Should I Stay or Should I Go? Religious (Dis)Affiliation and Depressive Symptomatology

Abstract

Religious affiliation is generally associated with better mental health. The nonreligious, however, currently constitute one of the fastest-growing religious categories in the United States. Since most of the nonreligious were raised in religious homes, their growth raises important questions about the mental health of those who consider dropping out of religion. In this article, I use longitudinal data from the Portraits of American Life Study to examine the impact of religious affiliation on mental health. Specifically, I compare individuals who dropped out of religion (leavers) with individuals who considered dropping out (stayers) and individuals who are more consistent in their religious (stable affiliates) and nonreligious (stable Nones) affiliations. I find that stayers experience more depressive symptoms than any other group and that they experience a greater increase in depressive symptoms over time. My findings are consistent with identity theories in sociology, and they provide evidence that a strong religious or secular identity is an important contributor to mental health.

Keywords

religion depression Portraits of American Life Study (PALS)identity theory

Nones—atheists, agnostics, and the unaffiliated—occupy one of the fastest growing “religious” groups in the United States today. According to recent reports, nearly one in four Americans is not affiliated with a religious institution (Pew Forum on Religion and Public Life 2015). Interestingly, though, nearly three quarters of these women and men grew up in religious homes (Kosmin and Keysar 2008). This means that the vast majority of Nones are religious dropouts who actively left religion behind. Are these women and men making the best decision for their mental well-being? What about those who consider dropping out but decide to stay?

The extraordinary rate of growth among the Nones in the United States raises important questions about the health and well-being of those who choose to disaffiliate from religious institutions. Active involvement in a religious community is generally associated with better health outcomes (Koenig 2012), but these benefits are not universal (Ellison 1995). A number of circumstances limit, or even reverse, the positive relationship between religious participation and mental health. On the one hand, disaffiliation is associated with more negative health outcomes (Fenelon and Danielsen 2016; Hayward et al. 2016). On the other hand, security in a particular nomos—whether the canopy is sacred or secular—appears to improve mental and physical health (Mochon, Norton, and Ariely 2011; Schnittker 2001). Thus, the decision to drop out of organized religion may play an important role in the relationship between religion and mental health. Moreover, since more people will consider dropping out of religion than will actually leave, it is important to consider the differences between those who depart and those who stay.

In this paper, I use longitudinal data from the Portraits of American Life Study (PALS) to examine the relationship between religious affiliation and mental health. Religious affiliation varies from unwavering devotion to a religious group (or no group at all) to the decision to completely abandon organized religion. Therefore, I compare individuals who considered dropping out of religion but remain affiliated with a religious group (stayers) to individuals who never considered dropping out of religion (stable affiliates), individuals who were never affiliated with a religious group (stable Nones), and individuals who dropped out of religion after being affiliated with a religious group (leavers). I find that stayers report more depressive symptoms than any other group and that they experience a greater increase in depressive symptoms over time. My findings are consistent with other research on identity and mental health (e.g., Cast and Burke 2002; Large and Marcussen 2000), and they provide evidence that a strong religious or secular identity is an important contributor to mental health.

Religious Affiliation

The number of Nones has grown from less than 10 percent in the early 1990s to nearly a quarter of the U.S. population (Pew Forum on Religion and Public Life 2015). Not surprisingly, this trend has sparked many attempts to understand who these Nones are and what they believe (e.g., Baker and Whitehead 2016; Manning 2013; Taylor and McPeek 2013). One Pew Forum study, for example, suggests that Nones are disproportionately concentrated among young adults (Pew Forum on Religion and Public Life 2010), and another indicates that a sizable majority—68 percent—still believe in God (Pew Forum on Religion and Public Life 2012). Today, Nones are the second largest “religious” group in the United States (22.8 percent) behind Evangelical Protestants (25.4 percent) and just ahead of Catholics (20.8 percent) (Pew Forum on Religion and Public Life 2015).

The growing number of religiously unaffiliated Americans raises important questions about the process of leaving organized religion behind. Nationally representative data from the Portraits of American Life Survey (Emerson and Sikkink 2012) indicate that 13 percent of religiously affiliated Americans considered dropping out of religion altogether between 2003 and 2006. Individuals who consider dropping out of religion altogether are generally motivated by religious skepticism, political attitudes, and the experience of stressful life events (Vargas 2012).

According to Vargas (2012), religious skepticism or a recent financial crisis increases the odds that a person will consider opting out of religion. Certain political attitudes tend to have opposite effects on religious commitment, though. More specifically, support for Christian political groups reduces the odds that women and men will consider dropping out of religion altogether, while “support for same-sex marriage is positively associated with considering and actually leaving religion” (Vargas 2012:214). Men and whites are also more likely to consider leaving religion altogether, but according to Vargas (2012) only 38 percent of the people who consider dropping out of religion altogether actually leave. This means that between 2003 and 2006, 8 percent of religiously affiliated Americans—or more than 6 percent of the U.S. population—considered opting out of religion but remained affiliated with an organized religious group.

Though disparate in their religious affiliations, “stayers” constitute a group of people that is larger than atheists (3.1 percent), agnostics (4.0 percent), and all of the major non-Christian religious traditions in the United States (5.9 percent total; includes Jews, Hindus, Muslims, and Buddhists) (Pew Forum on Religion and Public Life 2015). With the exception of Vargas (2012), whose comparison of “stayers,” “leavers,” and “stable affiliates” highlights the mechanisms that push leavers out of religion and pull stayers back into the pews, though, stayers are generally ignored amid a growing body of empirical research focused on the growing population of Nones. To date, no empirical research directly examines the mental health of stayers compared to stable affiliates and those who opt out of religion altogether. Identity theories in sociology and research on the relationship between organized religion and mental health suggest there may be important differences between these groups, though.

Identity Theory and Mental Health

Identity theory provides a useful framework for understanding the complex relationship between religious affiliation and mental health. According to identity theory, the self is the composite of multiple identities reflecting the many roles that an individual occupies in the society (Stryker 1980). Each identity is a control system composed of four components: an input, an identity standard, a comparator, and an output. Together, these components function to regulate the meanings perceived by the identity in the way a thermostat functions to regulate the temperature it perceives (Burke and Stets 2009). More specifically, people alter their behaviors (the outputs) in order to achieve semantic congruence between their self-perceptions of meaningful feedback from others (the inputs) and the meanings attached to the identity (the identity standard). When the comparator detects a disturbance between the input and the identity standard (i.e., an incongruence), distress is experienced in the form of negative emotions such as depression or anxiety (Burke 1991, 1996). The experience of negative emotions is more pronounced when core (i.e., salient) identities are threatened (Stryker 2002).

For many, religion is a core identity, “affecting how they understand themselves as religious or spiritual beings … as well as determining their social identification with a particular religious group” (Park and Edmondson 2012:150). Not surprisingly, the majority of empirical research on religion and mental health shows that people who are more religious experience better mental health than people who are less religious (Koenig 2012). Some notable examples reveal a positive relationship between religious engagement and subjective well-being (Ellison 1991; Ellison, Gay, and Glass 1989; Lim and Putnam 2010) and a negative relationship between religious engagement and various forms of psychological distress (Acevedo, Ellison, and Xu 2014; Ellison et al. 2012; Eliassen, Taylor, and Lloyd 2005). Acevedo et al. (2014), for example, find that religious involvement buffers the deleterious effects of perceived financial strain and neighborhood disadvantage on psychological distress. Similarly, Ellison et al. (2012) find that a secure attachment to God buffers against the pernicious effects of stressful life events on distress. It is increasingly clear, though, that religion does not improve the mental health of all participants.

Several conditions limit, or even reverse, the positive relationship between religion and mental health. According to Ellison and Lee (2010), there is a “dark side” to religion (see also Krause and Wulff 2004). More specifically, disturbances such as doubt (Ellison and Lee 2010; Krause 2006; Krause and Ellison 2009), disaffiliation (Fenelon and Danielsen 2016; Hayward et al. 2016), moderate levels of commitment (Eliassen et al. 2005), and negative interactions in the Church (Ellison et al. 2009) tend to diminish the otherwise positive relationship between religion and mental health. According to Speed and Fowler’s (2017) recent study of Ontario’s adult population, the relationship between religion and mental health is also moderated by religious affiliation. More specifically, there is a positive relationship between church attendance and well-being for Christians and a negative relationship between church attendance and well-being for the unaffiliated (Speed and Fowler 2017). The relationship between church attendance and mental health also varies within Christianity (Schwadel and Falci 2012) . Comparing Nebraska’s mainline Protestant, evangelical Protestant, and Catholic populations, Schwadel and Falci (2012) find that only mainline Protestants report fewer depressive symptoms when they attend church more often.

A couple of recent studies make use of identity theory to explain the contingent relationship between religion and mental health. First, Ellison et al. (2013) used identity theory to explain why the pernicious association between doubts about God and psychiatric symptoms is most pronounced among people who expressed strong religious commitments. Second, Galek and her colleagues used identity theory to explain why the belief that life lacks meaning and purpose caused more psychological distress among the highly religious than among the less religious or the highly religious who did not believe that life lacks meaning or purpose (Galek et al. 2015). Both of these studies echo prior research on the conditional relationship between core religious identities and mental health (e.g., Krause 1994, 1998; Krause and Wulff 2004), and all of these studies are consistent with Stryker’s (2002) argument that negative emotions are more pronounced when more salient roles are threatened.

My hypotheses are based on the propositions of Burke and Stets’s (2009) identity theory and the growing body of research on religion and mental health described above.¹ Generally speaking, disturbances produce negative emotions like anxiety and depression (Burke 1991, 1996). More specifically, religious doubts are the single largest predictor of who considers dropping out of religion (Vargas 2012), and they tend to increase psychological distress (Krause 1994, 1998; Krause and Wulff 2004). Thus, I hypothesize that people who consider dropping out of religion altogether will report more negative health outcomes than those who do not.

Hypothesis 1: Individuals who consider dropping out of religion altogether will experience more depressive symptoms than individuals who do not consider dropping out of religion altogether.

Hypothesis 2: Individuals who consider dropping out of religion altogether will experience more depressive symptoms than individuals who were never affiliated with a religious institution.

Considering dropping out of religion altogether is only the first step in the disaffiliation process. In order to disaffiliate, a person must actually leave. Research on religion and mental health suggests that completing the disaffiliation process can be better for the mental health of people who consider dropping out of religion altogether. Schnittker (2001), for example, describes an inverted U-shaped relationship between religious salience and depression. In other words, the negative relationship between religion and depression is strongest when religion is very important and when religion is not important at all. Similarly, Ross (1990) finds that stronger religious beliefs are associated with lower levels of distress than weaker religious beliefs, while no religious beliefs are also associated with low levels of distress (see also Eliassen, Taylor, and Lloyd 2005).

The research described above is consistent with Ebaugh’s (1988) role exit theory and the self-verification process (Cast and Burke 2002). Role exit is “the process of disengagement from a role that is central to one’s self-identity and the reestablishment of an identity in a new role that takes into account one’s ex-role” (Ebaugh 1988:1). Therefore, disaffiliation is the process of creating an ex-role that incorporates the old identity (i.e., affiliated) and the new identity (i.e., unaffiliated). By establishing an ex-role, disaffiliates can reduce the negative emotions caused by doubt and other disturbances. This process can be understood using Cast and Burke’s (2002) extension of identity theory to explain the self-verification process. Self-verification occurs when there is semantic congruence between the self-perceptions of meaningful feedback from others (the inputs) and the meanings attached to the identity (the identity standard) (Burke and Stets 2009). Generally, changes in behavior (outputs) are enough to bring these perceptions back into alignment with the identity standard (Burke 1991, 1996). In the case of persistent disturbances, however, people may abandon the identity in order to avoid the negative emotions caused by the disturbance (Cast and Burke 2002).

Some identities are easier to abandon than other identities, though (Thoits 2003). Ebaugh (1988) points out that some people never move out of the first stage of becoming an ex. Regarding religious affiliation, this includes people who consider dropping out of religion altogether but do not actually leave (i.e., stayers). People who consider dropping out of religion altogether and actually leave, on the other hand, are able to establish a new ex-role (i.e., leaver). According to role exit theory then, stayers are more likely to experience prolonged disturbances within a core identity, and they are also more likely to experience prolonged distress (Cast and Burke 2002; Ebaugh 1988). Therefore, I hypothesize that people who begin the disaffiliation process but remain affiliated will report more depressive symptoms than people who actually leave, and they will report a greater increase in negative health outcomes over time than people who complete the disaffiliation process.

Hypothesis 3: Individuals who consider dropping out of religion altogether but remain affiliated with a religious institution (stayers) will experience more depressive symptoms than individuals who drop out of organized religion altogether (leavers) after a period of several years.

Hypothesis 4: Individuals who consider dropping out of religion altogether but remain affiliated with a religious institution (stayers) will experience a larger increase in depressive symptoms than individuals who drop out of organized religion altogether (leavers).

Data and Methods

To examine the relationship between religious affiliation and depressive symptomatology, I use longitudinal data from PALS (Emerson and Sikkink 2012). PALS is a nationally representative multistage panel study of 2,610 American adults (wave 1, 2006) with follow-up data on 1,314 of the original respondents (wave 2, 2012). PALS’s multistage design generated a probability sample with a response rate of 58 percent at wave 1; this includes oversamples of African Americans, Hispanics, and Asian Americans. Specifically, PALS is designed to address the impact of religion and religious change across a number of individual and family-level experiences. PALS is particularly well suited for the present study because it is the only nationally representative longitudinal study with questions about the religious disaffiliation process and mental health.²

The longitudinal nature of PALS makes it possible to model the effect of changes in religious affiliation on mental health. Specifically, the wave 1 interview guide includes a 38-question health module gauging respondents’ physical, mental, and subjective well-being. During the wave 2 interview, respondents were asked the same set of questions about their physical, mental, and subjective health. Since PALS includes interviews with the same respondents at two points in time, I am able to examine the effect of religious affiliation on mental health at two points in time and the impact of religious affiliation on the change in mental health between waves. Thus, the current study provides a better test of the hypothesis that religious affiliation affects people’s mental health than studies utilizing cross-sectional data on religion and mental health.

Dependent Variable

My dependent variable is a count of the number of depressive symptoms experienced by respondents measured at two points in time. PALS includes three measures of behaviors commonly associated with depression. Specifically, PALS respondents were asked if they experienced feelings of depression, worthlessness, or hopelessness for a period lasting more than two weeks during the past 12 months. Each of these variables is a dichotomous measure wherein 1 indicates that the respondent experienced the symptom. I combined these measures to produce a count of respondents’ depressive symptoms during the past 12 months in 2006 (Chronbach’s α = .824) and 2012 (Chronbach’s α = .826). The variable ranges from 0 to 3. It is negatively skewed, with more than half of the wave 2 sample (57.6 percent) reporting no depressive symptoms lasting for a period of more than two weeks in the 12 months prior to the 2006 and 2012 PALS interviews.

Independent Variables

My key independent variable measures respondents’ religious affiliation at waves 1 and 2 of the PALS data set. PALS respondents were asked to provide their religious affiliation at wave 1 and wave 2. In total, 1,280 respondents provided their religious affiliation at both waves. Of these respondents, 1,028 (80.3 percent of the sample) were affiliated with a specific religious institution at wave 2. A total of 252 respondents (19.7 percent) reported no religious affiliation at wave 2. The percentage unaffiliated in the PALS data set is highly consistent with other summations of the total number of Nones in the United States in 2012 (e.g., Pew Forum on Religion and Public Life 2012). “Affiliated” and “unaffiliated” do not completely capture the nuance in these two groups, though.

The unaffiliated in this sample include 178 respondents with no religious affiliation at wave 1 and 74 respondents who cut ties with a specific religious institution between waves 1 and 2. The affiliated, on the other hand, include 138 respondents who seriously considered dropping out of religion between 2003 and 2012. More specifically, PALS respondents were asked the following two questions at waves 1 and 2, respectively: “In the past three years, have you seriously considered dropping out of religion altogether?” and “In the past six years, have you seriously considered dropping out of religion altogether?” Combining both questions with questions about respondents’ religious affiliation at waves 1 and 2 makes it possible to create a multinomial variable of respondents’ religious affiliation over a nine-year span.³ Following Vargas (2012), I refer to individuals who considered dropping out of religion altogether but were still affiliated with a specific religious group at wave 2 as stayers, and I refer to individuals with a religious affiliation at wave 1 and no religious affiliation at wave 2 as leavers. My analyses also include two unique, but stable, groups. Stable affiliates are religious affiliates who did not seriously consider dropping out of religion altogether between 2003 and 2012. Stable Nones are respondents with no religious affiliation at wave 1 or wave 2. Due to the small number of cases, the wave 2 analyses do not include women and men who were unaffiliated at wave 1 but are affiliated with a religious organization at wave 2. The mean for each group in Table 1 is the percentage of each group in the sample after accounting for missing data on the other variables in my analyses.

Table 1.

Descriptive Statistics for All Study Variables.

	Wave 1 (n =1,183)		Wave 2 (n = 926)
Variable	Mean/ Proportion	Standard Deviation	Mean/ Proportion	Standard Deviation	Range
Dependent variable
Depressive symptoms scale	0.57	1.02	0.53	0.98	0–3
Lagged (wave 1)			0.57	1.02	0–3
Depression	0.27	0.44	0.25	0.43	0–1
Lagged (wave 1)			0.26	0.44	0–1
Hopelessness	0.15	0.36	0.16	0.36	0–1
Lagged (wave 1)			0.15	0.35	0–1
Worthlessness	0.15	0.35	0.13	0.33	0–1
Lagged (wave 1)			0.16	0.37	0–1
Religious affiliation
Stable affiliate	0.77	0.42	0.70	0.46	0–1
Stayer	0.09	0.29	0.11	0.31	0–1
Leaver	0.04	0.20	0.06	0.25	0–1
Stable None	0.10	0.30	0.12	0.33	0–1
Control variable
Religious service attendance	3.71	2.17	3.64	2.28	1–7
Age	44.84	15.96	50.18	15.8	18–80^a
Female	0.51	0.50	0.52	0.50	0–1
White	0.69	0.46	0.70	0.46	0–1
Married	0.58	0.49	0.58	0.49	0–1
Education (1 = some college)	0.38	0.49	0.45	0.50	0–1
Income	3.16	1.67	3.25	1.78	1–6
Region (1 = South)	0.33	0.47	0.34	0.47	0–1
Subjective health	3.48	1.10	3.35	1.08	1–5

24–80 (wave 2).

Control Variables

In each of my models, I control for multiple measures related to religious commitment and mental health. Each of the variables described below is available in both waves of PALS. Religious service attendance is a common predictor of mental health (Barkan and Greenwood 2003; Childs 2010; Ellison 1991; Fenelon and Danielsen 2016). Some people retain their affiliation with a religious group without actively participating in any religious organization; others continue to participate in religious activities when they are unaffiliated.⁴ This phenomenon makes it important to consider the divergent experiences of active and inactive affiliates (both stable affiliates and stayers) and active and inactive disaffiliates (both leavers and stable Nones). Religious service attendance is an ordinal measure ranging from 1 (never) to 7 (more than once a week).

In addition to religious service attendance, I also control for age, sex (1 = female), race (1 = white), marital status (1 = married), education (1= some college), household income, region of the country (1 = South), and respondents’ subjective health. PALS respondents were asked to rate their overall health from 1 (poor) to 5 (excellent). Self-rated health is generally considered a measure of physical health (Krause and Jay 1994) and is a good indicator of respondents’ quality of life (Jylhä 2009). Several of the models described below also include a lagged measure of the dependent variable (i.e., the count of depressive symptoms reported at wave 1). Controlling for respondents’ mental health at wave 1 allows me to examine the effect of religious affiliation on changes in mental health over time. This is an important strength over cross-sectional studies of religion and health. Table 1 shows the range, mean, and standard deviation for all of the variables described above.

Missing Data

Wave 2 of PALS includes 1,314 cases. There are 1,183 cases with complete data at wave 1 and 926 cases with complete data at wave 2. The majority of missing cases (7 at wave 1 and 299 at wave 2) are missing on the dependent variable. Another 44 cases (10 at wave 1 and 34 at wave 2) are missing on the key independent variable (religious affiliation). I ran two separate analyses: one for the cases without missing data, and one using multiple imputation in STATA 14.0 to impute the eligible cases that were missing on one or more of the control variables (see Allison 2009 for a discussion of cases eligible for imputation with multiple imputation). The results of the imputed analyses produced results that are substantively and significantly similar to the analysis of complete cases. Therefore, I present only results for the primary analyses below since they are a more conservative estimate of the relationship between religious affiliation and mental health.

Analytic Strategy

I begin by plotting the pattern of change in mental health between waves for stable affiliates, stayers, leavers, and stable Nones. Comparing the average number of depressive symptoms across all four groups in 2006 and 2012 provides a simple test of all four of my hypotheses: (1) Do stayers and leavers experience more depressive symptoms than stable affiliates? (2) Do stayers and leavers experience more depressive symptoms than stable Nones? (3) Do stayers experience more depressive symptoms than leavers after several years? and (4) Do stayers experience a greater increase in depressive symptoms than leavers over time? Of course, comparing the means across groups does not account for possible covariates that might explain any observed relationship between religious affiliation and mental health. Therefore, I also estimate a series of regression models controlling for age, sex, race, marital status, education, household income, region of the country, subjective health, and religious service attendance, as described above.

In order to account for the clustered sample design of this dataset, I use the “svy” command in STATA 14.0 for each of the regressions described below. Since PALS oversamples African Americans, Asians, and Hispanics, I also use sample weights that more accurately reflect the racial makeup of the United States in each of my analyses. The weight variable also accounts for issues of nonresponse and the likelihood of being selected based on family size (Emerson, Sikkink, and James 2010). The first analysis examines the relationship between religious affiliation and depressive symptomatology at wave 1. The second analysis examines the same relationship at wave 2. The final analysis examines the change in depressive symptomatology between waves across the different types of religious affiliation.

My first set of analyses tests the hypothesis that stayers and leavers will experience more depressive symptoms than stable affiliates and stable Nones. Examining the relationship between religious affiliation and depressive symptomatology in the cross-section at wave 1 shows the short-term consequences of disturbances in religious affiliation on mental health. Identity theories in sociology (see Burke and Stets 2009) and Ebaugh’s (1988) role exit theory, however, are clear that any mental health problems observed in 2006 (wave 1) should be resolved by 2012 (wave 2) if the disturbance is removed and an ex-role identity is established. Therefore, my second set of analyses provides a test of my hypothesis that stayers will experience more depressive symptoms after a period of several years by examining the relationship between religious affiliation and depressive symptomatology at wave 2.

My final set of analyses tests the hypothesis that stayers will experience a greater increase in depressive symptoms than leavers over time. In order to examine the change in depressive symptoms between waves, I regress mental health on religious affiliation and respondents’ baseline mental health. Including depressive symptomatology at wave 1 shifts the interpretation of the other covariates in the model to the prediction of change in depressive symptomatology. The assumed causal effect of Y₁ on Y₂ plus the time-invariant covariates such as race and sex make the lagged dependent variable model the preferred model to test the effect of religious affiliation on changes in mental health over time (Finkel 1995; but also see Johnson 2005).⁵

The primary dependent variable in all of my analyses, the number of depressive symptoms experienced in the past 12 months, is a count that ranges from 0 to 3. Count variables typically require a Poisson or negative binomial model (Long 1997). Unlike the Poisson model, the negative binomial regression model does not assume that the conditional mean is equal to the conditional variance, a highly restrictive assumption. Therefore, I use the negative binomial regression model in order to account for overdispersion.⁶

Following the negative binomial regression, I examine each component of the count variable separately to determine whether the results of the negative binomial regressions are driven by a component of the count-dependent variable. Each component of my primary dependent variable is a dichotomous measure of respondents’ mental health. The construction of the count variable gives equal weight to each of these measures. Analyzing each of these measures separately is a way to test whether the count variable is biased by any of the unique measures. I use a series of logistic regression models to test the effect of religious commitment on each of these dichotomous outcomes. For each model, the results of postanalysis Wald tests are presented in order to test for differences between the model coefficients (e.g., stayers vs. leavers).

Results

Figure 1 plots the pattern of change in depressive symptoms between waves for stable affiliates, stayers, leavers, and stables Nones. Not surprisingly, stable affiliates report the lowest number of depressive symptoms at wave 1; stable affiliates and leavers are also the only groups to report a decrease in the number of depressive symptoms between waves. Stayers and stable Nones, on the other hand, report a modest increase in depressive symptoms between waves. Notably, stayers report the highest number of depressive symptoms at wave 1 (0.93 or 31 percent of all symptoms) and wave 2 (0.98 or 33 percent of all symptoms); and both stayers and stable Nones report an increase in the number of depressive symptoms experienced between waves. Based on these unadjusted means, it appears there is support for my hypotheses that individuals who consider dropping out of religion altogether will experience more depressive symptoms than individuals who do not consider dropping out of religion altogether and individuals who were never affiliated with a religious institution. There is also support for my hypothesis that stayers will experience more depressive symptoms than leavers when there is time for the latter to establish an ex-role identity. More specifically, stayers and leavers reported the greatest number of depressive symptoms at wave 1, and stayers reported the greatest number of symptoms at wave 2. In addition, the change in the unadjusted means between waves in Figure 1 provides support for my hypothesis that stayers will experience a greater increase in depressive symptoms between waves than leavers will. As previously noted, stayers report an increase in depressive symptoms between 2006 and 2012, but leavers report a decrease in depressive symptoms during the same period. Of course, the pattern in Figure 1 does not account for the possible covariates, nor does it show whether the relationships observed are statically significant. The models in Tables 2 through 4, however, provide support for all of my hypotheses and confirm most of the patterns shown in Figure 1.

Figure 1.

Mean depressive symptoms by wave and change in religious affiliation.

Table 2.

Mental Health Regressed on Religious Affiliation, Wave 1 (n = 1,183).

	Depressive Symptoms^a	Depression^b	Hopelessness^b	Worthlessness^b
	Model 1	Model 2	Model 3	Model 4
Religious commitment
Stayer	0.69***	0.95**	1.32**	1.17**
Leaver	0.33	0.25	0.91	0.73
Stable None	0.03	−0.11	0.21	0.53
Religious behavior
Religious service attendance	−0.02	−0.07	−0.02	0.04
Sociodemographics
Age	−0.01**	−0.02**	−0.02*	−0.02
Female	0.34*	0.52*	0.47	0.31
White	0.21	0.34	0.42	0.36
Married	−0.24	−0.31	−0.06	−0.50
Education (1 = some college)	−0.11	0.08	−0.18	−0.16
Income	−0.10*	−0.08***	−0.23**	−0.17
Region (1 = South)	0.11	0.16	0.25	0.23
Health
Subjective health	−0.36***	−0.48**	−0.62***	−0.71***
Wald test
Stayer vs. leaver	ns	ns	ns	ns
Stayer vs. stable None	*	**	*	ns
Leaver vs. stable None	ns	ns	ns	ns

Negative binomial coefficients.

Logistic regression coefficients.

p < .05. **p < .01. ***p < .001 (two-tailed tests).

Table 3.

Mental Health Regressed on Religious Affiliation, Wave 2 (n = 926).

	Depressive Symptoms^a	Depression^b	Hopelessness^b	Worthlessness^b
	Model 1	Model 2	Model 3	Model 4
Religious commitment
Stayer	0.97***	1.16***	1.52***	1.41**
Leaver	−0.19	−0.33	−0.21	−0.02
Stable None	−0.15	−0.01	−0.44	−0.25
Religious behavior
Religious service attendance	−0.12**	−0.11*	−0.15*	−0.13
Sociodemographics
Age	−0.02**	−0.03**	−0.02	−0.03*
Female	0.00	0.02	−0.18	−0.02
White	0.31	0.43	0.63	0.83*
Married	−0.37	−0.63**	−0.25	−0.32
Education (1 = some college)	−0.12	−0.05	−0.35	−0.33
Income	−0.05	−0.09	−0.05	−0.07
Region (1 = South)	0.21	0.20	0.50	0.13
Health
Subjective health	−0.36***	−0.43**	−0.62***	−0.55**
Wald test
Stayer vs. leaver	***	*	*	*
Stayer vs. stable None	***	**	***	***
Leaver vs. stable None	ns	ns	ns	ns

Negative binomial coefficients.

Logistic regression coefficients.

p < .05. **p < .01. ***p < .001 (two-tailed tests).

Table 4.

Mental Health Regressed on Religious Affiliation and Prior Mental Health, Wave 2 (n = 926).

	Depressive Symptoms^a	Depression^b	Hopelessness^b	Worthlessness^b
	Model 1	Model 2	Model 3	Model 4
Religious commitment
Stayer	0.70**	0.87**	1.30***	1.18*
Leaver	−0.08	−0.23	−0.11	0.07
Stable None	−0.25	−0.08	−0.51	−0.35
Religious behavior
Religious service attendance	−0.13***	−0.12*	−0.17*	−0.14
Sociodemographics
Age	−0.02**	−0.02**	−0.02	−0.03*
Female	−0.11	−0.13	−0.35	−0.18
White	0.27	0.31	0.52	0.73*
Married	−0.25	−0.51*	−0.10	−0.18
Education (1 = some college)	−0.06	0.12	−0.22	−0.20
Income	−0.03	−0.06	−0.03	−0.04
Region (1 = South)	0.16	0.13	0.43	0.04
Health
Subjective health	−0.27**	−0.29*	−0.49**	−0.44*
Lagged dependent variable	0.39***	0.64***	0.52***	0.47***
Wald test
Stayer vs. leaver	*	ns	ns	ns
Stayer vs. stable None	***	*	***	**
Leaver vs. stable None	ns	ns	ns	ns

Negative binomial coefficients.

Logistic regression coefficients.

p < .05. **p < .01. ***p < .001 (two-tailed tests).

The models in Table 2 (wave 1) are tests of the hypotheses that people who consider dropping out of religion altogether will experience more depressive symptoms than individuals who do not consider dropping out of religion altogether and individuals who were never affiliated with a religious institution. The models in Table 3 (wave 2), on the other hand, are tests of the hypothesis that stayers will experience more depressive symptoms than leavers after a period of several years. Each model in Tables 2 and 3 shows the effect of religious affiliation on a unique dependent variable. The first model in both tables displays the results for the negative binomial model. Each subsequent model in Tables 2 and 3 displays the results of a logistic regression model predicting one of the depressive symptoms that make up the count variable. The coefficients in the negative binomial regression models represent the percentage change in depressive symptoms for a unit change in the independent variable after computing [expβ – 1]. The coefficients in the logistic regression models, on the other hand, represent the odds of experiencing the dependent variable after computing [expβ]. Similar to Figure 1, the models in Table 2 provide partial support for the hypotheses that stayers and leavers will report more depressive symptoms than stable affiliates and stable Nones, and the models in Table 3 provide clear support for the hypothesis that stayers will experience more mental health problems than leavers over time.

The models in Table 2 indicate that people who seriously considered dropping out of religion between 2003 and 2006 but were still affiliated with a religious institution in 2006 experienced more depressive symptoms than religious affiliates who did not seriously consider dropping out of religion. Being a stayer at wave 1 increases the expected number of depressive symptoms by a factor of 1.98 (expβ) compared to stable affiliates (p< .001). A postanalysis Wald test also indicates that the difference in the expected number of depressive symptoms for stayers and stable Nones is statistically significant (p < .05). On the other hand, being a leaver does not significantly increase the expected number of depressive symptoms compared to stable affiliates or stable Nones. These same patterns repeat across all of the logistic regression models in Table 2 with one exception: The coefficient for stayers is not significantly different from the coefficient for stable Nones in model 4. Thus, stayers are more likely to feel depressed, hopeless, and worthless than stable affiliates, and they are more likely to feel depressed and hopeless than stable affiliates. Leavers, on the other hand, are not significantly different from stable affiliates or stable Nones, and they are less likely than stayers to report feelings of depression or hopelessness.

Models 1 through 4 in Table 3 show a similar pattern. The difference, however, is that religious affiliation in measured over a nine-year period (from 2003 to 2012). Model 1 in Table 3 indicates that being a stayer from 2003 to 2012 increases the expected number of depressive symptoms in 2012 by a factor of 2.63 (expβ) compared to stable affiliates (p < .001). Older persons, people in better physical health, and people who attend religious services more often also report fewer depressive symptoms. The latter variable demonstrates the unique relationship between religious affiliation and mental health. That is, net of religious service attendance, the expected number of depressive symptoms is higher for stayers than it is for stable affiliates. Like model 1 in Table 2, however, the differences between stable affiliates and leavers in model 1 of Table 2 is not statistically significant. This pattern persists across all of the models in Table 3 and is consistent with the assumption that stayers will fare the worst over a prolonged period of time (hypothesis 3).

Compared to stable affiliates, stayers are more likely to report having felt depressed in the past 12 months, they are more likely to report having felt worthless in the past 12 months, and they are more likely to report having felt hopeless in the past 12 months. Postanalysis Wald tests for all of the models in Table 3 also indicate that people who considered dropping out of religion between 2003 and 2012 experienced poorer mental health in the 12 months prior to the 2012 PALS interview when they remained affiliated with a religious group (stayers) than when they abandoned organized religion between 2006 and 2012 (leavers). Together, these models provide clear support for the hypothesis that individuals who consider dropping out of religion altogether but remain affiliated with a religious institution (stayers) will experience more depressive symptoms than individuals who drop out of organized religion altogether (leavers) over time. To test if stayers also experienced a greater increase in depressive symptoms than leavers did between 2006 and 2012, though, the models in Table 4 include a lagged measure of the dependent variable in models 1 through 4 of Table 3.

The inclusion of the lagged dependent variable in the models in Table 4 shifts the interpretation of the other covariates in the negative binomial model and the logistic regression models to the prediction of change in depressive symptoms between waves. These models provide a test of my hypothesis that stayers will experience a larger increase in depressive symptoms than leavers will. Being a stayer increases the expected change in the number of depressive symptoms by a factor of 2.01 (expβ) compared to stable affiliates. The postanalysis Wald tests for model 1 in Table 4 also indicate that stayers experience a greater increase in the expected number of depressive symptoms than leavers (p < .05) even though the difference between stayers and leavers is not statistically significant in models 2 through 4.⁷ Thus, model 1 in Table 4 provides support for my hypothesis that stayers will experience a larger increase in depressive symptoms than leavers will, and it confirms the pattern shown in Figure 1.

Discussion

The percentage of Americans with no religious affiliation has nearly tripled since the early 1990s (Pew Forum on Religion and Public Life 2012), yet most of the unaffiliated were raised in religious homes (Kosmin and Keysar 2008). Since active involvement in a religious community is generally associated with better mental health outcomes (Koenig 2012), it is important to consider the role that religious affiliation plays in the health and well-being of this growing population. My investigation was motivated by two suspicions. First I suspected that people who considered dropping out of religion (stayers) and people who left (leavers) would experience more mental health problems than people who did not consider dropping out of religion (stable affiliates and stable Nones). Second, I suspected that people who remained affiliated (stayers) would experience greater distress than people who did not (leavers) after several years. Indeed, I did find that religious affiliation affects depressive symptomatology, but not exactly in the expected way.

In sum, the results described above suggest that stayers experience more mental health problems than stable affiliates, leavers, and stables Nones and that they experience a greater increase in mental health problems over time. This is consistent with Burke and Stets’s (2009) identity theory and with Ebaugh’s (1988) role exit theory. More specifically, disturbances like considering dropping out of religion cause negative emotions like depression (Burke 1991, 1996). However, abandoning an identity can be a useful way to eliminate disturbances (Cast and Burke 2002). Therefore, it makes sense that stayers experience more depressive symptoms and a greater increase in depressive symptoms over time than the other groups, while leavers are not significantly different from stable affiliates or stable Nones. There are several important implications, limitations, and directions for future research to consider in light of these results.

According to Vargas (2012), less than half of the women and men who seriously consider dropping out of religion actually leave. Vargas based his assessment on data from wave 1 of PALS, but this still appears to be the case six years later. More than 23 percent of religiously affiliated Americans seriously considered dropping out of religion between 2003 and 2012. At wave 2 of PALS, 53.5 percent of these people still identified with a specific religious institution. Although this is a sizeable change from the 62 percent reported by Vargas, it means that 11.4 percent of the U.S. population is still positioned to experience “the dark side of religion” based on the current results. To some extent, this is consistent with previous research on religion and mental health, but the current study provides greater nuance to our understanding of the role that religious affiliation plays in the relationship between religion and mental health.

Few studies on religion and mental health actually consider the role of religious affiliation (see Fenelon and Danielsen 2016 for an exception). Instead, behaviors like religious service attendance or prayer/meditation are generally the key focal points (Koenig 2012). Like most of these studies, there is a negative relationship between religious service attendance and my measure of depressive symptomatology. However, my results also show there is a contingent relationship between religious affiliation and mental health net of religious behaviors like attending religious services. It is important for future research to consider the distinctive ways that religious affiliation and other religious behaviors affect mental health.

In addition to highlighting the contingent role that religious affiliation plays in shaping the mental health of the affiliated and the unaffiliated, this research calls into question the importance of religious affiliation and disaffiliation for other life outcomes. Since there is clear variability in the mental health of stayers and leavers, stable affiliates, and stable Nones based on my results, it is also likely that there is variability between these groups in other areas. Recent scholarship demonstrates religion’s important role in marriage markets (McClendon 2016), welfare attitudes (VanHeuvelen 2014), cross-racial interactions (Park and Bowman 2015), and the work-family interface (Reynolds and May 2014), to note just a few examples. Future research should consider the possibility that religious affiliation plays an important role in shaping people’s experiences in these and other domains.

This study is also an improvement over cross-sectional studies of religion and mental health. A key issue with cross-sectional studies of religion and mental health is causality. Do religious attitudes/behaviors cause changes in mental health, or do changes in mental health shape religious attitudes/behaviors? Cross-sectional studies of religion and health generally assume the former (see Ellison and Taylor 1996 for an exception), but a number of longitudinal studies show that some people turn to (or away from) religion during times of trial in their lives (Ferraro and Kelley-Moore 2001; Oates 2013; Pargament et al. 2004). One strength of the current study is my test of the relationship between religious affiliation and changes in depressive symptomatology over time.

Last, my findings raise important questions about religious identities and role exit. Identities are an important source of meaning and purpose (Thoits 1983, 1986), but some identities are more salient than others (Stryker 2002). Thus, it is easier to exit some identities (e.g., voluntary identities such as friend, religious/spiritual, volunteer/social) than others (e.g., obligatory identities such as worker, partner/spouse, or parent) (Thoits 2003). Although Thoits (2003) considers religious identities easier to exit, Park and Edmonson (2012) point out that religion is a core identity for most people. If stayers experience more depressive symptoms than leavers do because they are unable to give up the identity, religious identities are likely more than voluntary identities for these people. Understanding this difference is an important avenue for future research on identities and role exit. Vargas (2012) provides some insight into the process of staying versus leaving, but PALS does not specifically ask respondents to explain why they considered dropping out of religion. Likewise, PALS does not ask respondents who considered dropping out of religion why they decided to stay or why they decided to leave. Qualitative research can provide important insights into the reasons behind the commitment of religious participants and ex-participants (e.g., Packard and Hope 2015). It is possible that poor mental health motivates people to stick with their religious institution. This, in turn, might drive the observed increase in mental health problems for those who remain affiliated as they expected their feelings of depression, hopelessness, or worthlessness to wane. Future research should consider this complex relationship between religion, identity, and mental health.

A couple of additional limitations are also worth noting here. First, my analyses do not consider the role of religious tradition. Belonging to a religious institution is more important for the members of particular religious traditions (Bender et al. 2013). Unfortunately, less than 10 percent of PALS respondents belong to non-Christian faiths. Small sample sizes across different religious groups also make it difficult to raise meaningful comparisons between groups. It is possible that stayers experience better mental health outcomes when they belong to religious institutions that place a strong emphasis on community and formal worship. It is also possible that leavers experience better health outcomes when they leave religious traditions that place less emphasis on the congregational aspects of religious life. Future research should examine the roles that context and religious tradition play in the relationship between religious affiliation and mental health.

Second, my measure of depressive symptomatology is based on just three items: depression, worthlessness, and hopelessness. More traditional measures of depressive symptomatology like the Center for Epidemiological Studies Depression Scale (Radloff 1977) or the Hospital Anxiety and Depression Scale (Zigmond and Snaith 1983) also include behavioral (e.g., “I talked less than usual.”), physiological (e.g., “My sleep was restless.”), and attitudinal (e.g., “I have lost interest in my appearance.”) measures of depression. These items are not available in PALS, and it is possible that the experiences of stayers, leavers, stable affiliates, and stable Nones described here are limited to the emotional components of depression that are captured by my dependent variable. Nevertheless, my findings provide an important starting point for thinking about the contingent relationship between religious affiliation and mental health. Future research should consider the relationship between religious affiliation and the other dimensions of depressive symptomatology and other measures of mental health.

Last, there is no way to know if repeated changes in religious affiliation affect the observed changes in mental health in different ways. PALS’s multistage panel design makes it possible to model changes in mental health during a six-year period, but we are constantly adjusting our behaviors in order to achieve semantic congruence between our perceptions of meaningful feedback from others and the meanings we attach to an identity (Burke and Stets 2009). Thus, leavers might rejoin organized religion, and stayers might decide to leave. It is possible, for example, that an individual considered dropping out of religion in 2006, dropped out of religion in 2007, and returned to religion in 2010. This individual is considered a stayer in the current analysis if he or she is still affiliated in 2012, but his or her experience is likely different from that of someone who persistently thinks about dropping out of religion but maintains his or her affiliation with the religious group. Reducing the number of years between the measures of religious affiliation and mental health can provide a better test of the control system described above.

Conclusion

In this article, I examined the role that religious affiliation plays in the relationship between religion and depressive symptomatology. My findings indicate that individuals who seriously consider dropping out of religion altogether but maintain their affiliation with a religious institution experience more depressive symptoms than people who do not seriously consider dropping out of religion altogether, people who were never affiliated, and people who actually leave. Likewise, they experience a greater increase in depressive symptoms than the members of these other groups do over time. Moving forward, it is important to consider how the different causes of variability in religious affiliation are associated with health and well-being and how the relationship between religious affiliation and mental health may vary across different religious traditions, different settings, and when there are different motivations to stay or leave. Research on these topics will further our understanding of one of the fastest-growing religious groups in the United States during the past 25 years and the mental health of these Nones and almost Nones.

Footnotes

Acknowledgements

The author thanks Ashley Barr, Dennis Condron, Katie James, and participants in the Georgia Workshop on Culture, Power, and History for their helpful comments on earlier drafts of this manuscript.

Notes

References

Acevedo

Gabriel A.

Ellison

Christopher G.

Xiaohe H.

2014. “Is It Really Religion? Comparing the Main and Stress-buffering Effects of Religious and Secular Civic Engagement on Psychological Distress.” Society and Mental Health 4(2):111–28.

Allison

Paul D.

2009. “Missing Data.” Pp. 72–89 in The Sage Handbook of Quantitative Methods in Psychology, edited by Missap

R. E.

Maydeu-Olivares

Thousand Oaks, CA: Sage.

Allison

Paul D.

2012. “Do We Really Need Zero-inflated Models?” Statistical Horizons. Retrieved March 20, 2017 (http://statisticalhorizons.com/zero-inflated-models).

Baker

Joseph O.

Whitehead

Andrew L.

2016. “Gendering (Non)Religion: Politics, Education, and Gender Gaps in Secularity in the United States.” Social Forces 94(4):1623–45.

Barkan

Steven E.

Greenwood

Susan F.

2003. “Religious Attendance and Subjective Well-being among Older Americans: Evidence from the General Social Survey.” Review of Religious Research 45(2):116–29.

Bender, Courtney, Wendy Cadge, Peggy Levitt, and David Smilde. 2013. “Religion on the Edge: De-centering and Re-centering.” Pp. 1–20 in Religion on the Edge: De-centering and Re-centering the Sociology of Religion, edited by Bender

Cadge

Levitt

Smilde

New York: Oxford University Press.

Burke

Peter J.

1991. “Identity Processes and Social Stress.” American Sociological Review 56(6):836–49.

Burke

Peter J.

1996. “Social Identities and Psychosocial Stress.” Pp. 141–74 in Psychosocial Stress: Perspectives on Structure, Theory, Life Course, and Methods, edited by Kaplan

H. B.

San Diego, CA: Academic Press.

Burke

Peter J.

Stets

Jan E.

2009. Identity Theory. New York: Oxford University Press.

10.

Cast

Alicia D.

Burke

Peter J.

2002. “A Theory of Self-esteem.” Social Forces 80(3):1041–68.

11.

Childs

Ellen

. 2010. “Religious Attendance and Happiness: Examining Gaps in the Current Literature: A Research Note.” Journal for the Scientific Study of Religion 49(3):550–60.

12.

Ebaugh

Helen Rose Fuchs

. 1988. Becoming an Ex: The Process of Role Exit. Chicago: University of Chicago Press.

13.

Eliassen

A. Henry

Taylor

John

Lloyd

Donald A.

2005. “Subjective Religiosity and Depression in the Transition to Adulthood.” Journal for the Scientific Study of Religion 44(2):187–99.

14.

Ellison

Christopher G.

1991. “Religious Involvement and Subjective Well-being.” Journal of Health and Social Behavior 32(1):80–99.

15.

Ellison

Christopher G.

1995. “Race, Religious Involvement and Depressive Symptomatology in a Southeastern U.S. Community.” Social Science & Medicine 40(11):1561–72.

16.

Ellison

Christopher G.

Bradshaw

Matt

Kuyel

Nilay

Marcum

Jack P.

2012. “Attachment to God, Stressful Life Events, and Changes in Psychological Distress.” Review of Religious Research 53(4):493–511.

17.

Ellison

Christopher G.

Fang

Qijuan

Flannelly

Kevin J.

Steckler

Rebecca A.

2013. “Spiritual Struggles and Mental Health: Exploring the Moderating Effects of Religious Identity.” International Journal for the Psychology of Religion 23(3):214–29.

18.

Ellison

Christopher G.

Gay

David A.

Glass

Thomas A.

1989. “Does Religious Commitment Contribute to Individual Life Satisfaction?” Social Forces 68(1):100–23.

19.

Ellison

Christopher G.

Lee

Jinwoo

. 2010. “Spiritual Struggles and Psychological Distress: Is There a Dark Side of Religion?” Social Indicators Research 98(3):501–17.

20.

Ellison

Christopher G.

Taylor

Robert Joseph

. 1996.“Turning to Prayer: Social and Situational Antecedents of Religious Coping among African Americans.” Review of Religious Research 38(2):111–31.

21.

Ellison

Christopher G.

Zhang

Wei

Krause

Neal

Marcum

John P.

2009. “Does Negative Interaction in the Church Increase Psychological Distress? Longitudinal Findings from the Presbyterian Panel Survey.” Sociology of Religion 70(4):409–31.

22.

Emerson

Michael O.

Sikkink

David

. 2012. “Portraits of American Life Survey, 2nd Wave.”

23.

Emerson

Michael O.

Sikkink

David

James

Adele D.

2010. “The Panel Study on American Religion and Ethnicity: Background, Methods, and Selected Results.” Journal for the Scientific Study of Religion 49(1):162–71.

24.

Fenelon

Andrew

Danielsen

Sabrina

. 2016. “Leaving My Religion: Understanding the Relationship between Religious Disaffiliation, Health, and Well-being.” Social Science Research 57:49–62.

25.

Ferraro

Kenneth F.

Kelley-Moore

Jessica A.

2001. “Religious Seeking among Affiliates and Non-affiliates: Do Mental and Physical Health Problems Spur Religious Coping?” Review of Religious Research 42(3):229–51. doi: 10.2307/3512568.

26.

Finkel

Steven E.

1995. Causal Analysis with Panel Data. Thousand Oaks, CA: Sage.

27.

Galek

Kathleen

Flannelly

Kevin J.

Ellison

Christopher G.

Silton

Nava R.

Jankowski

Katherine R. B.

2015. “Religion, Meaning and Purpose, and Mental Health.” Psychology of Religion and Spirituality 7(1):1–12.

28.

Hayward

R. David

Krause

Neal

Ironson

Gail

Hill

Peter C.

Emmons

Robert

. 2016. “Health and Well-being among the Non-religious: Atheists, Agnostics, and No Preference Compared with Religious Group Members.” Journal of Religion & Health 55(3):1024–37.

29.

Heise

David R.

2007. Expressive Order: Confirming Sentiments in Social Actions. New York: Springer.

30.

Johnson

David

. 2005. “Two-wave Panel Analysis: Comparing Statistical Methods for Studying the Effects of Transitions.” Journal of Marriage and Family 67(4):1061–75.

31.

Jylhä

Marja

. 2009. “What Is Self-rated Health and Why Does It Predict Mortality? Towards a Unified Conceptual Model.” Social Science & Medicine 69(3):307–16.

32.

Koenig

Harold G.

2012. “Religion, Spirituality, and Health: The Research and Clinical Implications.” ISRN Psychiatry 2012:1–33.

33.

Kosmin

Barry

Keysar

Ariela

. 2008. American Nones: The Profile of the No Religion Population. Hartford, CT: Institute for the Study of Secularity in Society.

34.

Krause

Neal

. 1994. “Stressors in Salient Social Roles and Well-being in Later Life.” Journal of Gerontology 49(3):P137–48.

35.

Krause

Neal

. 1998. “Stressors in Highly Valued Roles, Religious Coping, and Mortality.” Psychology and Aging 13(2):242–55.

36.

Krause

Neal

. 2006. “Religious Doubt and Psychological Well-being: A Longitudinal Investigation.” Review of Religious Research 47(3):287–302.

37.

Krause

Neal

Ellison

Christopher G.

2009. “The Doubting Process: A Longitudinal Study of the Precipitants and Consequences of Religious Doubt in Older Adults.” Journal for the Scientific Study of Religion 48(2):293–312.

38.

Krause

Neal

Wulff

Keith M.

2004. “Religious Doubt and Health: Exploring the Potential Dark Side of Religion.” Sociology of Religion 65(1):35–56.

39.

Krause

Neal M.

Jay

Gina M.

1994. “What Do Global Self-rated Health Items Measure?” Medical Care 32(9):930–42.

40.

Large

Michael D.

Marcussen

Kristen

. 2000. “Extending Identity Theory to Predict Differential Forms and Degrees of Psychological Distress.” Social Psychology Quarterly 63(1):49–59.

41.

Lim

Chaeyoon

Putnam

Robert D.

2010. “Religion, Social Networks, and Life Satisfaction.” American Sociological Review 75(6):914–33.

42.

Long

J. Scott

. 1997. Regression Models for Categorical and Limited Dependent Variables. Thousand Oaks, CA: Sage.

43.

Manning

Christel J.

2013. “Unaffiliated Parents and the Religious Training of Their Children.” Sociology of Religion 74(2):149–75.

44.

McClendon

David

. 2016. “Religion, Marriage Markets, and Assortative Mating in the United States.” Journal of Marriage and Family 78(5):1399–421.

45.

Mochon

Daniel

Norton

Michael I.

Ariely

Dan

. 2011. “Who Benefits from Religion?” Social Indicators Research 101(1):1–15.

46.

Oates

Gary L.

2013. “Enhanced Religiosity Following Illness? Assessing Evidence of Religious Consolation among Black and White Americans.” Review of Religious Research 55(4):597–613.

47.

Packard

Josh

Hope

Ashleigh

. 2015. Church Refugees: Sociologists Reveal Why People Are DONE with Church but Not Their Faith. Loveland, CO: Group.

48.

Pargament

Kenneth I.

Koenig

Harold G.

Tarakeshwar

Nalini

Hahn

June

. 2004. “Religious Coping Methods as Predictors of Psychological, Physical and Spiritual Outcomes among Medically Ill Elderly Patients: A Two-year Longitudinal Study.” Journal of Health Psychology 9(6):713–30. doi: 10.1177/1359105304045366.

49.

Park

Crystal L.

Edmondson

Donald

. 2012. “Religion as a Source of Meaning.” Pp. 145–62 in Meaning, Mortality, and Choice: The Social Psychology of Existential Concerns, edited by Shaver

P. R.

Mikulincer

Washington, DC: American Psychological Association.

50.

Park

Julie J.

Bowman

Nicholas A.

2015. “Religion as Bridging or Bonding Social Capital: Race, Religion, and Cross-racial Interaction for College Students.” Sociology of Education 88(1):20–37.

51.

Pew Forum on Religion and Public Life. 2010. “Religion among the Millennials.” Retrieved December 19, 2016 (http://www.pewforum.org/files/2010/02/millennials-report.pdf).

52.

Pew Forum on Religion and Public Life. 2012. “‘Nones’ on the Rise: One-in-five Adults Have No Religious Affiliation.” Retrieved December 19, 2016 (http://www.pewforum.org/uploadedFiles/Topics/Religious_Affiliation/Unaffiliated/NonesOnTheRise-full.pdf).

53.

Pew Forum on Religion and Public Life. 2015. “America’s Changing Religious Landscape.” Retrieved December 19, 2016 (http://www.pewforum.org/2015/05/12/americas-changing-religious-landscape/).

54.

Radloff

Lenore Sawyer

. 1977. “The CES-D Scale: A Self-report Depression Scale for Research in the General Population.” Applied Psychological Measurement 1(3):385–401.

55.

Reynolds

Jeremy

May

Matthew

. 2014. “Religion, Motherhood, and the Spirit of Capitalism.” Social Currents 1(2):173–88.

56.

Robinson

Dawn T.

2007. “Control Theories in Sociology.” Annual Review of Sociology 33(1):157–74.

57.

Ross

Catherine E.

1990. “Religion and Psychological Distress.” Journal for the Scientific Study of Religion 29(2):236–45.

58.

Schnittker

Jason

. 2001. “When Is Faith Enough? The Effects of Religious Involvement on Depression.” Journal for the Scientific Study of Religion 40(3):393–411.

59.

Schwadel

Philip

Falci

Christina D.

2012. “Interactive Effects of Church Attendance and Religious Tradition on Depressive Symptoms and Positive Affect.” Society and Mental Health 2(1):21–34.

60.

Speed

David

Fowler

Ken

. 2017. “Good for All? Hardly! Attending Church Does Not Benefit Religiously Unaffiliated.” Journal of Religion and Health 56(3):986–1002.

61.

Stryker

Sheldon

. 1980. Symbolic Interactionism: A Social Structural Version. Menlo Park, CA: Benjamin-Cummings.

62.

Stryker

Sheldon

. 2002. “Traditional Symbolic Interactionism, Role Theory, and Structural Symbolic Interactionism: The Road to Identity Theory.” Pp. 202–21 in Handbook of Sociological Theory, edited by Turner

J. H.

New York: Springer.

63.

Taylor

Peter D.

McPeek

Mark A.

2013. “The Rise of the ‘Nones’: Does Education Explain the Decline in Religious Affiliation?” American Journal of Education 119(3):335–40.

64.

Thoits

Peggy A.

1983. “Multiple Identities and Psychological Well-being: A Reformulation and Test of the Social Isolation Hypothesis.” American Sociological Review 48:174–87.

65.

Thoits

Peggy A.

1986. “Multiple Identities: Examining Gender and Marital Status Differences in Distress.” American Sociological Review 51:259–72.

66.

Thoits

Peggy A.

2003. “Personal Agency in the Accumulation of Multiple Role-identities.” Pp. 179–94 in Advances in Identity Theory and Research, edited by Burke

P. J.

Owens

T. J.

Serpe

Thoits

P. A.

New York: Kluwer-Plenum.

67.

VanHeuvelen

Tom

. 2014. “The Religious Context of Welfare Attitudes.” Journal for the Scientific Study of Religion 53(2):268–95.

68.

Vargas

Nicholas

. 2012. “Retrospective Accounts of Religious Disaffiliation in the United States: Stressors, Skepticism, and Political Factors.” Sociology of Religion 73(2):200–23.

69.

Zigmond

Anthony S.

Snaith

R. Philip

. 1983. “The Hospital Anxiety and Depression Scale.” Acta Psychiatrica Scandinavica 67(6):361–70.