Sage Journals: Discover world-class research

Abstract

Many studies have used retrospective survey measures to examine changes in individuals’ religious affiliation, but studies examining changes to individuals’ core religious beliefs are comparatively rare. This is likely because surveys rarely contain measures of both current and past beliefs. Using data from a probability sample of U.S. adults that includes a measure of individuals’ current and age 16 belief in God, the authors examine the predictors of an individual’s adopting an atheistic worldview. Overall, 6 percent of the sample report moving from a nonatheistic worldview to an atheistic worldview. This rate is higher among those who said they had weaker belief in God at age 16, men, those with higher incomes, and some sexually minoritized groups. This rate is lower among older individuals, political conservatives, and some racially or ethnically minoritized groups. The authors discuss the implications of these findings for the study of nonbelief and the measurement of religious change.

Keywords

atheism surveys belief in god religion

Socially constructed systems of meanings and beliefs, or worldviews, provide a shared lens by which individuals interpret and understand the world (Berger 1967; Johnson, Hill, and Cohen 2011; Sire 2015). When these worldviews are shared by entire societies, they are taken for granted (Berger 1967). In societies with worldview pluralism, individuals come to adopt worldviews from the possible options known to them (Berger 1967). Individuals do not just decide to adopt a worldview in the way they would choose to purchase a car (Bankston 2002); instead social interactions and structures shape not only the worldviews available to a person but also their plausibility (Berger 1967; Corcoran 2013; Smith and Emerson 1998; Stark and Finke 2000; Wellman 2008). One’s position in the social structure facilitates some worldviews, while constraining others. That is, the costs of adopting a particular worldview will vary by one’s social location.

In the United States, the percentage of people who believe in God has declined over time though it still represents a large majority of the population (Jones 2022). This decline may reflect an increase in those who have never believed in God as well as those who shifted from a nonatheistic worldview (e.g., believing in God or being agnostic) to an atheistic worldview: explicitly not believing in God. General Social Survey (GSS) data¹ descriptively show that percentages for both have increased since 2004 (Davern et al. 2021). Although there is research on predictors of theistic beliefs and certainty in those beliefs (Brinkerhoff and Mackie 1993; Corcoran 2013; Durkin and Greeley 1991; Nelsen 1981; Wollschleger 2021), there is scarce research on correlates of atheistic worldviews (Gervais, Najle, and Caluori 2021; Sherkat 2008), particularly as distinct from identifying as an atheist (Bainbridge 2005; Beit-Hallahmi 2007; Hunter 2010; Langston 2014; LeDrew 2013; Smith 2011). We know little about the factors that affect adopting an atheistic worldview.

Understanding changes in religiosity—whether at the societal or individual level—is a core concern of the social scientific study of religion. Measuring such changes, though, is methodologically challenging. At the individual level, data produced from panel studies do allow some assessment of changes in individuals’ religious identities, beliefs, and behaviors. The intensive and costly nature of panel studies, though, means that individuals are often only followed for a relatively short period of time. Retrospective questions offered within cross-sectional surveys provide an alternative strategy for measuring changes in an individual’s religion. Such retrospective questions have tended to focus on an individual’s religious identity or affiliation or, to a lesser extent, their religious behaviors at a younger age, often at around age 12 or 16 (e.g., Davern et al. 2021). It is much rarer for surveys to include retrospective questions about core religious beliefs (see Davern et al. 2021 for an exception), such as an individual’s belief in God at those earlier ages. Indeed, we are not aware of any previous major probability survey of U.S. adults that has included such a question.²

In this study, we use unique items included in a probability survey of U.S. adults asking individuals both their current belief in God and their belief in God at age 16. On the basis of past theory and research, we propose expectations concerning the forces that increase or decrease an individual’s odds of adopting an atheistic worldview between age 16 and the time of the survey. Logistic regression models reveal that an individual’s level of certainty in their belief in God at age 16 is significantly associated with their odds of adopting an atheistic worldview. Net of this age 16 level of belief, our analysis finds that men, those with higher incomes, and those identifying as bisexual have greater odds of adopting an atheistic worldview relative to women, those with lower incomes, and those identifying as heterosexual.

Atheism as Identity and/or Worldview

The increase in the proportion of the U.S. population who do not identify with a religion in the past 30 years has been one of the more dramatic trends in social life (Burge 2020; Mitchell 2019; Wiertz and Lim 2021). Not all of these “nones,” of course, identify as atheists. The proportion of U.S. adults explicitly identifying as atheists has grown, although it remains low overall, at about 5 percent (Mitchell 2019).

Identifying as an atheist is a somewhat distinct concept from holding an atheistic worldview—that is, not believing in God. Indeed, surveys and studies have often conflated measures of identity and belief when examining atheism (Stewart 2016). The percentage of the U.S. population who hold an atheistic worldview has also increased slightly to about 7 percent in recent surveys (Davern et al. 2021). Studies have shown that only about half of those who hold an atheistic worldview identify as atheists, while practically all of those who identify as atheists hold an atheistic worldview. In other words, holding an atheistic worldview appears to be a necessary but not sufficient condition for identifying as an atheist. The same studies have identified a number of social forces that increase or decrease the likelihood that an individual who holds an atheistic worldview will identify as an atheist (Scheitle, Corcoran, and Hudnall 2019). Less research has considered the forces that increase or decrease the likelihood of adopting an atheistic worldview.

Religious Doubt and Adopting an Atheistic Worldview

The increase in those who do not identify with religion in the United States (Burge 2020; Mitchell 2019; Wiertz and Lim 2021) has spurred research on possible causes of religious disaffiliation and adopting nonreligious identities. Berger (1967) suggested that in religiously pluralistic contexts, all religious worldviews become less plausible. Because one’s certainty in religious explanations is not externally provided, one must provide it themselves, which he predicts should increase levels of uncertainty and, consequently, decrease religiosity (Berger 1967).

Leaving aside the particular causes of religious uncertainty, some studies have investigated the role of religious skepticism or doubting religious beliefs and practices in disaffiliation and adopting an atheist identity. Religious doubt or skepticism is positively associated with religious disaffiliation (Perry and Longest 2019; Vargas 2012). Religious doubt is also an important stage in the process of adopting an atheistic identity for many atheists (Hunsberger and Altemeyer 2006; LeDrew 2013; Smith 2011). Past research on religious doubt and atheism has primarily been qualitative in nature and thus unable to identify whether religious doubt in adolescence is associated with adopting an atheistic worldview as an adult. Brinkerhoff and Mackie (1993) found that the older people are “before encountering doubts, the greater their belief persistence,” which suggests that experiencing doubts at younger ages may be negatively associated with continuing to believe. Given this, we propose the following hypothesis:

Hypothesis 1: Level of doubt about belief in God in adolescence is positively associated with adoption of an atheistic worldview.

Social Location and Adopting an Atheistic Worldview

Berger (1970) proposed that social location shapes the stake people have in modernity, which affects their likelihood of holding nonreligious worldviews, including atheistic worldviews. Certain social locations make an atheistic worldview more or less costly, thereby serving as supports or barriers to adopting it, independent of an individual’s starting level of religious doubt. That is, we would expect individuals with the same initial level of religious doubt to have different probabilities of adopting an atheistic worldview depending on their social location.

In a study examining an individual’s likelihood of identifying as an atheist if they do not believe in God, Scheitle et al. (2019) identified a number of “supports and barriers” to atheist identification. Although the question we are examining in this study is about the adoption of an atheistic worldview—that is, moving toward saying that one does not believe in God—not atheist identification, many of the supports and barriers identified by Scheitle et al. are relevant to our question as well.

Supports to Adoption of an Atheistic Worldview

Both social theory and popular narratives have often seen education as an agent of secularization. Education, particularly higher education, is assumed to erode individual’s religious beliefs and commitment (Caplovitz and Sherrow 1977; Lehman 1972; Weber 2004; Wilson 1982). The mechanism for this is said to at least in part be intellectual, as individuals are exposed to a scientific worldview that undercuts religious claims about the world or are exposed to a pluralistic peer group that undercuts their religious certainty. Another mechanism is said to be more social in nature, as pursuing higher education might disrupt the networks and communities that support an individual’s religious life. Other research suggests that college may support religiosity through the numerous opportunities to participant in religious organizations (Cherry, DeBerg, and Porterfield 2001). Research on the effects of education on religion have been mixed. One study showed that education is not associated with religiosity among emerging adults (Arnett and Jensen 2002). Another showed that those who do not attend college have higher rates of declining religiosity (i.e., service attendance, importance of religion, and disaffiliation) compared with those currently attending four-year colleges (Uecker, Regnerus, and Vaaler 2007). Baker and Whitehead (2016) identified that education is associated with lower levels of religious service attendance and affiliation. However, as conservative religious traditions are less likely to seek higher educational attainment (Darnell and Sherkat 1997; Massengill 2008), the causal direction may be reverse. Schwadel (2016), using longitudinal data, found that obtaining a bachelor’s degree is associated with reduced frequency of prayer, religious certainty, and religious belief, although higher education did not appear to influence service attendance or a measure of personal religiosity. Other experimental research has demonstrated a negative association between education and religiosity (Dilmaghani 2019; Hungerman 2014).

Past research has focused primarily on associations between education and religious behaviors, affiliations, or identities with less focus on beliefs and worldviews. Mayrl and Uecker (2011) found increased doubts about religious beliefs among college students compared with nonstudents but also found that college students are more likely to retain beliefs in a personal God. Examining superempirical beliefs together (e.g., believing in God, demons, and angels), Hill (2011) found that they are negatively associated with attending and graduating from college. In terms of the education and religious belief literature specifically, the weight of the evidence falls on the side of a negative association between education and religious beliefs. Thus, we hypothesize as follows:

Hypothesis 2: Net of any adolescent religious doubt, education is positively associated with the adoption of an atheistic worldview.

Believing in God specifically or religion in general has been seen by some social theorists as a way to cope with deprivation, danger, fear, or uncertainty (Glock 1967; Niebuhr 1962; Weber 1968). This has led some to hypothesize that as an individual’s or a society’s level of existential security increases, their religiosity will decrease (Norris and Inglehart 2012). Although a sense of existential security can be influenced by a number of factors, socioeconomic well-being is often highlighted as a key predictor. Socioeconomic status is a social location that provides more security and thus a reduced demand for belief in a God. Household income is associated with lower levels of religiosity (Storm 2017). Given this, we might expect that individuals with greater income will be more likely to adopt an atheistic worldview.

Hypothesis 3: Net of any adolescent religious doubt, income is positively associated with the adoption of an atheistic worldview.

Although individuals can and do hold multiple social identities simultaneously (Stryker and Burke 2000), some identities and their associated beliefs and values have formal or informal proscriptions against others. This makes holding an identity that is forbidden by one’s community both socially and psychologically difficult. Religious communities have often proscribed certain identities from being held by their members, as well as the beliefs and behaviors that might be associated with those identities. Most obviously, many religious traditions and their communities expect members to identify with only their community. Many religious communities, especially those grounded in more conservative theological traditions, have and often still do proscribe adherents from holding nonheterosexual identities (Klarman 2013; Linneman 2003; Rimmerman 2002; Thumma 1991). This has led many sexually minoritized people to experience hostility from their religious communities (Barringer 2019; Beagan and Hattie 2015; Levy and Harr 2018; Schuck and Liddle 2001), which in turn has led sexually minoritized individuals to disaffiliate from religion at higher rates than their peers (Scheitle and Wolf 2017). Some sexually minoritized individuals report questioning, doubting, or moving away from their religious beliefs because of their sexuality (Sherry et al. 2010). Given such dynamics, we expect that sexually minoritized individuals will be more likely than those identifying as heterosexual to adopt an atheistic worldview.

Hypothesis 4: Net of any in adolescent religious doubt, being a sexually minoritized person is positively associated with the adoption of an atheistic worldview.

Barriers to Adoption of an Atheistic Worldview

Some social locations and identities may serve as resistance to adopting an atheistic worldview, including being politically conservative, racially or ethnically minoritized, or a woman.

Since the 1960s, public and policy debates surrounding moral issues such as gender, sexuality, reproductive rights, family structures, and similar topics strongly aligned political conservatism with religious conservatism (Putnam and Campbell 2010). This led those who were politically liberal to view religion as being opposed to their own values and identity, which has led such individuals to disavow religious identification (Hout and Fischer 2002, 2014; Sumerau and Cragun 2016). Indeed, individuals identifying as atheists are particularly likely to identify as politically liberal relative to even other religiously unaffiliated individuals (Schwadel 2020). Smith (2013) found that some atheist groups participate in liberal political/social activism as a means of forging collective identity. Individuals who are politically conservative, then, may be hesitant to adopt an atheistic worldview independent of other factors, as adopting such a worldview may indirectly associate them with a political identity and community that they view as aversive (Frimer, Skitka, and Motyl 2017).

Hypothesis 5: Net of any adolescent religious doubt, political conservatism is negatively associated with the adoption of an atheistic worldview.

Although an increasing proportion of the U.S. population is religiously unaffiliated, explicit rejection of a belief in God (i.e., holding an atheistic worldview) has grown much slower. This is likely in part because of the stigmatized nature of atheism in the United States. Indeed, a number of studies have shown that atheists are seen as morally and culturally suspect (Edgell, Gerteis, and Hartmann 2006; Edgell et al. 2016; Gervais 2014). Some have argued that those with more privilege will be more easily and willingly able to accept the costs of holding an atheistic worldview (Miller 2013). Moreover, certain social locations may make atheism more or less costly. Atheism is often perceived to be culturally masculine and White (Abbott et al. 2020; Schnabel et al. 2016). Women who adopt an atheistic worldview may be viewed as violators of gender norms (Schnabel et al. 2016), whereas Black people who adopt an atheistic worldview may be viewed as “race traitors” because of the centrality of the Black church in the Black community (Hutchinson 2011:20; Miller 2013). The connection between religiosity and ethnicity for some ethnic groups may also make adopting an atheistic worldview more costly (Jeung and Calvillo 2018; Kim 2011). For this reason, we might expect that women and racially or ethnically minoritized people will be less likely to adopt an atheistic worldview, all else being equal.

Hypothesis 6: Net of any adolescent religious doubt, being a racially or ethnically minoritized person is negatively associated with the adoption of an atheistic worldview.

Hypothesis 7: Net of any adolescent religious doubt, being a woman is negatively associated with the adoption of an atheistic worldview.

The religious composition of the United States varies significantly across its geography. This produces distinct cultures that can influence the individuals within them, whether or not they are actually involved in that culture or religion at all (Silk and Walsh 2008). Although one can point to a number of such geographically distinctive areas in the United States—from the historical and social dominance of Latter-day Saints in Utah to highly secular pockets of the Pacific Northwest—the South presents a particularly large and influential regional culture (Ellison, Burr, and McCall 2003; Lindsey 2005; Silk and Walsh 2008). Compared with all other regions in the United States, the South has the least non-Christian religious diversity (Lindsey 2005; Silk and Walsh 2008). Evangelical Protestantism dominates in the South, which makes it particularly salient when considering the adoption of an atheistic worldview. Nonreligious people, including atheists, experience higher rates of hostility and discrimination in the South than in other U.S. regions (Manning 2015; Scheitle and Corcoran 2018). Holding an atheistic worldview in the South is more costly, thus we expect that those residing in the South would be less likely to have adopted an atheistic worldview compared with other regions.

Hypothesis 8: Net of any adolescent religious doubt, residing in the South is negatively associated with the adoption of an atheistic worldview relative to other regions.

Data

The data for this study come from a survey fielded in late 2018 using the Gallup Panel. The Gallup Panel is a probability panel of nearly 100,000 U.S. adults who complete surveys online or by mail. Individuals are selected into the panel using random-digit dialing and address-based sampling methods. Because panelists are selected at random and with an equal and known probability, the Gallup Panel is representative of all U.S. households. More information about the Gallup Panel can be found in Gallup (2023).

For the survey used in the current study, a stratified sample of 29,345 panelists were selected to participate. This sample included oversamples of 752 individuals who had been preidentified as Muslim and 882 individuals who had been preidentified as Jewish. All those selected were provided a $2 prepaid cash incentive. Surveys were completed online or through the mail and were offered in both English and Spanish. A total of 13,270 individuals completed the survey. Responses rates for panel surveys, however, must take into account all stages of selection into the sample. The overall response rate for this particular survey (American Association for Public Opinion Research RR3) is 1.2 percent.

Measures

The survey used here was not designed specifically for the purposes of this study. Much of the instrument focused on individuals’ attitudes and behaviors concerning their work and how their faith (or lack of) influenced their work lives. This presents some limitations, as is the case when using any secondary data source. We use the data from this survey, though, because it happened to include a relatively unique measure of an individual’s belief in God at age 16.

Dependent Measure: Adopting an Atheistic Worldview

Individuals were asked, “Please indicate which statement below comes closest to expressing what you believe about God.” Six statements were provided: (1) “I don’t believe in God”; (2) “I don’t know whether there is a God, and I don’t believe there is any way to find out”; (3) “I do believe in a Higher Power of some kind”; (4) “I find myself believing in God some of the time, but not at others”; (5) “While I have doubts, I feel that I do believe in God”; and (6) “I know God really exists and I have no doubts about it.” This question mirrors ones found on other surveys, such as the GSS (Davern et al. 2021).

Respondents were also asked about their belief in God at age 16. Specifically, individuals were asked, “Please indicate which statement below comes closest to expressing what you believed about God at age 16.” Again, six slightly modified statements were offered: (1) “I didn’t believe in God”; (2) “I didn’t know whether there was a God and I didn’t believe there was any way to find out”; (3) “I didn’t believe in a personal God, but I did believe in a Higher Power of some kind”; (4) “I found myself believing in God some of the time, but not at others”; (5) “While I had doubts, I felt that I did believe in God”; and (6) “I knew God really existed and I had no doubts about it.”

We operationalized the adoption of an atheistic worldview as representing individuals who selected the first response (“I don’t believe in God”) for the current belief in God question but who did not select the first response (“I didn’t believe in God”) for the age 16 belief in God question. Moreover, because this study is focused on the adoption of an atheistic worldview, we exclude individuals from our analysis who said that they already had an atheistic worldview at age 16—that is, those who said that “I didn’t believe in God.” In other words, our analysis is limited to those who were eligible for or had the potential to adopt an atheistic worldview since age 16. This coding is summarized in Table 1.

Table 1.

Operationalization of Adopting an Atheistic Worldview Outcome Measure.

Age 16 Belief in God Response	Current Belief in God Response	Outcome Coding
1. I didn’t believe in God.		Not eligible for adopting an atheistic worldview; removed from sample
2. I didn’t know whether there was a God and I didn’t believe there was any way to find out.3. I didn’t believe in a personal God, but I did believe in a Higher Power of some kind,4. I found myself believing in God some of the time, but not at others.5. While I had doubts, I felt that I did believe in God.6. I knew God really existed and I had no doubts about it.	1. I do not believe in God.	(1) Adopted an atheistic worldview
	2. I don’t know whether there is a God, and I don’t believe there is any way to find out.3. I do believe in a Higher Power of some kind.4. I find myself believing in God some of the time, but not at others.5. While I have doubts, I feel that I do believe in God.6. I know God really exists and I have no doubts about it.	(0) Did not adopt an atheistic worldview

Predictors

We include a number of measures representing the hypotheses stated earlier to predict an individual’s odds of adopting an atheistic worldview. One of these is the individual’s certainty of belief in God at age 16, which we just described in constructing the dependent variable of interest. Also included are measures representing the individual’s education, their mother’s education, and their father’s education. This is measured from less than high school diploma (1) to postgraduate or professional degree, including master’s doctorate, medical, or law degree (8). We also include a measure representing the individual’s “total annual household income, before taxes.” This is measured from less than $12,000 (1) to more than $1,000,000 (11).

Political conservatism is measured with a question asking, “There is a lot of talk these days about liberals and conservatives. On a scale of extremely liberal to extremely conservative, where would you place yourself?” Responses ranged from “extremely liberal” (1) to “extremely conservative” (7).

Measures for race, ethnicity, and immigrant status are also included in our analyses. Race is coded into one of five categories: White (1), Black (2), Asian (3), other (4), and multiple races (5). Ethnicity is measured with a question asking, “Are you, yourself, of Hispanic origin or descent, such as Mexican, Puerto Rican, Cuban, or other Spanish background?” Responses were coded as no (0) or yes (1).

Gender is measured using a question asking, “What is your current gender?” Offered responses were “man” (1), “woman” (2), “prefer to self-describe” (3), and “prefer not to say” (4). The latter two were combined given the relatively low proportions within them. Sexual identity was measured with a question asking, “Which of the following describes you?” Offered responses were “heterosexual or straight” (1); “gay, lesbian, or homosexual” (2); “bisexual” (3); “prefer to self-describe” (4); “prefer not to say” (5); and “don’t know” (6). The latter were combined into a single category for the purposes of our analyses.

We include indicators representing the region that the individual resides in as defined by the census regions: Northeast, Midwest, South, or West. The South is used as the comparison category in our analyses.

We also control for age (measured continuously), individuals’ marital status, number of children, and immigrant status. Age and marriage has been found to be positively associated with certain belief in God (Sherkat 2008). Additionally, age and having one child are negatively associated with identifying as an atheist (Hunter 2010). Marital status is measured with six categories: married (1), living with a partner (2), single/never married (3), divorced (4), separated (5), and widowed (6). We also control for immigration status, as past research has suggested connections between immigration and religiosity (Connor 2008, 2009; Massey and Higgins 2011). Immigrant status is measured with a question asking, “Did you immigrate to the United States?” Responses are coded as no (0) or yes (1).³

Missing Data and Analysis Plan

As noted earlier, in this study we are predicting individuals’ odds of shifting from a nonatheistic worldview at age 16 to an atheistic worldview at the time of the survey. Because individuals who already had an atheistic worldview at age 16 are not at risk of or eligible for such a shift, these individuals (n = 1,096) are excluded from our analyses. In other words, our analysis focuses on the population of individuals who did not have an atheistic worldview at age 16. Missing values on other measures included in the analysis are imputed using chained equations (m = 10), such that the full sample within this target population is included in the analyses. This represents 12,174 individuals.

All analyses were conducted in Stata/SE 15.2 and use the data weights constructed by Gallup to account for the sample design and nonresponse patterns. We begin our analysis by examining descriptive statistics for our focal dependent variable and all independent variables. We then examine the rate of adopting an atheistic worldview across some key independent variables as an initial assessment of patterns. Finally, we estimate logistic regression models predicting our outcome of adopting an atheistic worldview.

Results

Table 2 presents descriptive statistics for all measures used in our analyses, including our focal independent variable indicating that an individual shifted from a nonatheistic worldview to an atheistic worldview. We see that 6 percent of individuals who said they did not have an atheistic worldview at age 16 now express such a worldview. We also see that, among those who did not have an atheistic worldview at 16, 9 percent said that at age 16, they “didn’t know whether there was a personal God and [they] didn’t believe there was anyway to find out.” Eight percent said that they “didn’t believe in a personal God, but [they] did believe in a Higher Power of some kind.” Twelve percent said that they believed “in God some of the time, but not at others.” Twenty-six percent said that although they “had doubts, [they] felt that [they] did believe in God” at age 16. Finally, 43 percent of individuals who did not have an atheistic worldview at age 16 said that they “knew God really existed” and they “had no doubts about it” at that age.

Table 2.

Descriptive Statistics.

	Proportion or Mean	SE	Minimum, Maximum
Adopted Atheistic worldview since age 16	.06	.002	0, 1
Certainty of belief in god at age 16
I didn’t know whether there was a personal God and I didn’t believe there was any way to find out	.09	.003	—
I didn’t believe in a personal God, but I did believe in a Higher Power of some kind	.08	.003	—
I found myself believing in God some of the time, but not at others	.12	.003	—
While I had doubts, I felt that I did believe in God	.26	.004	—
I knew God really existed and I had no doubts about it	.43	.005	—
Mother education	3.58	.022	1, 8
Father education	3.42	.024	1, 8
Respondent education	3.98	.022	1, 8
Income	5.48	.026	1, 11
Political conservatism	4.09	.018	1, 7
Race
White	.78	.004	—
Black	.12	.003	—
Asian	.01	.001	—
Other	.03	.002	—
Multiple	.04	.002	—
Hispanic/Latino (yes)	.16	.004	0, 1
Immigrant (yes)	.06	.003	0, 1
Age	48.75	.174	19, 101
Gender
Man	.46	.005	—
Woman	.51	.005	—
Other/self-describe	.01	.001	—
Marital status
Married	.58	.005
Living with partner	.05	.002
Single/never married	.20	.004
Separated	.01	.001
Divorced	.09	.003
Widowed	.04	.002
Number of children	.77	.01	0-7
Sexuality
Heterosexual	.89	.003	—
Gay, lesbian, homosexual	.02	.001	—
Bisexual	.03	.002	—
Other/self-describe	.03	.002	—
Region of residence
South	.37	.005	—
Northeast	.18	.004	—
Midwest	.21	.003	—
West	.23	.004	—

Source: Data are from the Faith at Work Survey (n = 12,174).

Note: Individuals who had an atheistic worldview at age 16 are excluded from analyses.

As noted in hypothesis 1, we expect that those who said they were less certain about their belief in God at age 16 should be more likely to have an atheistic worldview now, as this shift would represent a smaller shift in worldview than for those who had greater belief in God at age 16. Figure 1 presents an initial examination of this hypothesis alongside some other independent variables of interest. We see that 20.3 percent of individuals who said at age 16 that they did not know whether there was a God and did not believe there was any way to find out have adopted an atheistic worldview. As we look across the other categories of belief in God at age 16, we find that the rate of adopting an atheistic worldview steadily declines. Only 1.6 percent of those who said that at age 16 they “new God really existed” currently hold an atheistic worldview.

Figure 1.

Percentage adopting an atheistic worldview between age 16 and now, overall and by select characteristics.

Figure 1 also shows the rate of adopting an atheistic worldview by education, race, and gender. We see that the rate of adoption appears to be somewhat higher among those with a bachelor’s degree (8.7 percent) or graduate degree (8.7 percent) compared with those with a high school degree (3.3 percent) or less than a high school degree (3.4 percent). This provides some initial support for hypothesis 2. Looking at the racial groups, we see that the rate of adopting an atheistic worldview is higher among White (6.7 percent) individuals and lowest among Black (2.5 percent) and Asian (4.0 percent) individuals. This provides some initial support for hypothesis 6. Finally in Figure 1, we see the rate of adopting an atheistic worldview across gender identities. We see that the rate is highest among those identifying with a gender other than man or woman (11.7 percent). The rate among those identifying as men (7.3 percent) is also relatively high compared with the rate among individuals identifying as women (4.8 percent). Again, this provides some initial support for hypothesis 7.

Of course, the patterns seen in Figure 1 do not account for the overlapping associations between these and other variables. For instance, the apparent racial differences in the rate of adopting an atheistic worldview in Figure 1 could be a product of varying levels of certainty in the belief in God at age 16 across the racial groups. To assess whether the rate of adopting an atheistic worldview is higher in some groups, net of other differences between groups, we need to examine findings from our logistic regression models. The results of these models are shown in Table 3.

Table 3.

Logistic Regression Models Predicting the Adoption of an Atheistic Worldview between Age 16 and Now.

Predictors	Outcome: Adopted Atheistic Worldview since Age 16
	Model 1		Model 2		Model 3
	Odds Ratio	95% Confidence Interval	Odds Ratio	95% Confidence Interval	Odds Ratio	95% Confidence Interval
Belief in God at age 16
I didn’t know whether there was a personal God and I didn’t believe there was any way to find out (reference)	—	—	—	—	—	—
I didn’t believe in a personal God, but I did believe in a Higher Power of some kind	.51**	(.39–.67)	—	—	.61**	(.46–.82)
I found myself believing in God some of the time, but not at others	.39**	(.30–.50)	—	—	.50**	(.38–.67)
While I had doubts, I felt that I did believe in God	.20**	(.16–.26)	—	—	.34**	(.26–.45)
I knew God really existed and I had no doubts about it	.06**	(.04–.08)	—	—	.15**	(.11–.22)
Supports to adopting an atheistic worldview
Mother education	—	—	.95	(.89–1.01)	.95	(.90–1.01)
Father education	—	—	1.01	(.97–1.07)	1.01	(.96–1.06)
Respondent education	—	—	1.03	(.98–1.09)	1.03	(.98–1.09)
Income	—	—	1.06**	(1.01–1.11)	1.05*	(1.01–1.11)
Sexuality
Heterosexual (reference)	—	—	—	—	—	—
Gay, lesbian, homosexual	—	—	.80	(.52–1.24)	.77	(.49–1.20)
Bisexual	—	—	1.84**	(1.26–2.68)	1.71**	(1.16–2.50)
Other/self-describe	—	—	.99	(.58–1.68)	1.00	(.57–1.73)
Barriers to adopting and atheistic worldview
Political conservatism	—	—	.50**	(.47–.54)	.54**	(.50–.58)
Race
White (reference)	—	—	—	—	—	—
Black	—	—	.26**	(.17–.40)	.36**	(.24–.55)
Asian	—	—	.24**	(.10–.59)	.25**	(.10–.63)
Other	—	—	.92	(.47–1.80)	1.12	(.56–2.23)
Multiple	—	—	.65	(.42–1.01)	.71	(.46–1.10)
Hispanic\Latino (yes)	—	—	.50**	(.35–.70)	.56**	(.39–.80)
Gender
Man (reference)	—	—	—	—	—	—
Woman	—	—	.42**	(.34–.50)	.46**	(.37–.56)
Other/self-describe	—	—	.90	(.43–1.87)	.87	(.42–1.80)
Region of residence
South (reference)	—	—	—	—	—	—
Northeast	—	—	1.02	(.78–1.34)	.95	(.72–1.26)
Midwest	—	—	.91	(.70–1.17)	.92	(.71–1.19)
West	—	—	1.30*	(1.02–1.67)	1.24	(.96–1.59)
Controls
Marital status
Married (reference)	—	—	—	—	—	—
Living with partner	—	—	1.03	(.69–1.54)	.90	(.59–1.39)
Single/never married	—	—	1.00	(.74–1.35)	.95	(.70–1.31)
Separated	—	—	.68	(.26–1.77)	.64	(.24–1.73)
Divorced	—	—	1.48*	(1.03–2.14)	1.34	(.92–1.95)
Widowed	—	—	.98	(.51–1.87)	1.06	(.58–1.95)
Immigrant (yes)	—	—	1.35	(.92–1.99)	1.40	(.94–2.07)
Age	—	—	.98**	(.97–.99)	.98**	(.97–.99)
Number of children	—	—	.96	(.86–1.06)	.98	(.88–1.09)
n	12,174		12,174		12,174

Source: Data are from the Faith at Work Survey.

Note: Individuals who had an atheistic worldview at age 16 are excluded from analyses.

p < 0.05. **p < 0.01.

Model 1 in Table 3 predicts the adoption of an atheistic worldview simply as a function of an individual’s certainty of belief in God at age 16. Remember that our findings are presented in the form of odds ratios (ORs), so coefficients above 1 represent an increase in the odds of adopting an atheistic worldview, and coefficients below 1 represent a decrease in the odds of adopting an atheistic worldview. We see that relative to those who say that at age 16 they did not know whether there was a God and did not believe there was a way to find out, all other belief categories have lower odds of adopting an atheistic worldview. Looking at the ORs, we can see that they decrease in size as we move down the levels of belief certainty. This reiterates what was seen in Figure 1, but we can also see from the nonoverlapping 95 percent confidence intervals that many of the adjacent categories are significantly different from each other. This means that these categories are not simply different from the comparison category but also largely different from each other.

Model 2 includes all of our measures other than the measure for certainty of belief at age 16. We see that net of the other measures, there is no significant association between the respondent’s education and the odds of adopting an atheistic worldview. Nor is there a significant independent association between either measure of parental education and the odds of adopting an atheistic worldview since age 16. We do find, though, that income is significantly associated with an increase (OR = 1.06, p < .01) in the odds of having adopted an atheistic worldview since age 16. Political conservatism, on the other hand, is significantly associated with a reduction in the odds of having adopted an atheistic worldview since age 16 (OR = 0.50, p < .01).

Turning to the indicators for race, ethnicity, and immigration status, we see that relative to individuals identifying as White, individuals identifying as Black (OR = 0.26, p < .01) or Asian (OR = 0.24, p < .01) have reduced odds of having adopted an atheistic worldview since age 16 independent of the other measures in the model. The model also shows that individuals identifying as Hispanic or Latino have significantly reduced odds (OR = 0.50, p < .01) of having adopted an atheistic worldview since age 16. We do not find a significant difference in the odds when comparing immigrants with nonimmigrants.

Looking at the indicator for age, we see that older individuals have significantly reduced odds (OR = 0.98, p < .01) of adopting an atheistic worldview since age 16. Individuals identifying as women, relative to those identifying as men, also have significantly reduced odds (OR = 0.42, p < .01) of having adopted an atheistic worldview since age 16. The indicators for marital status show that relative to those currently married, individuals who are currently divorced have greater odds (OR = 4.48, p < .05) of having adopted an atheistic worldview since age 16. We do not find a significant association between our outcome and an individual’s number of children. We do find, though, that individuals identifying as bisexual have greater odds (OR = 1.84, p < .01) of having adopted an atheistic worldview since age 16 compared with those identifying as straight or heterosexual. The other sexuality categories do not significantly differ from this comparison group. Finally, looking at the indicators for an individual’s region of residence, we find that individuals living in the West have significantly greater odds (OR = 1.30, p < .05) of having adopted an atheistic worldview compared with individuals living in the South.

Model 3 includes all of our measures, including our measure of belief in God at age 16. This model allows us to assess whether any of the patterns seen in model 2 become insignificant once we control for an individual’s certainty of belief at age 16. Comparing the findings in model 3 with those in model 2 reveals few notable differences. We do see that the previously significant difference in the odds of adopting an atheistic worldview between individuals residing in the West and those residing in the South becomes nonsignificant in model 3. This suggests that this regional effect is a function of differences in belief in God at age 16 between the regions. Similarly, we see that the significant difference in the odds of adopting an atheistic worldview between those currently married and those currently divorced in model 2 becomes nonsignificant once we account for individuals’ level of belief in God at age 16.

Most of the findings in model 3, though, mirror those seen in model 2. That is, even after accounting for differences in age 16 belief in God, individuals with higher incomes and those identifying as bisexual have greater odds of adopting an atheistic worldview relative to those with lower incomes and those identifying as heterosexual. Political conservatives, Black and Asian individuals, women, and older individuals have lower odds of adopting an atheistic worldview relative to liberals, White individuals, men, and younger individuals. We discuss these findings further below.

Discussion

Although assessing changes in individuals’ religious identities, beliefs, and behaviors is a central concern for the social scientific study of religion, it is often difficult to do so given the practical challenges of collecting panel data. There is a long tradition of using retrospective survey measures of religious affiliation to examine dynamics of switching of denominations and traditions, including the dropping out of religion (Adamczyk 2005; Hadaway and Marler 1996; Loveland 2003; Scheitle and Adamczyk 2010). Retrospective measures of religious beliefs, however, are less common (see Davern et al. 2021 for an exception).

Our expectations for the predictors of adopting an atheistic worldview, derived from parallel theory and research on the supports and barriers to adopting an atheistic identity, were largely supported. Racially and ethnically minoritized people largely had lower odds of adopting an atheistic worldview relative to White individuals. Women also had lower odds relative to men, and political conservatives had lower odds of adopting an atheistic worldview relative to liberals. As we argued, these groups face social and cultural barriers to adopting an atheistic worldview. On the other hand, individuals with greater income had greater odds of adopting an atheistic worldview, which we argue is a function of the greater existential security that comes with more financial resources.

One hypothesis that was not supported was our expectation that individuals with more education would have greater odds of adopting an atheistic worldview. We examined this issue further and found that in reduced models an individual’s education is associated with greater odds of adopting an atheistic worldview. This independent association, however, is greatly weakened to the point of being statistically insignificant when we account for an individual’s political conservatism. This may be due to a selection effect whereby fewer political conservatives obtain higher education or due to higher education politically liberalizing students. Data from the IDEALS study suggest stronger support for the former (Mayhew and Singer 2020).

Although perhaps not surprising, our findings regarding the association between prior strength of belief in God on the odds of adopting an atheistic worldview is worth a brief discussion. As seen in the analysis, the stronger an individual’s reported belief in God at age 16 the lower the odds are that they will have adopted an atheistic worldview. Indeed, the odds are fairly consistently reduced across each level of age 16 belief. This suggests that the adoption of an atheistic worldview is often incremental, with those weaker belief levels being the most likely to ultimately adopt an atheistic worldview. This could have intergenerational consequences, as the percentage of U.S. adults saying that they have the strongest level of belief in God has been slowly declining, while those expressing somewhat lower levels of belief have increased (Davern et al. 2021; Jones 2022).

This study, of course, has its limitations. Maybe most apparent is that our measure of adopting an atheistic worldview is retrospective and cross-sectional in nature. It would be ideal to have panel data following individual’s belief in God over many years, but such data are challenging to collect. On the other hand, this limitation could also be seen as an opportunity to consider the meaning and utility of such retrospective measures for future research. The obvious concern of such measures is that they are not reliable or valid. Concerning reliability, it is possible that adults’ responses concerning their belief in God at age 16 will not match what they would have actually said at age 16 if parallel measures were available. Regarding validity, it is possible that those responses would not match because individuals are not reporting their actual past belief in God but are instead reporting some idealized narrative of their biography and identity, possibly to better fit with their current sense of self. Individuals who did not strongly believe in God at age 16, for instance, but who now do strongly believe in God might project that confidence into the past to better ground their identity. Or, someone who did not actually have an atheistic worldview at age 16 but who does currently might report that they have always held such a worldview for the same reason.

On the other hand, past research has suggested that retrospective measures could produce more valid responses of change in cases where an individual’s sense of the response scale is altered by time and experience, a process called response shift bias (Howard 1980). Consider, for example, an individual who says at 16 that they hold a certain level of belief (or disbelief) in God. As they age, however, their perception or understanding of the response scale offered in a panel survey may change. They might, for instance, have previously responded that they had no doubts about God’s existence, for instance, only to realize latter in their life that they had more doubts than they realized. Panel measures may have assessed this individual as having had no change in their level of belief (i.e., strong belief then and now), while retrospective measures may have better captured such change (i.e., weaker belief then and stronger belief now). That is, the retrospective measure allows respondents to reflect on their current and past belief in God with the same understanding of the question and measurement scale, such that changes in the responses are not due to changes in subjective interpretation of the questions and responses. However, changes in responses could still reflect retrospective bias. Thus, there are advantages and disadvantages to both longitudinal and retrospective measures examining changes in belief.

Although past research has used retrospective measures of affiliation and religious behaviors (e.g., religious service attendance and prayer), such research is generally lacking for retrospective measures of beliefs and worldviews. Given the cost of longitudinal surveys, retrospective measures represent a cost-efficient method for examining religious change. This is one of the first studies we know of to use a retrospective measure of belief in God for a representative U.S. adult sample to examine adopting an atheistic worldview. Future research would benefit from including retrospective measures on other forms of religious (e.g., angels, demons, spirits), and parareligious (e.g., luck, parapsychology, magic) beliefs, which would be informative for understanding factors associated with declining belief in God as well as changes in religious identification and belonging.

Finally, we note that this study was focused on the adoption of an atheistic worldview, but this is only one potential outcome from changes in belief in God. The most obvious alternative outcome would be the adoption of a theistic worldview. Such other outcomes, though, are likely shaped in part by distinct social processes and would produce their own expectations. Given this, it was not possible to consider every possible change in this single study, but this is clearly an area for future studies to examine and assess the similarities and differences between the findings presented here.

Supplemental Material

sj-docx-1-srd-10.1177_23780231231204850 – Supplemental material for Predictors of Adopting an Atheistic Worldview: An Analysis of Survey Data Containing a Retrospective Measure of Belief in God

Supplemental material, sj-docx-1-srd-10.1177_23780231231204850 for Predictors of Adopting an Atheistic Worldview: An Analysis of Survey Data Containing a Retrospective Measure of Belief in God by Christopher P. Scheitle and Katie E. Corcoran in Socius

Footnotes

We thank Elaine Howard Ecklund and Denise Daniels for providing access to the data used in this study.

Funding

The author(s) disclosed receipt of the following financial support for the research,authorship,and/or publication of this article: Funding for this study was provided by the Explaining Atheism program led by Queen’s University Belfast in collaboration with the University of Kent,Brunel University London,Coventry University,St. Mary’s University Twickenham London,and the University of Notre Dame Australia. The program has been made possible through the support of grant 61928 from the John Templeton Foundation.

ORCID iDs

Christopher P. Scheitle

Katie E. Corcoran

Supplemental Material

Supplemental material for this article is available online.

1

Descriptively using raw unweighted GSS data,the percentage of U.S. adults who indicated “I don’t believe in God now,and I never have” and “I don’t believe in God now,but I used to” increased from 2004 to 2008 and then again in 2018.

2

In 1991,1998,2004,2008,and 2018,the GSS asked respondents,“Which best describes your beliefs about God?” with the following response choices: “I don’t believe in God now and I never have,” “I don’t believe in God now,but I used to,” “I believe in God now,but I didn’t used to,” and “I believe in God now,and I always have” (

). There are a couple of reasons why we believe our measures differ and are likely stronger than the GSS measure. First,the GSS measure of change in belief in God collapses differing levels of certainty of belief in God into currently or previously believing in God. Individuals who were agnostic or had weaker levels of belief in God,for instance,are forced into either the “used to believe in God” or “never believed in God” categories. In short,the single-item GSS measure of change in belief in God loses a great deal of variation and nuance. Asking for six-level certainty of belief in God at age 16 and six-level certainty of belief now maintains much more of this nuance. Second,the GSS measure of change in belief in God does not specify a time point. As a result,if someone says that they “don’t believe in God,but I used to,” it is unclear what time period the phrase “used to” refers to or what came before that time period.

3

These data do not contain a measure of individuals’ religious affiliation or identity at age 16. However,we do not believe that the absence of these indicators in our analysis has an impact on our findings. Religious tradition at 16 is likely to influence one’s certainty of belief in God,and we have a measure of certainty of belief in God at age 16. In other words,certainty of belief in God at age 16 is likely to mediate the effect of religious tradition at 16. Although adding certainty of belief in God at age 16 to a model would undoubtedly affect the findings for religious tradition indicators,we would not expect adding religious tradition at age 16 to affect the findings for certainty of belief in God at 16.

Author Biographies

Christopher P. Scheitle (PhD,Penn State University,2008) is an associate professor of sociology at West Virginia University. His work has been funded by the National Science Foundation,the John Templeton Foundation,and other agencies. His newest book is The Faithful Scientist: Experiences of Anti-religious Bias in Scientific Training (New York University Press,2023).

Katie E. Corcoran (PhD,University of Washington,2012) is a professor of sociology at West Virginia University. Her work explores religious belief,identity,and behavior focusing on health outcomes and the relationship between religion and science.

References

Abbott

Dena M.

Mollen

Debra

Mercier

Caitlin

Anaya

Elyxcus J.

Rukus

Victoria A.

2020. “‘Isn’t Atheism a White Thing?’ Centering the Voices of Atheists of Color.” Journal of Counseling Psychology 67(3):275.

Adamczyk

Amy

. 2005. “Religious Switching: Does Parents’ Education Matter?” Pp. 51–70 in Research in the Social Scientific Study of Religion, Vol. 15, edited by Piedmont

R. L.

Moberg

D. O.

Leiden, the Netherlands: Brill.

Arnett

Jeffrey Jensen

Jensen

Lene Arnett

. 2002. “A Congregation of One: Individualized Religious Beliefs among Emerging Adults.” Journal of Adolescent Research 17(5):451–67.

Bainbridge

William Sims

. 2005. “Atheism.” Interdisciplinary Journal of Research on Religion 1(2):1–24.

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.

Bankston

Carl L.

III . 2002. “Rationality, Choice and the Religious Economy: The Problem of Belief.” Review of Religious Research 43(4):311–25.

Barringer

Mandi N.

2019. “Lesbian, Gay, and Bisexual Individuals’ Perceptions of American Religious Traditions.” Journal of Homosexuality 67(9):1173–96.

Beagan

Brenda L.

Hattie

Brenda

. 2015. “Religion, Spirituality, and LGBTQ Identity Integration.” Journal of LGBT Issues in Counseling 9(2):92–117.

Beit-Hallahmi

Benjamin

. 2007. “Atheists: A Psychological Profile.” Pp. 300–17 in The Cambridge Companion to Atheism, edited by Martin

New York: Cambridge University Press.

10.

Berger

Peter L.

1967. The Sacred Canopy: Elements of a Sociological Theory of Religion. New York: Doubleday.

11.

Berger

Peter L.

1970. A Rumor of Angels: Modern Society and the Rediscovery of the Supernatural. New York: Anchor.

12.

Brinkerhoff

Merlin B.

Mackie

Marlene M.

1993. “Casting Off the Bonds of Organized Religion: A Religious Careers Approach to the Study of Apostasy.” Review of Religious Research 34(3):235–58.

13.

Burge

Ryan P.

2020. “How Many ‘Nones’ Are There? Explaining the Discrepancies in Survey Estimates.” Review of Religious Research 62(1):173–90.

14.

Caplovitz

David

Sherrow

Fred

. 1977. The Religious Drop-Outs: Apostasy among College Graduates. Beverly Hills, CA: Sage.

15.

Cherry

Conrad

DeBerg

Betty A.

Porterfield

Amanda

. 2001. Religion on Campus. Chapel Hill: University of North Carolina Press.

16.

Connor

Phillip

. 2008. “Increase or Decrease? The Impact of the International Migratory Event on Immigrant Religious Participation.” Journal for the Scientific Study of Religion 47(2):243–57.

17.

Connor

Phillip

. 2009. “International Migration and Religious Participation: The Mediating Impact of Individual and Contextual Effects.” Sociological Forum 24(4):779–803.

18.

Corcoran

Katie E.

2013. “Divine Exchanges: Applying Social Exchange Theory to Religious Behavior.” Rationality and Society 25(3):335–69.

19.

Darnell

Alfred

Sherkat

Darren E.

1997. “The Impact of Protestant Fundamentalism on Educational Attainment.” American Sociological Review 62(2):306–15. doi:10.2307/2657306.

20.

Davern

Michael

Bautista

Rene

Freese

Jeremy

Morgan

Stephen L.

Smith

Tom W.

2021. “General Social Surveys, 1972–2021 Cross-Section [Machine-Readable Data File, 68,846 Cases].” Chicago: NORC at the University of Chicago.

21.

Dilmaghani

Maryam

. 2019. “Education and Religious Decline: Evidence from the Canadian Compulsory Schooling Laws.” Education Economics 27(3):287–307.

22.

Durkin

John T.

Jr. Greeley

Andrew M.

1991. “A Model of Choice Under Uncertainty: On Responding Rationally to the Nonrational.” Rationality and Society 3(2):178–96.

23.

Edgell

Penny

Gerteis

Joseph

Hartmann

Douglas

. 2006. “Atheists as ‘Other’: Moral Boundaries and Cultural Membership in American Society.” American Sociological Review 71(2):211–34.

24.

Edgell

Penny

Hartmann

Douglas

Stewart

Evan

Gerteis

Joseph

. 2016. “Atheists and Other Cultural Outsiders: Moral Boundaries and the Non-religious in the United States.” Social Forces 95(2):607–38.

25.

Ellison

Christopher G.

Burr

Jeffrey A.

McCall

Patricia L.

2003. “The Enduring Puzzle of Southern Homicide: Is Regional Religious Culture the Missing Piece?” Homicide Studies 7(4):326–52.

26.

Frimer

Jeremy A.

Skitka

Linda J.

Motyl

Matt

. 2017. “Liberals and Conservatives Are Similarly Motivated to Avoid Exposure to One Another’s Opinions.” Journal of Experimental Social Psychology 72(1):1–12.

27.

Gallup. 2023. “How Does the Gallup Panel Work?” Retrieved July 7, 2023. https://www.gallup.com/174158/gallup-panel-methodology.aspx.

28.

Gervais

Will M.

2014. “Everything Is Permitted? People Intuitively Judge Immorality as Representative of Atheists.” PLoS One 9(4):e92302.

29.

Gervais

Will M.

Najle

Maxine B.

Caluori

Nava

. 2021. “The Origins of Religious Disbelief: A Dual Inheritance Approach.” Social Psychological and Personality Science 12(7):1369–79.

30.

Glock

Charles Y.

1967. To Comfort and to Challenge. Berkeley: University of California Press.

31.

Hadaway

C. Kirk

Marler

Penny Long

. 1996. “The Problem with Father as Proxy: Denominational Switching and Religious Change, 1965–1988.” Journal for the Scientific Study of Religion 35(2):156–64.

32.

Hill

Jonathan P.

2011. “Faith and Understanding: Specifying the Impact of Higher Education on Religious Belief.” Journal for the Scientific Study of Religion 50(3):533–51.

33.

Hout

Michael

Fischer

Claude S.

2002. “Why More Americans Have No Religious Preference: Politics and Generations.” American Sociological Review 67(2):165–90.

34.

Hout

Michael

Fischer

Claude S.

2014. “Explaining Why More Americans Have No Religious Preference: Political Backlash and Generational Succession, 1987–2012.” Sociological Science 1:423–47.

35.

Howard

George S.

1980. “Response-Shift Bias: A Problem in Evaluating Interventions with Pre/Post Self-Reports.” Evaluation Review 4(1):93–106.

36.

Hungerman

Daniel M.

2014. “The Effect of Education on Religion: Evidence from Compulsory Schooling Laws.” Journal of Economic Behavior & Organization 104(1):52–63.

37.

Hunsberger

Bruce E.

Altemeyer

Bob

. 2006. Atheists: A Groundbreaking Study of America’s Nonbelievers. Amherst, NY: Prometheus.

38.

Hunter

Laura A.

2010. “Explaining Atheism: Testing the Secondary Compensator Model and Proposing an Alternative.” Interdisciplinary Journal of Research on Religion 6:6.

39.

Hutchinson

Sikivu

. 2011. Moral Combat: Black Atheists, Gender Politics, and the Value Wars. Los Angeles: Infidel.

40.

Jeung

Russell

Calvillo

Jonathan E.

2018. “Race, Immigration, Ethnicity, and Religion in America.” In Oxford Encyclopedia of Religion in America, edited by Corrigan

New York: Oxford University Press.

41.

Johnson

Kathryn A.

Hill

Eric D.

Cohen

Adam B.

2011. “Integrating the Study of Culture and Religion: Toward a Psychology of Worldview.” Social and Personality Psychology Compass 5(3):137–52.

42.

Jones

Jeffrey M.

2022. “Belief in God in U.S. Dips to 81%, a New Low.” Retrieved July 7, 2023. https://news.gallup.com/poll/393737/belief-god-dips-new-low.aspx.

43.

Kim

Rebecca Y.

2011. “Religion and Ethnicity: Theoretical Connections.” Religions 2(3):312–29.

44.

Klarman

Michael

. 2013. From the Closet to the Altar: Courts, Backlash, and the Struggle for Same-Sex Marriage. New York: Oxford University Press.

45.

Langston

Joseph

. 2014. “Explaining Atheism: Testing Hunter’s Durkheimian Theory.” Interdisciplinary Journal of Research on Religion 10:Article 10.

46.

LeDrew

Stephen

. 2013. “Discovering Atheism: Heterogeneity in Trajectories to Atheist Identity and Activism.” Sociology of Religion 74(4):431–53.

47.

Lehman

Edward C.

1972. “The Scholarly Perspective and Religious Commitment.” Sociological Analysis 33(4):199–216.

48.

Levy

Denise L.

Harr

Jennifer

. 2018. “‘I Never Felt Like There Was a Place for Me:’ Experiences of Bisexual and Pansexual Individuals with a Christian Upbringing.” Journal of Bisexuality 18(2):186–205.

49.

Lindsey

William

. 2005. “Religious Affiliation in the Southern Crossroads and the Nation—Religious Self-Identification and Adherents Claimed by Religious Groups, National and Regional Comparisons.” Pp. 23–26 in Religion and Public Life in the Southern Crossroads: Showdown States, edited by Lindsey

C. W.

Silk

Walnut Creek, CA: AltaMira.

50.

Linneman

Thomas J.

2003. Weathering Change: Gays and Lesbians, Christian Conservatives, and Everyday Hostilities. New York: New York University Press.

51.

Loveland

Matthew T.

2003. “Religious Switching: Preference Development, Maintenance, and Change.” Journal for the Scientific Study of Religion 42(1):147–57.

52.

Manning

Christel

. 2015. Losing Our Religion: How Unaffiliated Parents Are Raising Their Children. New York: New York University Press.

53.

Massengill

Rebekah Peeples

. 2008. “Educational Attainment and Cohort Change among Conservative Protestants, 1972–2004.” Journal for the Scientific Study of Religion 47(4):545–62.

54.

Massey

Douglas S.

Higgins

Monica Espinoza

. 2011. “The Effect of Immigration on Religious Belief and Practice: A Theologizing or Alienating Experience?” Social Science Research 40(5):1371–89.

55.

Mayhew

Matthew

Singer

Kevin

. 2020. “Do Conservative Students Turn Liberal When They Go to College? It’s Complicated.” The Independent. Retrieved July 13, 2023. https://www.independent.co.uk/voices/conservative-college-students-liberal-defund-trump-biden-election-b423824.html.

56.

Mayrl

Damon

Uecker

Jeremy E.

2011. “Higher Education and Religious Liberalization among Young Adults.” Social Forces 90(1):181–208.

57.

Miller

Ashley F.

2013. “The Non-religious Patriarchy: Why Losing Religion Has Not Meant Losing White Male Dominance.” CrossCurrents 63(2):211–26.

58.

Mitchell

Travis

. 2019. “In U.S., Decline of Christianity Continues at Rapid Pace.” Pew Research Center. Retrieved July 7, 2023. https://www.pewresearch.org/religion/2019/10/17/in-u-s-decline-of-christianity-continues-at-rapid-pace/.

59.

Nelsen

Hart M.

1981. “Religious Conformity in an Age of Disbelief: Contextual Effects of Time, Denomination, and Family Processes Upon Church Decline and Apostasy.” American Sociological Review 46(5):632–40.

60.

Niebuhr

Helmut Richard

. 1962. The Social Sources of Denominationalism. Cleveland, OH: Meridian.

61.

Norris

Pippa

Inglehart

Ronald

. 2012. Sacred and Secular: Religion and Politics Worldwide. 2nd ed. Cambridge, UK: Cambridge University Press.

62.

Perry

Samuel L.

Longest

Kyle C.

2019. “Examining the Impact of Religious Initiation Rites on Religiosity and Disaffiliation over Time.” Journal for the Scientific Study of Religion 58(4):891–904.

63.

Putnam

Robert D.

Campbell

David E.

2010. American Grace: How Religion Divides and Unites Us. New York: Simon & Schuster.

64.

Rimmerman

Craig A.

2002. From Identity to Politics: The Lesbian and Gay Movements in the United States. Philadelphia: Temple University Press.

65.

Scheitle

Christopher P.

Adamczyk

Amy

. 2010. “High-Cost Religion, Religious Switching, and Health.” Journal of Health and Social Behavior 51(3):325–42.

66.

Scheitle

Christopher P.

Corcoran

Katie E.

2018. “Religious Tradition and Workplace Religious Discrimination: The Moderating Effects of Regional Context.” Social Currents 5(3):283–300.

67.

Scheitle

Christopher P.

Corcoran

Katie E.

Hudnall

Erin B.

2019. “Adopting a Stigmatized Label: Social Determinants of Identifying as an Atheist beyond Disbelief.” Social Forces 97(4):1731–55.

68.

Scheitle

Christopher P.

Wolf

Julia Kay

. 2017. “The Religious Origins and Destinations of Individuals Identifying as a Sexual Minority.” Sexuality & Culture 21(3):719–40.

69.

Schnabel

Landon

Facciani

Matthew

Sincoff-Yedid

Ariel

Fazzino

Lori

. 2016. “Gender and Atheism: Paradoxes, Contradictions, and an Agenda for Future Research.” Pp. 75–97 in Annual Review of the Sociology of Religion, edited by Cipriani

Garelli

Leiden, the Netherlands: Brill.

70.

Schuck

Kelly D.

Liddle

Becky J.

2001. “Religious Conflicts Experienced by Lesbian, Gay, and Bisexual Individuals.” Journal of Gay & Lesbian Psychotherapy 5(2):63–82.

71.

Schwadel

Philip

. 2016. “Does Higher Education Cause Religious Decline? A Longitudinal Analysis of the within- and between-Person Effects of Higher Education on Religiosity.” Sociological Quarterly 57(4):759–86.

72.

Schwadel

Philip

. 2020. “The Politics of Religious Nones.” Journal for the Scientific Study of Religion 59(1):180–89.

73.

Sherkat

Darren E.

2008. “Beyond Belief: Atheism, Agnosticism, and Theistic Certainty in the United States.” Sociological Spectrum 28(5):438–59.

74.

Sherry

Alissa

Adelman

Andrew

Whilde

Margaret R.

Quick

Daniel

. 2010. “Competing Selves: Negotiating the Intersection of Spiritual and Sexual Identities.” Professional Psychology: Research and Practice 41(2):112.

75.

Silk

Mark

Walsh

Andrew

. 2008. One Nation, Divisible: How Regional Religious Differences Shape American Politics. Lanham, MD: Rowman & Littlefield.

76.

Sire

James W.

2015. Naming the Elephant: Worldview as a Concept. Downers Grove, IL: InterVarsity.

77.

Smith

Christian

Emerson

Michael O.

1998. American Evangelicalism: Embattled and Thriving. Chicago: University of Chicago Press.

78.

Smith

Jesse M.

2011. “Becoming an Atheist in America: Constructing Identity and Meaning from the Rejection of Theism.” Sociology of Religion 72(2):215–37.

79.

Smith

Jesse M.

2013. “Creating a Godless Community: The Collective Identity Work of Contemporary American Atheists.” Journal for the Scientific Study of Religion 52(1):80–99.

80.

Stark

Rodney

Finke

Roger

. 2000. Acts of Faith: Explaining the Human Side of Religion. Berkeley: University of California Press.

81.

Stewart

Evan M.

2016. “The True (Non) Believer? Atheists and the Atheistic in the United States.” Annual Review of the Sociology of Religion 7:137–60.

82.

Storm

Ingrid

. 2017. “Does Economic Insecurity Predict Religiosity? Evidence from the European Social Survey 2002–2014.” Sociology of Religion 78(2):146–72.

83.

Stryker

Sheldon

Burke

Peter J.

2000. “The Past, Present, and Future of an Identity Theory.” Social Psychology Quarterly 63(4):284–97.

84.

Sumerau

Jason E.

Cragun

Ryan T.

2016. “‘I Think Some People Need Religion’: The Social Construction of Nonreligious Moral Identities.” Sociology of Religion 77(4):386–407.

85.

Thumma

Scott L.

1991. “Negotiating a Religious Identity: The Case of the Gay Evangelical.” Sociological Analysis 52:333–47.

86.

Uecker

Jeremy E.

Regnerus

Mark D.

Vaaler

Margaret L.

2007. “Losing My Religion: The Social Sources of Religious Decline in Early Adulthood.” Social Forces 85(4):1667–92.

87.

Vargas

Nicholas

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

88.

Weber

Max

. 1968. Economy and Society. Edited by Roth

Wittich

New York: Bedminster.

89.

Weber

Max

. 2004. “Science as a Vocation.” Pp. 1–31 in Max Weber: The Vocation Lectures, edited by Owen

Strong

T. B.

Indianapolis, IN: Hackett.

90.

Wellman

James K.

Jr.

2008. Evangelical vs. Liberal: The Clash of Christian Cultures in the Pacific Northwest. New York: Oxford University Press.

91.

Wiertz

Dingeman

Lim

Chaeyoon

. 2021. “The Rise of the Nones across the United States, 1973 to 2018: State-Level Trends of Religious Affiliation and Participation in the General Social Survey.” Sociological Science 8:429–54.

92.

Wilson

Bryan

. 1982. Religion in Sociological Perspective. Oxford, UK: Oxford University Press.

93.

Wollschleger

Jason

. 2021. “Religious Certainty as Social Capital.” Review of Religious Research 63:325–42.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.01 MB