‘pick-and-mix’ orientation towards religious and spiritual engagement (Kuzma et al. 2009; Watts 2019). While not a new, or as of now large, phenomenon, it is a growing one (Cho 2018). Such classifications would seem to reflect arguments regarding the decline of religious authority, as well as a blurring of the boundaries between distinct religions, and the liminal space between religion and non-religion (Lim et al. 2010). Moreover, echoing a combination of several of the dynamics laid out in this section of the paper, religious identity can often be more indicative of ethno-cultural connection than religious connection (e.g., Voas and Bruce 2004). This may also lead to multiple-religious identification (e.g., Davenport 2016; Lipka 2016). Ultimately, given socio-cultural environments of simultaneous growth in secularization, spirituality, and multiple-religious identification, it may be unsurprising for specific religious or denominational organizations to no longer possess the same kind of religious authority as in the past, or indeed, benefit from the same amount of institutional trust.

3. Research Questions and Hypotheses

We set out to understand how institutional trust, including in religious organizations, impacts upon both public and private religious practice. Our first research questions are necessarily foundational ones for such a study, and ask:

RQ1: What are the substantive groupings and contours of institutional trust in the contemporary United States?

RQ2: What are the substantive groupings of public and private religious practice in the contemporary United States?

We do not propose a hypothesis for these questions beyond presuming that there will be different groups that represent high and low levels of trust and religious practice along with substantively distinct groupings that fall between these.

The second set of research questions relate to how elements of trust and religion interact. To investigate these relationships, we ask the following set of research questions:

RQ3: What is the relationship between institutional trust and the self-claimed religiosity and spirituality of individuals? Do these relationships differ in substance or intensity?

RQ4: What is the relationship between institutional trust and public and private religious practice? Does this relationship differ between these two?

By way of hypotheses, we may first expect that self-claimed spirituality and self-claimed religiosity have different relationships with trust in institutions generally, but especially in religious institutions, with the former having lower trust than the latter. This is because if one has declining trust in religious institutions, they may be less likely to claim religiosity, even as they may still claim spirituality. Therefore, for this same reason we might expect those with higher institutional trust to participate more in public forms of religious activity, because these are most likely to be associated with a religious institution and organized forms of religion. Conversely, then, we hypothesize that those with lower institutional trust are less likely to participate in public religious activities and those with higher institutional trust to claim more religiosity.

4. Materials and Methods

This study draws its data from the General Social Survey (GSS) dataset organized and overseen by the NORC (National Opinion Research Center), housed at the University of Chicago. The GSS is a prominent and well-regarded social survey of the United States population that seeks to capture prevailing opinions, beliefs, and behaviors of the public. The GSS has been consistently administered since 1972, spanning 34 distinct survey cycles, with the three most recent iterations coming from 2018, 2021, and 2022. Because the survey questions often adapt to current societal trends and the shifting research foci of academic and public policy researchers, the research for this paper utilizes the dataset from the 2018 GSS (n = 2348). The 2018 survey cycle was selected because it includes the most comprehensive set of questions and information for answering the research questions of this paper. This is especially the case for many of the religious practices and measures of spirituality investigated here, as 2018 was the only survey year in which some of them were inquired into in the past decade or more. Namely, inquiring into respondents’ visiting of holy sites, reading of scriptures, having religious objects or shrines in their homes, engagement in religious activities, and assessment of their spirituality in relation to religion were only conducted in this survey year. Along with these five measures, we also include in the analysis of this paper more regular GSS questions of respondents’ prayer habits, religious service attendance, assessment of respondents’ religiosity, and respondents’ confidence in various institutions of American society.

To produce a metric that encompasses the general dispositions of institutional trust among respondents, we utilize the following question from the GSS survey:

I am going to name some institutions in this country. As far as the people running these institutions are concerned, would you say you have a great deal of confidence, only some confidence, or hardly any confidence at all in them?

As the question suggests, there are three response categories for this question, and thirteen US institutions were asked about. These thirteen included respondents’ confidence in the military, in major companies, in organized religion, in education, in the executive branch of the federal government, in banks and financial institutions, in the United States Supreme Court, in organized labor, in the US Congress, in medicine, in the press, in the scientific community, and in television. The relative frequencies of respondents’ answers to these questions can be seen in

Table 1. Latent class profile of institutional trust variables.

	Cluster1	Cluster2	Cluster3	Cluster4	Cluster5	Cluster6	Cluster7	Overall
								Relative Freq.
Cluster Size	0.2697	0.1866	0.1510	0.1090	0.1083	0.1031	0.0722
Military
A great deal	0.5166	0.8282	0.4878	0.9600	0.8787	0.3124	0.1093	0.6075
Only some	0.4298	0.1668	0.4500	0.0397	0.1189	0.5450	0.5199	0.3259
Hardly any	0.0536	0.0050	0.0623	0.0002	0.0024	0.1426	0.3709	0.0666
Major companies
A great deal	0.1093	0.2180	0.1147	0.6033	0.3487	0.0295	0.0029	0.1943
Only some	0.7823	0.7342	0.7821	0.3918	0.6293	0.6729	0.3168	0.6692
Hardly any	0.1084	0.0479	0.1032	0.0049	0.0220	0.2976	0.6804	0.1365
Organized religion
A great deal	0.1486	0.2513	0.1551	0.5024	0.3758	0.0140	0.0053	0.2077
Only some	0.6272	0.6194	0.6289	0.4617	0.5548	0.3291	0.2194	0.5399
Hardly any	0.2242	0.1293	0.2160	0.0359	0.0694	0.6569	0.7754	0.2524
Education
A great deal	0.2120	0.3871	0.1438	0.7091	0.1171	0.1069	0.0378	0.2549
Only some	0.6487	0.5568	0.6501	0.2829	0.6386	0.6315	0.4995	0.5782
Hardly any	0.1392	0.0562	0.2062	0.0079	0.2443	0.2616	0.4627	0.1669
Executive branch
A great deal	0.0507	0.0028	0.0009	0.6285	0.3844	0.0003	0.0006	0.1246
Only some	0.7540	0.3270	0.2039	0.3678	0.5993	0.1151	0.1744	0.4246
Hardly any	0.1953	0.6703	0.7952	0.0038	0.0163	0.8847	0.8249	0.4508
Financial institutions
A great deal	0.1407	0.2248	0.1075	0.5451	0.2224	0.0068	0.0019	0.1805
Only some	0.6980	0.6795	0.6876	0.4384	0.6806	0.3148	0.1831	0.5861
Hardly any	0.1613	0.0957	0.2049	0.0164	0.0970	0.6784	0.8150	0.2334
Supreme Court
A great deal	0.1703	0.4901	0.0257	0.8076	0.5432	0.2446	0.0082	0.3140
Only some	0.7299	0.4940	0.5706	0.1910	0.4452	0.6928	0.3923	0.5440
Hardly any	0.0998	0.0159	0.4037	0.0014	0.0116	0.0626	0.5995	0.1420
Organized labor
A great deal	0.1322	0.1441	0.0373	0.4236	0.0332	0.1562	0.0153	0.1352
Only some	0.7245	0.7245	0.6068	0.5506	0.5893	0.7230	0.4679	0.6544
Hardly any	0.1433	0.1314	0.3559	0.0258	0.3775	0.1207	0.5169	0.2104
US Congress
A great deal	0.0403	0.0071	0.0000	0.4501	0.0026	0.0006	0.0000	0.0616
Only some	0.8300	0.6039	0.0080	0.5449	0.4407	0.2469	0.0648	0.4750
Hardly any	0.1297	0.3890	0.9920	0.0050	0.5566	0.7525	0.9351	0.4633
Medicine
A great deal	0.2301	0.7108	0.2044	0.7269	0.3446	0.2544	0.0187	0.3698
Only some	0.6284	0.2801	0.6337	0.2651	0.5760	0.6208	0.3721	0.4996
Hardly any	0.1414	0.0091	0.1619	0.0080	0.0794	0.1248	0.6092	0.1306
Press
A great deal	0.0668	0.2932	0.0299	0.2695	0.0002	0.2659	0.0059	0.1345
Only some	0.4527	0.5469	0.3452	0.5528	0.0365	0.5535	0.1750	0.4102
Hardly any	0.4805	0.1599	0.6249	0.1777	0.9633	0.1806	0.8191	0.4553
Science
A great deal	0.2104	0.8063	0.1466	0.8061	0.4467	0.7741	0.0778	0.4510
Only some	0.7031	0.1920	0.7223	0.1923	0.5302	0.2235	0.6940	0.4861
Hardly any	0.0865	0.0017	0.1311	0.0017	0.0232	0.0024	0.2282	0.0629
Television
A great deal	0.0522	0.1787	0.0292	0.1704	0.0113	0.0563	0.0008	0.0775
Only some	0.5317	0.6434	0.4467	0.6432	0.3152	0.5425	0.0982	0.4982
Hardly any	0.4161	0.1779	0.5241	0.1864	0.6735	0.4011	0.9010	0.4243

. However, for our direct analysis of trust in religious institutions and self-rated spirituality and religiosity, we rather use a different question that includes five response options, ranging from complete confidence to no confidence at all, which was only asked in the 2018 survey round in recent decades.

Our analytical approach utilizes two methods of statistical analysis. First, we deploy latent class analysis (LCA), which is a type of structural equation modelling and a potent technique for discerning distinct clusters within datasets. The elucidation of these distinct groupings is useful primarily in terms of understanding the substantive differences between these groups, but secondarily the number of groups that are most representative of the dataset. Indeed, these two factors are interconnected. Within the design of this study, we use LCA to consolidate the numerous variables of institutional trust, public religious practice, and public religious practice to understand distinct patterns and dispositions towards these metrics for those in the GSS sample.

After our LCA analysis of institutional trust, private religious practice, and public religious practice variables, we go on to conduct multinomial logistic regression analyses. In executing these regression analyses, we position our institutional trust metrics as a primary independent variable. We include our public and private aspects of religiosity as the dependent variables in some of these regressions, and religiosity and spirituality as both dependent and independent variables in others. This is to say that we first see the predictive relationship of institutional trust for respondents’ reported religiosity and spirituality, and then include these variables as controls (along with demographic information) when assessing the predictive effect of institutional trust on public and private religious practice.

5. Results

5.1. LCA Analyses of Institutional Trust

The first independent variable latent class analysis uses the thirteen questions of institutional trust to create groups of substantive patterns among the thirty-nine total response options. Using multiple commonly used and up-to-date measures of model fit (e.g., Bayes Information Criterion (BIC), Consistent Akaike Information Criterion (CAIC)), we chose the seven-class model solution (see also

Table A1. Model fit statistics for latent class analysis of institutional trust measures.

		LL	BIC(LL)	AIC(LL)	AIC3(LL)	CAIC(LL)	SABIC(LL)	Class.Err.
Model13	1-Cluster	−17,017.4591	34,223.3460	34,086.9181	34,112.9181	34,249.3460	34,140.7536	0.0000
Model14	2-Cluster	−16,134.5709	32,559.0307	32,349.1417	32,389.1417	32,599.0307	32,431.9655	0.0865
Model15	3-Cluster	−15,831.1585	32,053.6672	31,770.3171	31,824.3171	32,107.6672	31,882.1291	0.0893
Model16	4-Cluster	−15,664.2301	31,821.2714	31,464.4601	31,532.4601	31,889.2714	31,605.2605	0.1415
Model17	5-Cluster	−15,572.2968	31,738.8660	31,308.5936	31,390.5936	31,820.8660	31,478.3823	0.1661
Model18	6-Cluster	−15,472.7034	31,641.1404	31,137.4069	31,233.4069	31,737.1404	31,336.1839	0.1793
Model19	7-Cluster	−15,389.6335	31,576.4617	30,999.2670	31,109.2670	31,686.4617	31,227.0324	0.1809
Model20	8-Cluster	−15,339.4793	31,577.6144	30,926.9586	31,050.9586	31,701.6144	31,183.7122	0.2032

). We therefore take a statistically based, primarily inductive approach to assessing institutional trust in this specific dataset, in line with widely considered best practice (Bonikowski and DiMaggio 2022), even as we recognize that critiques of inductive approaches to LCA exist (e.g., Eger and Hjerm 2022). Using this statistically based perspective, strongest fit arguments can be made for the seven latent class model in the first instance, and the eight latent class model in the second instance. Because this is the case, even as these two have little between them, the seven latent class model has the advantage of parsimony over the model producing eight latent classes. Although we recognize that a purely parsimonious or more theoretically deductive approach could argue for other latent classes, we opt here for what we consider as the most statistically justifiable approach for this specific dataset. It is also the case that we are most concerned in this paper with identifying the relationship to different dispositions of institutional trust by those of different religious and spiritual beliefs and practices, rather than endeavoring to establish a definitive typology of institutional trust more broadly in the United States. However, it has also been argued that there is an increased need to capture the continually nuanced landscape of trust within a diversely secular society Kasselstrand et al. (2017).

The seven latent class groups comprised approximately 27%, 19%, 15%, 11%, 11%, 10%, and 7% of the sample, respectively. The latent class profile of this seven-class model solution appears in Table 1. The first and largest of these latent class groups is characterized by moderate levels of institutional trust in all thirteen of the institutions, although also some lower levels of trust in the press, science, and medicine. For these reasons, this group is the ‘moderate trust’ group.

The second latent class has more of a range in trust levels across institutions. This group has a high level of trust in the military, education, supreme court, medicine and science, while also having moderate to high levels of trust in the press, television, organized labor, financial institutions, major companies, and organized religion. Those in this group have the lowest amount of trust in the federal government and Congress, with low amounts of trust in the former and moderate trust in the latter. Because this group has lower levels of trust in institutions whose heads are elected officials, but moderate to high trust in all others, this group is the ‘elected government skeptical’ group. This moniker reflects the complexity of people, and for many of these groups, the complexity of their substantive make-up necessitates relatively clunky monikers.

The third latent class group has moderate levels of trust in the military, major companies, organized religion, financial institutions, and science, while moderate to low trust in education, the Supreme Court, and medicine. This group has low trust in the press, television, organized labor, congress, and the federal government. While recognizing that this group has even lower trust in institutions of government with elected heads, we emphasize that this group also has low levels of trust in organized labor (opposed to business and finance), and in the media. Therefore, we label this group the ‘labor and media skeptical’ group.

The substantive make-up of the fourth latent class group, however, is more consistent. This group has the highest aggregate institutional trust, with high levels of trust in all thirteen institutions, although also some moderate trust in Congress, television, and the press. For these reasons, this group is the ‘high trust’ group.

The fifth latent class group again shows a varied trust profile. This group has high trust in the military, in major companies, and in the Supreme Court. They also have moderate to high trust in organized religion, the federal government, and financial institutions. They also have moderate levels of trust in science and moderate to low levels of trust in education. For these reasons, we label this group the ‘science and education skeptical’ group.

With the sixth latent class group we see high levels of trust in science and moderate to high levels of trust in the press and in organized labor. This group has moderate trust in the Supreme Court, in medicine, and in television, while also moderate to low amounts of trust in the military, major companies, and education. This group has low levels of trust in financial institutions, Congress, the federal government, and organized religion. Again, at the risk of a clunky moniker, we decided to label this group as the ‘business and religion skeptical’ group, even as it is again emphasized here that elected institutions of government have even lower amounts of trust than the ‘elected government skeptical’ group. We have done so in part because this group is the opposite of group three in terms of their trust in labor versus business, finance, and organized religion, even as both share the lowest amounts of trust in Congress and the federal government.

Lastly, the seventh latent class group is on the other end of the institutional trust spectrum from latent class four. This group has low levels of trust in all thirteen institutions, while also having some moderate levels of trust in science and in the military. Because of this, we label this group as the ‘low trust’ group. It is notable that this is the smallest of the groups.

5.2. LCA Analysis of Public and Private Religious Practice

For the third latent class analysis, we now move onto formulating variables of public and private religious practice that will serve as the dependent variables for subsequent regression analyses. This LCA includes three variables, with multiple response options, which correspond to the variables from the GSS encompassing public religious practice. From model fit statistics for this LCA, we chose the four-class model as the one of best fit. The variables and latent class profile from this LCA can be seen in

Table 2. Latent class profile of public religious practice variables.

	Cluster1	Cluster2	Cluster3	Cluster4	Overall
Cluster Size	0.3863	0.2595	0.1795	0.1747
Visit holy sites
Never	0.4846	0.9865	0.0132	0.9152	0.6054
Less than once a year	0.1829	0.0130	0.0239	0.0680	0.0902
About once or twice a year	0.2114	0.0005	0.1326	0.0155	0.1083
Several times a year	0.0805	0.0000	0.2424	0.0012	0.0748
About once a month or more	0.0407	0.0000	0.5879	0.0001	0.1212
Attend religious services
Every week or more	0.0355	0.0000	0.5820	0.7151	0.2431
Almost every week	0.0882	0.0001	0.2955	0.2321	0.1277
Once a month	0.1191	0.0013	0.0816	0.0409	0.0681
Several times a year	0.2163	0.0165	0.0303	0.0097	0.0950
Once a year or less	0.3262	0.1746	0.0093	0.0019	0.1733
Never	0.2147	0.8074	0.0013	0.0002	0.2927
Religious activities
Never	0.4838	0.9763	0.0581	0.0692	0.4627
Less than once a year	0.0838	0.0187	0.0222	0.0253	0.0456
About 1–2 times a year	0.1904	0.0047	0.1115	0.1216	0.1160
Several times a year	0.1343	0.0004	0.1736	0.1812	0.1148
About once a month	0.0520	0.0000	0.1486	0.1484	0.0727
2–3 times per month/several x/yr.	0.0381	0.0000	0.2407	0.2300	0.0981
Once a week or more	0.0176	0.0000	0.2453	0.2243	0.0900

, which shows that the four latent class groups represent 39%, 26%, 18%, and 17% of the sample, respectively.

The first latent class group shows moderate levels of religious service attendance, of visitations to holy sites, and infrequent engagement in religious activities. For these reasons, we label this group the ‘moderate public (religious) practice’ group. The second latent class group shows virtually no public religious practice across the three measures, while the third group shows elevated religious practice across all three measures. Therefore, we label these two groups the ‘no public practice’ and ‘high public practice’ groups, respectively. Finally, the fourth latent class group shows even higher religious service attendance than the high public practice group, as well as similar levels of engagement in religious activities. However, this last group shows low levels of visiting holy sites, like those of the no public practice group. Because of this combination, and at the risk of another clunky moniker, we have chosen to characterize this group as the ‘non-holy site public practice’ group.

The final latent class analysis utilized three measures of more private religious practices to characterize groups of such practice amongst the respondents. Using the model fit statistics, for this LCA we chose the model corresponding to a three latent class group solution (see

Table A4. Relative frequencies of non-LCA produced variables included in the regression analyses of this paper (other than institutional and interpersonal trust metrics).

	Relative Frequency
	(GSS)
Religiosity
Very religious	15.5%
Moderately religious	37.7%
Slightly religious	25.2%
Not religious	21.6%
Spirituality
Follow a religion and are also spiritual	45.1%
Follow a religion but are not spiritual	21.0%
Do not follow a religion but are spiritual	21.1%
Do not follow a religion and are not spiritual	12.8%
Age
18–24	11.8%
25–34	19.0%
35–49	25.8%
50–64	24.9%
65+	18.3%
Sex
Male	45.5%
Female	54.5%
Race
Other	12.8%
Black	14.8%
White	72.4%

). These groups represented approximately 43%, 31%, and 25% of the sample, respectively, and the corresponding latent class profile is displayed in

Table 3. Latent class profile of private religious practice variables.

	Cluster1	Cluster2	Cluster3	Overall
Cluster Size	0.4325	0.3130	0.2545
Pray Frequency
Never	0.0002	0.4052	0.0729	0.1455
Twice/yr. or less	0.0003	0.0834	0.0259	0.0328
Several/yr. to 1/month	0.0023	0.1157	0.0620	0.0530
Almost every week	0.0102	0.0875	0.0808	0.0524
Every week	0.0531	0.0774	0.1235	0.0786
Several times/week	0.9339	0.2308	0.6350	0.6377
Religious object
Yes	0.4274	0.1020	0.7594	0.4100
No	0.5726	0.8980	0.2406	0.5900
Religious scripture
Yes	0.9674	0.2314	0.3293	0.5746
No	0.0326	0.7686	0.6707	0.4254

. The first of these three latent class groups, the largest such group, is characterized by increased levels of private religious practice. This group prays relatively frequently, reads scriptures, and has religious objects in the home. We therefore characterize this group as the ‘high private (religious) practice’ group. Conversely, the second latent class group is the most unlikely of the groups to pray frequently, to read scripture, or to have religious objects in the home. This second group is therefore characterized as the ‘low private practice’ group. Lastly, the third latent class group is by a large margin the most likely to have religious objects in the home, although they are unlikely to pray frequently or read scriptures. For these reasons, we have chosen to characterize this final group as the ’object focused private practice’ group.

5.3. Regression Analyses of Religiosity, Spirituality, and Public and Private Religious Practice Using Institutional Trust Measures

We proceed with simple regression models to understand the direct relationship between first institutional trust in ‘churches and religious organizations’ and then institutional trust profiles and respondents’ religiosity and spirituality. We use both these variables because one variable shows the relationship between people’s understanding of their religiosity versus their spirituality (or not), while the other shows the level of intensity to religiosity. These can both be viewed in

Table 4. Log odds ratios from regression analyses of trust in religious institutions and religiosity, along with religious institutional trust and spirituality.

Spirituality a	Religion + Spiritual	Religion, Not Spiritual	No Religion but Spiritual
Confidence in churches and religious orgs.
Complete confidence	5.326 ***	5.359 ***	1.910
A great deal of confidence	4.150 ***	4.498 ***	0.787
Some confidence	2.671 ***	3.315 ***	0.686 *
Very little confidence	1.420 **	2.617 ***	0.818 *
No confidence at all	-	-	-
Nagelkerke R-squared	(0.221)
Religiosity b	Very Religious	Moderately Religious	Slightly Religious
Confidence in churches and religious orgs.
Complete confidence	2.962 ***	4.449 ***	2.424 ***
A great deal of confidence	2.913 ***	4.835 ***	2.806 ***
Some confidence	1.378 ***	3.793 ***	2.520 ***
Very little confidence	0.303	2.389 ***	1.807 ***
No confidence at all	-	-	-
Nagelkerke R-squared	(0.206)

^a Reference group = no religion, not spiritual; ^b reference group = not religious; * p < 0.05; ** p < 0.01; *** p < 0.001.

and

Table 5. Log odds ratios from regression analyses of institutional trust and religiosity, along with institutional trust and spirituality.

Spirituality a	Religion + Spiritual	Religion Not Spiritual	No Religion but Spiritual
Institutional trust
Moderate trust	−0.583	−0.644	−0.807
Elected government skeptical	−0.593	0.077	−0.688
Labor and media skeptical	−0.792	−0.664	−0.767
High trust	−0.353	0.154	−0.236
Science and education skeptical	0.026	0.511	−0.999
Business and religion skeptical	−1.900 *	−1.723	−0.906
Low trust	-	-	-
Nagelkerke R-squared	(0.065)
Religiosity b	Very Religious	Moderately Religious	Slightly Religious
Institutional trust
Moderate trust	0.416	1.270 ***	0.890 **
Elected government skeptical	0.025	0.597	0.405
Labor and media skeptical	0.298	1.176 ***	0.622
High trust	0.822 *	1.637 ***	0.556
Science and education skeptical	1.680 ***	1.943 ***	1.119 **
Business and religion skeptical	−22.289	−0.800 *	−0.494
Low trust	-	-	-
Nagelkerke R-squared	(0.130)

^a Reference group = no religion, not spiritual; ^b reference group = not religious; * p < 0.05; ** p < 0.01; *** p < 0.001.

.

First, as Table 4 shows, there is a strong positive relationship between trust in religious organizations and churches and those who report following a religion, whether they also consider themselves religious or not. However, there is little to no positive relationship between trust in religious institutions and spirituality without also following a religion (i.e., spiritual but not religious), as this group shows very little to some confidence in religious institutions. This is approximately confirmed by the regression of trust in religious institutions and self-rated religiosity. There is a strong positive relationship between confidence in religious institutions and religiosity, although the strongest positive relationships are with the moderately religious rather than the very religious. However, this relationship is indeed stronger with the very religious than with the slightly religious, although perhaps not by as great a degree as might be assumed.

Regarding the variable of self-rated spirituality, Table 5 shows little in the way of statistically significant relationships between broader institutional trust and respondents’ religiosity. Indeed, there is only one, with those in the business and organized religion skeptical group less likely than those with the lowest aggregate institutional trust group to say that they both follow a religion and are spiritual, rather than not being either religious or spiritual. Conversely, the different levels of religiosity show many strong statistically significant relationships with institutional trust profiles. Compared to those who are not religious, those with high institutional trust, and those in the science and education skeptical group, are much more likely to say that they are very religious than those with low institutional trust. Besides the business and organized religion skeptical group, which shows the inverse relationship, the other institutional trust groups, and especially again the high institutional trust and science and education skeptical groups, are much more likely to say they are moderately religious than non-religious. Finally, both the group of moderate institutional trust, along with the science and education skeptical group, are also more likely than those with low institutional trust to say that they are slightly religious than not religious. Therefore, we generally see strong positive predictive relationships between higher institutional trust and self-reported religiosity, while we do not see this kind of relationship when it comes also to self-reported spirituality.

The third regression model that we create assesses the predictive power of our institutional trust metric on public religious practice. With public religious practice, therefore, as the dependent variable, and institutional trust as the primary independent variable, we also include other variables as controls. These include respondents’ assessments of their religiosity and spirituality, which in theory are also primary drivers of religious practice generally, along with simple demographic measures of age, sex, and race (see also Table 5), because these too may predict differential religious practice. The results of this first regression analysis are displayed in

Table 6. Log odds ratios from regression analysis of public religious practice using the GSS survey, and including variables of institutional trust, religiosity, spirituality, and basic demographic information ^a.

	Moderate Public Practice	High Public Practice	Non-Holy Site Public Practice
Institutional trust
Moderate trust	−0.738	0.960	−1.210
Elected government skeptical	0.050	0.938	0.564
Labor and media skeptical	0.420	1.126	0.875
High trust	−0.189	0.125	−0.618
Science and education skeptical	0.853	2.920 **	0.891
Business and religion skeptical	0.543	3.433 **	0.669
Low trust	---	---	---
Religiosity
Very religious	0.868	3.724 **	4.721 ***
Moderately religious	0.881	1.962 *	2.127 *
Slightly religious	1.339 **	1.127	0.483
Not religious	---	---	---
Spirituality
Religion + spiritual	2.198 ***	21.227 ***	21.055
Religion but not spiritual	2.009 **	20.842 ***	20.902
Spiritual but not religion	1.249 *	19.012	18.496
Neither religion nor spiritual	---	---	---
Age
18–24	−0.745	1.284	−1.311
25–34	0.018	1.128	−1.610
35–49	1.290 *	0.975	0.1550
50–64	−0.155	0.482	−1.130
65+---	---	---
Sex
Male	0.284	0.074	−0.994
Female	---	---	---
Race
Other	1.005	0.391	1.063
Black	0.105	1.187 *	1.122
White	---	---	---
Nagelkerke R-squared	0.608

^a Reference group = latent class 2, ‘no public (religious) practice’; * p < 0.05; ** p < 0.01; *** p < 0.001.

.

As the results of this table show, relative to the ‘no public practice’ group, institutional trust is not a predictive variable for the latent class group of moderate religious practice. Rather, elevated levels of religiosity and spirituality primarily predict this profile of public religious practice. Some institutional trust profiles are, however, prominent predictors of high levels of public religious practice. Specifically, those with science and education skeptical trust profile, along with those with business and organized religion skeptical profiles, are more likely to have high public religious practice than those with the lowest institutional trust. This group of high public practice is, perhaps less surprisingly, strongly predicted by high levels of reported religiosity and spirituality among respondents. Black Americans are also more likely to be in this group of high public religious practice than white Americans. Finally, compared to those in the no public practice group, those in the non-holy site public practice group are not predicted by any of the institutional trust measures and are rather predicted almost exclusively by self-rated religiosity, with those who are more religious being many times more likely to be in this group than those who are not religious.

With the final regression model of our analysis, we observe the predictive power of our institutional trust metric and associated control variables on the private religious practice groups found in this paper (see

Table 7. Log odds ratios from regression analysis of social engagement in the US using the GSS survey data (n = 2348), including socio-economic variables and religion-based variables of religious preference, strength of affiliation, and service attendance ^a.

	High Private	Non-Object
	Religious Practice	Private Practice
Institutional trust
Moderate trust	−0.734	−1.068
Elected government skeptical	−0.582	−1.940 *
Labor and media skeptical	−1.277	−1.255
High trust	−2.441 **	−1.862 *
Science and education skeptical	1.823	−1.419
Business and religion skeptical	−0.659	−1.233
Low trust	---	---
Religiosity
Very religious	4.822 ***	2.939 *
Moderately religious	2.350 ***	1.282 *
Slightly religious	0.526	0.382
Not religious	---	---
Spirituality
Religion + spiritual	4.302 ***	1.485 *
Religion but not spiritual	3.356 **	1.153
Spiritual but not religion	2.988 *	1.413 *
Neither religion nor spiritual	---	---
Age
18–24	−0.096	−0.801
25–34	0.327	−1.201
35–49	1.366 *	−0.350
50–64	0.521	−0.383
65+	---	---
Sex
Male	−0.513	0.087
Female	---	---
Race
Other	0.143	0.777
Black	0.196	−1.732 *
White	---	---
Nagelkerke R-squared	0.573

^a Reference group = latent class two, ‘low private (religious) practice’; * p < 0.05; ** p < 0.01; *** p < 0.001.

). As the results show, there appears to be an inverse relationship between institutional trust and private religious practice, at least at either end. This is to say that those with the highest aggregate institutional trust are much less likely to have high private religious practice compared to those with the lowest aggregate institutional trust. Both religiosity and spirituality are highly predictive, although like for public religious practice, the predictiveness of each decreasing level of religiosity drops much more sharply than for spirituality, suggesting that across both measures of public and private religious practice, spirituality can in some ways compensate for feelings of religiosity when it comes to practices otherwise associated with religions. The middle age group (35–49) is also more likely to have high private practice compared to the oldest age group, to statistically significant levels, something that we also saw with the moderate group of public religious practice. Finally, there is again an inverse relationship between institutional trust and non-object private religious practice, for both the high institutional trust group and the elected government skeptical group compared to those with low institutional trust across the board. Indeed, including non-statistically significant results, results suggest that those with the lowest institutional trust are more likely to be in this group than any group with higher institutional trust, compared to those with no private religious practice. This group is still predicted by higher levels of religiosity and spirituality, although the relationships are not as strong as for other groups. Finally, this group is significantly less likely to include Black Americans than it is to include White Americans.

6. Discussion

We set out in this study to investigate questions regarding the relational dynamics of institutional trust, religiosity, spirituality, and religious practices in contemporary American society. Through the analysis of this paper, we sought to answer four research questions guided by theories of the relationships between institutional trust and religion in contemporary societies. Largely confirming our hypotheses, the results show notable relationships between broader institutional trust and religious practice, and strong relationships between trust in religious institutions and the prevalence of religiosity and spirituality. Our emphasis on aggregate institutional trust, as opposed to investigating specific institutions individually, is to provide a more comprehensive view of the societal trust landscape, reflecting the varied and nuanced relationships of religiosity and spirituality to trust in the dataset for this paper.

In answer to our first research question regarding the groupings and contours of institutional trust in the US, we found a varied trust profile across the survey sample, with a group of moderate trust as the largest. Another notable group exhibited distrust towards elected government bodies, especially Congress, but otherwise has higher levels of trust in other institutions. Additionally, distrust in the elected branches of government (especially Congress) persisted across multiple groups, including those with additional skepticism in organized labor, the media, science, education, business, and organized religion. There was finally a group of high trust across institutions and low trust across institutions, with the latter group as the smallest group. Therefore, while these patterns generally align with Putnam’s (e.g., Putnam 1993, 2000) arguments on the contemporary erosion of trust in American society, these results may ultimately be more accurately described to portray a lack of trust in the most democratic institutions of contemporary American society rather than a wholesale mistrust of institutions. However, this distrust may indeed have broader societal implications, such as decreased levels of voting or increased levels of crime (Jones 2015; Messner et al. 2004).

In answer to our second research question, we find distinct substantive categories of both public and private religious practice, with four categories of public religious practice and three for private religious practice. The size of these groups supports the idea that public religious engagement may be less than the rates of private religious practice in contemporary American society. The empirical finding of less public engagement would support assertions of secularization along these more public facing metrics. However, when considered alongside private practice and rates of spirituality, these findings support theories and empirical findings asserting a type of privatization of religious engagement (Charles 2010), whether through a loss of religious authority (Chaves 1994) or disillusionment with organized religion (McDowell 2018). Therefore, it is privatization of religion, rather than an obvious secularization, that is most supported by the data of this paper. This is especially true given the cross-sectional nature of this data. However, while privatization is different from secularization, it may arguably offer scholars a phenomenon more connected to issues of institutional trust because it may better control for this dynamic given a continuation of belief. This might be particularly the case in the context of the United States, especially compared to (Western) Europe, because of relatively lower levels of institutional trust (Hitlin and Shutava 2023) and relatively higher levels of religious and spiritual belief (Berger et al. 2008; Norris and Inglehart 2007). While this is particularly true for religious institutions, a more widely studied dynamic, the results of this paper provide initial findings for understanding how connected these trust dynamics are to non-religious institutions and if they are generally connected in contemporary American society.

In answer to our third research question regarding the relationship of institutional trust to religiosity and spirituality, we observed first that individuals who trust in religious institutions are much more likely to report themselves as religious, and progressively so as trust increased. However, when focusing on elements of spirituality in comparison to religion, the predictiveness of trust wanes. This generally supports our hypothesis, suggesting that spirituality, being more personal, and generally less ties to religious structures, could potentially be associated with diminished trust in religious institutions. The nature of spirituality, often being less institutionally bound (Fuller 2001), seems to indeed slightly decouple it from the influence of institutional trust, or suggest support for a concept of ‘spiritual but not religious’ with relation to institutional trust. While more general institutional trust shows little significant influence on self-rated spirituality, it shows varying degrees of significant relationships to religiosity. Notably, those with high institutional trust or belonging to the ‘science and education skeptical’ group are more likely to claim religiosity than those with low trust. Therefore, the relationships do differ in substance, with religiosity being more deeply connected to trust in religious institutions than spirituality. Our findings, showing varied relationships between institutional trust and religiosity, suggest complexities of institutional trust extending beyond individual actions to encompass broader organizational systems and cultural narratives (Bachmann 2021).

Finally, to understand the relationship between institutional trust and public and private religious practice and if those relationships differ, we found first that some groups with elevated levels of institutional trust are more engaged in public religious practice, although this was not true for the highest institutional trust group. However, because broader institutional trust was generally negatively associated with increased private religious practice, this would suggest confirmation of the hypothesis that private religious practices would be more associated with those of lower institutional trust levels, not only lower trust in religious institutions. Both of these findings may therefore suggest a type of secularization among those with generally high institutional trust, which would mirror the high institutional trust yet high secularization dynamics of Western and Northern European countries. Declining trust in institutions broadly could motivate individuals inclined to distance themselves from public religious institutions by turning towards more unique individual practices, reflecting privatization (Charles 2010), perhaps especially because they do not trust other institutions to provide answers to what they seek.

7. Conclusions

In conclusion, this study investigated the dynamics of institutional trust, religiosity, and spirituality in the contemporary American context of secularization and privatization of belief. Key findings highlight an outsized distrust in democratic institutions, particularly Congress, while religious participation shows more strongly in private than public settings. Trust in religious institutions directly relates to professed religiosity, while professed spirituality is less connected to this trust. These results suggest evidence of both privatization and a type of spirituality without religion, and this paper concludes that there is a meaningful but varied connection between institutional trust and religiosity, spirituality, and religious practice in the United States today. The privatization of religion, as our findings suggest, may lead to a nuanced understanding of institutional trust, where personal spiritual beliefs become decoupled from traditional institutional reliance.