Correlates of Christian Religious Identification and De-identification among Sexual and Gender Minorities: A U.S. Probability Sample

Abstract

Using a U.S. nationally representative sample of 1,529 sexual and gender minorities (SGMs), we examined the demographic and developmental correlates of Christian religious de-identification. We found that SGMs who were older, Black, cisgender men, and/or lived in the American South were more likely to identify as Christian in adulthood, relative to other SGMs. Those who were never Christian reported being more out to family and friends at earlier ages than those who were raised Christian. SGMs who were raised Christian, but did not identify as Christian in adulthood reported, more adverse childhood experiences and bullying than other SGMs. Sexual minorities who identified as Christian in adulthood reported more internalized homophobia in adulthood than other sexual minorities. Finally, sexual minorities who were raised Christian, regardless of their adulthood religious identification, reported experiencing more stigma than sexual minorities who were not raised Christian. No differences were noted in measures of adulthood resilience or health. Taken together, results indicate that identifying as Christian—in childhood and/or adulthood—is associated with greater experiences of stigma and stress than not identifying as Christian in adulthood.

Sexual and gender minorities (SGMs)—individuals who identify as lesbian, gay, bisexual, transgender, nonbinary, or other similar terms (e.g., queer)—often describe a complex relationship with religion/spirituality (Benson et al., 2018; Lefevor, Davis et al., 2021; Levy & Lo, 2013). Although many U.S. SGMs are religious (Pew Research Center, 2015), many de-identify from their childhood religion as adults. SGMs have described traumas and stressors associated with their prior religiousness—such as rejection from religious family and friends or religiously promoted sexual orientation change efforts—as an impetus for their religious de-identification (Exline et al., 2021; Skidmore, Lefevor, Larsen et al., 2022). SGMs who continue to identify religiously sometimes describe shifting their religious identification by finding congregations that are explicitly affirming of their sexual identities and experiences (Gattis et al., 2014). Other SGMs report continued identification with traditional (i.e., cis/heteronormative) religious communities and traditions, describing their connection with them as important aspects of their religious identification (Fredriksen-Goldsen et al., 2011; Lefevor, McGraw et al., 2022).

A recent meta-analysis found that the empirical relationship between religiousness and health among sexual minorities is complex, reflecting the varied experiences sexual—and likely gender—minorities have with religious communities and traditions (Lefevor, Davis et al., 2021). The analysis found a small but positive association between religiousness and health, but also that nearly as many studies reported a negative health effect for religiousness as reported a positive effect. Most likely, this variation is due to several moderating factors, such as the degree to which religiousness promotes or protects from some stressful experiences (e.g., congregational affirmativeness), the degree to which religiousness pervades an individual’s worldview and family culture (e.g., family religious expectations may vary for SGMs of color), and variation in the ways in which individuals identify religiously (e.g., identifying with formal religion versus a personally formulated and practiced spirituality).

In the present study we build on research on Christian identification and de-identification through an analysis of a U.S. probability sample of SGMs. Specifically, we investigate (a) which demographic characteristics are associated with SGMs’ religious de-identification as adults; (b) what childhood experiences are related to religious de-identification; and (c) how religious de-identification may relate to SGMs’ experiences of minority stressors, resilience resources, and health indicators as adults. To set the stage for our study, we first describe what it means to identify or de-identify religiously, and why both the general population and SGMs specifically may religiously de-identify. We then explore what is known about the relationship between religiousness and mental health among SGMs.

Understanding Religious Identification and De-identification

Religiousness is a multi-faceted construct that is broadly understood as people’s search for the sacred in the context of culturally sanctioned rituals, belief/value systems, and institutions (Harris et al., 2018). Religiousness comprises four cross-culturally observed, interrelated dimensions: believing (cognitive), bonding (emotional), behaving (moral), and belonging (social; Saroglou, 2011; Saroglou et al., 2020). Believing is a cognitive dimension of religiousness that describes the content and strength of individuals’ beliefs about the sacred (e.g., importance of religion, religious belief). Bonding is an emotional dimension of religiousness that describes the connection that individuals feel with the sacred and the activities individuals engage in to foster that connection (e.g., prayer, service attendance). Behaving is a moral dimension of religiousness that describes the behaviors individuals engage in or do not engage in because they are viewed as right or wrong from a religious or moral perspective. Belonging is a social dimension of religiousness that describes the connections that individuals feel to coreligionists, religious leaders, and/or their religious community. Religious individuals typically describe their religiousness as encompassing all four of these dimensions, although individuals vary in the degree to which they emphasize each dimension (Saroglou, 2011).

If identifying as religious involves experiencing some degree of believing, bonding, behaving, and belonging in context of a search for the sacred, then de-identifying religiously can be understood to involve “any change in which an individual may shift from identifying as religious to nonreligious” (van Tongeren & DeWall, 2021, p. 2). The religious deidentification process has corresponding cognitive, emotional, moral, and social dimensions (van Tongeren & DeWall, 2021): disbelief (cognitive), disengagement (emotional), discontinuance (behavioral), and disaffiliation (social). Disbelief involves a cognitive relinquishment of sacred beliefs. Disengagement involves an emotional cessation from engaging in bonding rituals aimed to bring closeness with the divine. Discontinuance involves morally separating oneself from religious pre- and pro-scriptions. Disaffiliation involves socially removing oneself from a religious group or community. Although individuals may religiously de-identify by disbelieving, disengaging, discontinuing, or disaffiliating (Chaves, 2010), many individuals religiously de-identify by disbelieving, disengaging, discontinuing, and disaffiliating (van Tongeren & DeWall, 2021).

Why Do Some SGMs Identify Religiously?

Many—and perhaps most—SGM adults identify religiously, with most religious SGMs in the U.S. identifying as Christian (Pew Research Center, 2015; Williams Research Center, 2020). Like heterosexual and cisgender individuals, one reason why SGMs may identify religiously is because of the believing, bonding, or belonging that their faith provides.

Christianity may provide SGMs with a sense of meaning in life and an understanding of which goals are worth pursuing (i.e., believing). Empirical studies find that religiousness is consistently related to meaning making (Park, 2010; Skidmore et al., under review), which may lead SGMs raised in both SGM-affirmative and non-affirmative Christian environments to attempt to preserve the cognitive and emotional meaning-making framework that their faith provides them, if possible. Christianity may also provide SGMs with opportunities to experience connection with the divine (i.e., bonding). Research has found that connection with the divine, often described as spirituality, is experienced as much more unambiguously positive by SGMs (Halkitis et al., 2009; Lefevor, Davis et al., 2021). Finally, Christianity may provide SGMs with a sense of social support (i.e., belonging). SGMs raised Christian may access support from their families (Lefevor, McGraw et al., 2021), congregations (Skidmore et al., 2022), and communities (Lassiter & Mims, 2021; Mizock & Mueser, 2014) through engagement with Christianity.

Why Do Some SGMs Religiously De-identify?

Despite being raised religious, many SGMs ultimately religiously de-identify (Exline et al, 2021; Woodell & Schwadel, 2020). SGMs may religiously de-identify for many of the same reasons as heterosexual/cisgender individuals (van Tongeren & DeWall, 2021); however, they may also religiously de-identify because they experience unique stressors in religious environments (often related to whether religious organizations affirm SGM experiences). SGMs report experiencing both distal (i.e., external) and proximal (i.e., internal) minority stressors from religious environments (Dyer & Goodman, 2022; Lefevor, Davis et al., 2021; Meyer, 2003). Indeed, many SGMs report increased exposure to stressors such as internalized homophobia/transphobia, concealment, and discrimination from being around religious people and places (Lefevor, Huffman et al., 2021). These minority stressors in turn may relate to an increase in internalized homophobia/transphobia and a decrease in life satisfaction (e.g., Barnes & Meyer, 2012; Gibbs, 2015; Meyer, 2003; Pachankis et al., 2020; Testa et al., 2017). SGMs also report experiencing religiously based discrimination or prejudice, which may help explain why, although the majority of sexual minority adults in the U.S. identify religiously (59%), only about one-third of those individuals report religion as being very important to them (as compared to 72% of heterosexual individuals who identify religiously and two-thirds of that group who report religion as being important to them; Pew Research Center, 2013, 2015).

Minority stressors may ultimately lead SGMs to religiously de-identify by promoting disaffiliation, disbelief, discontinuance, and/or disengagement. Because homophobic/transphobic attitudes are often congregationally embedded (Campbell et al., 2019; Lefevor, Tamez Guerrero et al., 2021), many SGMs experience rejection from fellow congregants (Lefevor, Huffman et al., 2021; Lefevor, Sorrell et al., 2019). This rejection may lead to feelings of disbelonging among SGMs. Similarly, many SGMs internalize negative messages about their identities in religious spaces (i.e., internalized homophobia/transphobia; Barnes & Meyer, 2012) and report concealing their identities as a way to avoid discrimination in religious spaces (Skidmore, Lefevor, Larsen et al., 2021; Lefevor, Huffman et al., 2021). As SGMs challenge these internal, negative processes (i.e., proximal stressors), they may find themselves at odds with religious doctrines, which may ultimately promote disbelief as a way to resolve cognitive dissonance (Gibbs & Golbach, 2021). Particularly as many places of worship do not allow individuals in same-gender relationships or with gender-expansive expressions to be full-fledged members or leaders of their congregation (Chavez & Anderson, 2014), SGMs may be motivated to challenge the idea that religious institutions provide a moral framework for life more generally, leading SGMs to discontinue following religious proscriptions (Woodell & Schwadel, 2020). Finally, because many places of worship endorse cis/heteronormative doctrines and policies that they believe are divinely inspired (Lefevor, Sorrell et al., 2020), challenging any aspect of faith may lead SGMs to emotionally disengage from deity.

How Does Religiousness Impact the Health of SGMs?

Religiousness may provide community, meaning, connection and isolation, loss, and rejection for SGMs. It is therefore unsurprising that empirical studies vary in their findings about the relationship between religiousness and health among SGMs (Lefevor, Davis et al., 2021; Lefevor, Sprague et al., 2019). Rather than focusing on whether religiousness promotes or hinders well-being, we focus on the diverse experiences of religiousness among SGMs.

Religiousness may relate to health differently for SGMs based on their lived experiences. Some SGMs may be more likely than others to religiously identify, due to the identities they hold and the environments in which they live. For example, SGMs of color and SGMs who live in the American South are more likely than other SGMs to live in communities where religiousness is a larger presence (Lefevor, Smack et al., 2019). Older SGMs may be more likely than younger SGMs to have been raised in religious families and communities. Similarly, SGMs socialized as women may be more likely to internalize religious values and traditions than other SGMs (Pew Research Center, 2015). Because religiousness is likely more salient for these SGMs, it may be that identifying religiously enables these SGMs to access the believing, bonding, and belonging that religiousness may provide (Newcomb et al., 2019). These individuals may thus be more likely to remain religious and to benefit from their religiousness (Rosati et al., 2020).

In childhood, religiousness may provide SGMs with a framework for understanding life as well as a sense of purpose/meaning. Given that many religious traditions endorse cis/hetero-normative ideologies (Etengoff & Lefevor, 2021), SGMs who are raised religious may also be more likely than those not raised religious to experience bullying around their sexual and gender identities, which may cause increased stress. Religiousness may also serve to complicate sexual/gender identity development for SGMs, leading SGMs raised religious to experience greater identity conflict, report later awareness of same-gender sexual desire or gender incongruence, and report later adoption of an LGBTQ+ identity label (Delhin et al., 2014).

In adulthood, religiousness may also lead to greater discrimination, more internalized negative beliefs, and less connection with other SGMs (Exline et al., 2021; Lefevor, Huffman et al., 2021; Skidmore, Lefevor, & Dillon, 2022), particularly in congregations that are not experienced as affirming (Barnes & Meyer, 2012); however, it may at the same time provide a sense of connection with religious others, enhanced sense of life purpose/meaning, and critical coping resources (Barringer & Gay, 2016; Brewster et al., 2013; Skidmore, Lefevor, Golightly et al., 2022; Skidmore et al., under review).

The Present Study

SGMs’ experiences of religion as children may lead to decisions about religious identification when they get older. These decisions may influence their exposure to minority stressors and accessibility of support, thereby impacting their health and well-being. The present study engages research on religious identification and de-identification among SGMs by asking three research questions (RQs): (RQ1) Which sociodemographic variables relate to religious de-identification among SGMs? (RQ2) What childhood variables relate to religious de-identification? and (RQ3) What adult variables relate to religious de-identification?

Based on our review of the literature, we offer the following hypotheses: (H1) Following general trends, SGMs who are older, people of color, and women will be more likely to religiously identify. (H2) SGMs who were not raised Christian will report an earlier awareness and acceptance of their SGM experience. (H3) SGMs raised Christian will report more childhood bullying and adverse childhood events than SGMs not raised Christian. (H4) Adult SGMs who stayed Christian will report greater degrees of minority stress and distress than SGMs who left Christianity or who were not raised Christian. (H5) Adult SGMs who stayed Christian will report less connectedness to SGM communities and resources than SGMs who left Christianity or who were not raised Christian.

Methods

Procedure

The data for this analysis were collected as part of the Generations and the TransPop studies. Generations is a national probability survey of sexual minority adults aged 18-60 in the U.S. Data collection occurred from April 2016 through March 2018. TransPop is the first national probability sample of transgender adults aged 18 to 60 in the U.S. Data collection occurred in two phases, from April to August 2016 and from June 2017 to December 2018. For both studies, respondents were recruited by Gallup using a dual-frame sampling procedure, including random-digit dialing and address-based sampling.

In Generations, people were included if they were in target age cohorts (18-25, 34-41, or 52-59). Black and Latinx respondents were oversampled to increase the number of racial and ethnic minority respondents; however, American Indian/Alaskan Native and Asian/Asian American/Pacific Islander respondents were not recruited (except those who were multi-racial and also identified as White, Black, or Latinx) because estimates of recruitment showed that the researchers could not recruit a sufficient number in the target age groups to allow for meaningful statistical analyses. In TransPop all ages and race/ethnicities were included. Because of the different race/ethnicity eligibility requirements in Generations and TransPop, the categories of American Indian/Alaskan Native and Asian/Asian American/Pacific Islander are not fully representative of the U.S. population. Respondents completed at least sixth grade to ensure reading comprehension for self-administration of the survey.

The samples were combined and weighted to account for response biases to better represent the population of adults SGMs in the United States. Eligible respondents who agreed to participate received a survey questionnaire via web link by email or a printed questionnaire by mail to complete by self-administration, and received a $25 gift certificate (by email) or cash (by mail) together with their survey materials. The study was approved by the Gallup IRB, the UCLA IRB, and the IRBs of other participating institutions by reliance on the UCLA IRB’s determination. Further information about these studies and extensive methodological notes can be found at www.Generationsstudy.com and www.TransPop.org.

Participants

Respondents in our sample identified as cisgender men (40.0%), cisgender women (42.6%), nonbinary (9%), and transgender men and women (8.4%). Most respondents were gay/lesbian (53.9%), White (63.4%), and college-educated (79.5%). Respondents were on average 36.4 years old (SD = 0.4), and one-third of respondents lived in the South (33.3%). See Table 1 for more detail.

Table 1.

Sample Demographic Characteristics

	%	N
Gender Identity
Cis Woman	42.6	651
Cis Man	40.0	612
Nonbinary	9.0	138
Trans Man and Woman	8.4	128
Race/Ethnicity
White	63.4	965
Latinx	17.3	265
Black	14.8	226
Other	4.5	69
Sexual Orientation
Gay/Lesbian	53.9	822
Bisexual/Queer/Pansexual	40.4	617
Other	2.9	44
Straight	2.8	43
Education
High school or less	20.5	313
More than high school	79.5	1,214
Marital Status
Married/registered domestic partnership	15.1	595
In committed relationship but not married	45.4	598
Not in a committed relationship	39.6	320
Parental status
No children	84.2	1,235
Child not in household/is >18 yrs	7.0	152
Child <18 yrs in household	8.7	129
Region
South	33.3	508
All else	66.7	1,019
	M	SD
Age	36.4 (0.4)	1,529

Measures

Religious Identification

Respondents were asked, “What is your present religion, if any?” and “Thinking about when you were a child, in what religion were you raised, if any?”. For this analysis, respondents were categorized as “Stayed Christian” if they were raised Christian (Protestant, Catholic, Mormon, or Orthodox) and also currently identified as Christian. Respondents were categorized as “Left Christianity” if they were raised Christian but did not currently identify as Christian (Jewish, Muslim, Buddhist, Hindu, spiritual, Agnostic, Atheist, or Nothing in particular). Those who were not raised Christian and did not currently identify as Christian were categorized as “Never Christian”. Respondents who converted to Christianity in adulthood (n = 27) were excluded from analyses due to the small sample size. Further, respondents who chose “Something else” for either question were excluded from the analysis (n = 162) because the experiences of SGMs in nonChristian religions are characteristically different than the experiences of those in Christian religions.

Demographic Characteristics

Sexual Identity.

Respondents were recruited in a two phase design. Respondents were first asked, “Do you, personally, identify as lesbian, gay, bisexual, or transgender?” Those who responded, “yes,” were then invited to complete a survey where they were asked further about their sexual identity with the answer options of “Straight/heterosexual”, “Lesbian”, “Gay”, “Bisexual”, “Queer”, “Same-gender loving”, or “Other.” For this analysis, same-gender loving was combined with lesbian and gay (i.e., monosexual), whereas write-in answers such as pansexual were grouped with bisexual and queer (i.e., plurisexual), and asexual and anti-label were combined with other. Straight was kept as its own category. The majority of straight individuals are trans identifying (n=33) but the rest are respondents from the Generations study (n=11). These respondents identified as LGBT and were therefore eligible but then chose straight as the sexuality they most identify with (these respondents also reported same gender attraction/behavior).

Gender.

Respondents were asked a two-step question to determine their gender identity: First, “On your original birth certificate, was your sex assigned as female or male?” and then, “Do you currently describe yourself as a man, woman, or transgender?” Respondents who said they were transgender were then asked, “Are you a trans woman (male-to-female), trans man (female-to-male), or nonbinary or genderqueer”. Respondents were classified as transgender if they chose transgender in the second question or if their current gender identity was different than their sex assigned at birth.

Race/Ethnicity.

Respondents indicated all of the racial/ethnic identities that applied to them. Respondents who identified as Latinx regardless of additional races/ethnicities chosen were categorized as Latinx; respondents who identified as Black regardless of any additional races chosen, apart from Latinx, were categorized as Black; respondents who identified as White and chose no other race/ethnic categories were categorized as White. The remaining respondents were multiracial and included those who were Asian/Pacific Islander, American Indian/Native American, and Middle Eastern/North African. The sampling design for the Generations study was based on a planned cohort analysis stratified by gender and race/ethnicity. To allow meaningful analyses investigators required a minimum number of respondents in each cohort by gender by race/ethnicity group. Preliminary predictions based on the smaller number of Asian/Pacific Islanders and American Indian/Native Alaskan in the overall U.S. population precluded recruiting these two groups in sufficient numbers over the 1-year recruitment period despite screening over 360,000 individuals (Meyer, Marken, Russell, Frost, & Wilson, 2020). To facilitate meaningful analyses the investigator then grouped Latinx people in one group, as is done in most U.S. surveys, and then non-Hispanic Black and non-Hispanic white respondents in groups that included also multi-racial Black and White respondents. This system allowed for the inclusion of diverse groups of non-Hispanic individuals in the sample (rather than using a mono-racial category).

Education.

Education was dichotomized as a high school degree or less and attended some college or more.

Geographic Region.

Using respondents’ states of residence, respondents were assigned to their corresponding Census regions. We believe the states represent a good approximation of ideology and because people from the South tend to be more religious and more frequently identify as Christian (Pew Research Center, 2015), respondents were assessed as two groups: those from the South and those from all other regions.

Childhood Experiences

Coming Out Milestones:

Throughout the questionnaire respondents were instructed to think of the identity that most corresponds with they used for their sexual identity. The instruction for respondents stated: Below are some questions about growing up and your sexual feelings when you were younger. Again, by “LGB” we mean a sexual minority identity that you identify with. Sexual minority respondents reported (a) the age when they first realized they were LGB, (b) the age when they first told a family member that they were LGB, and (c) how “out” they were in high school on a 5-point Likert scale with anchors being to no one (1) and to everyone (5).

Bullying.

Respondents reported how often they experienced bullying as a child on a 4-point Likert scale ranging from never (1) to often (4).

Adverse Childhood Experiences.

Respondents reported how many of eight adverse childhood experiences they had experienced, including emotional abuse, physical abuse, household substance use, and household mental illness. The adverse childhood experiences were originally identified by the Center for Disease Control and Kaiser Permanente and selected based on their ability to predict health and risk behaviors (Felitti et al., 1998).

Adulthood Minority Stress, Resilience, and Health Indicators

Minority Stressors.

We assessed five minority stressors: Expectations of Stigma, Internalized Homophobia, and Everyday Discrimination.

Expectations of Stigma.

Expectations of Stigma were measured using the 3-item Felt Stigma scale (Herek, 2008). Sexual minority respondents indicated their agreement with items such as “most employers where I live will hire openly LGB people if they are qualified for the job” and “most people where I live would not want someone who is openly LGB to take care of their children” on a 5-point Likert scale. Internal consistency was acceptable (α = .70).

Internalized Homophobia.

Internalized homophobia was measured using the 5-item Internalized Homophobia Scale (Herek et al., 2009). Sexual minority respondents indicated their agreement with items such as “I have tried to stop being attracted to people who are the same sex as me” and “I feel that being LGB is a personal shortcoming for me” on a 5-point Likert scale. Internal consistency in the present study was acceptable (α = .73).

Everyday Discrimination.

Everyday Discrimination was assessed by respondents’ reports of how many out of nine experiences of discrimination respondents had experienced in the past year such as being treated with less courtesy than other people and being called names or insulted. Everyday Discrimination has evidenced concurrent validity with ill health and psychological well-being (Williams et al., 1997).

Sexual Orientation and Gender Identity Change Efforts.

Respondents reported whether they had “ever received treatment from someone who tried to change your sexual orientation (such as try to make you straight/heterosexual)?” or “ever received treatment from someone who tried to make you identify only with your sex assigned at birth (in other words, try to stop you from being transgender)?” Approximately 9% of the sample received this treatment (N = 131).

Resilience.

We assessed three variables related to resilience for SGM individuals: Outness, LGBT Community Connectedness, and Social Support.

Outness.

Sexual minority respondents reported to what degree they were “out” about their sexual orientation to family and friends on a scale ranging from none (1) to all (4). Single-item measures of outness have been demonstrated to be as valid as multi-item indicators of outness (Wilkerson et al., 2016).

LGBT Community Connectedness.

LGBT Community Connectedness was assessed using Frost and Meyer’s (2012) 7-item Connectedness to the LGBT Community scale. Respondents indicated their agreement with items such as “You feel you’re a part of the LGBT community” and “You are proud of the LGBT community” on a 4-point Likert scale. Internal consistency in the present study was good (α = .86).

Social Support.

Social Support was assessed using the 12-item, Multidimensional Scale of Perceived Social Support (Zimet et al., 1988). Respondents indicated their agreement with items such as “My family really tries to help me” and “I can talk about my problems with my friends” on a 5-point Likert scale. Internal consistency was excellent (α = .93).

Health Indicators.

We assessed two health indicators: Psychological Distress and Social Wellbeing.

Psychological Distress.

Psychological Distress was assessed using the 6-item Kessler-6 Distress Scale (Kessler et al., 2003). Respondents indicated how often in the past 30 days they felt a variety of things such as “nervous” or “hopeless”, with answers ranging from none of the time (1) to all of the time (5). Internal consistency in the present study was good (α = .89).

Social Wellbeing.

Social Wellbeing was assessed using the 15-item Social Well-Being Scale (Keyes, 1998). Respondents indicated their agreement with items such as “My community is a source of comfort” and “I don’t feel I belong to anything I’d call a community” on a 7-point Likert scale. Internal consistency in the present study was good (α = .81).

Analysis Plan

We conducted three sets of multinomial logistic regression analyses using three religious categories (i.e., Stayed Christian, Left Christianity, and Never Christian). We present proportions with 95% confidence intervals (CIs) for categorical independent variables and means with standard deviations for continuous independent variables. Where the overall regression equation was significant, we also report multinomial logistic regression coefficients (Relative Risk Ratios) and 95% CIs for continuous and binary independent variables comparing the three groups. Relative risk ratios (RRRs) are a measure of effect size and can be interpreted as the probability that an outcome occurs (e.g., someone stays Christian vs. leaves Christianity) based on another characteristic. Since an RRR is a comparison of risk between groups, a null effect for an RRR will be 1, with a reduction of risk being less than 1 and a increase of risk being greater than 1.

For significant models involving independent variables with multiple categories, we use confidence intervals to assess differences between groups. In the first set of analyses, we examined demographic variables related to Christian religious de-identification, the second examined childhood variables related to Christian religious de-identification, and the third set examined adult variables related to Christian religious de-identification.

As a supplement to our analysis, we assessed differences among the 3 religious groups for LGB people and transgender people separately. If the results were different from SGM people as a whole, we mention that in the respective portion of the results section and include supplemental tables with means, proportions, and F statistics from multinomial logistic regression results. All analyses were weighted with national sampling weights where possible. The study was not pre-registered, but all data and code are available by request.

Results

Demographic Characteristics

We first examined whether there were demographic differences by Christian identification through six weighted, multinomial, logistic regressions (see Table 2). Overall, 76.8% of SGMs were raised Christian but only 36.9% of SGMs identified as Christian at the time of the survey. Compared to White SGM adults, more Black SGM adults stayed Christian (22.3% vs. 46.5). More cisgender sexual minority men than nonbinary individuals stayed Christian (34.2% vs. 9.6%). More gay/lesbian individuals stayed Christian than bisexual/pansexual/queer individuals or other/asexual/anti-label (36.4% vs 19.2% vs. 23.0%). More SGMs from the South than those not from the South stayed Christian (33.1% vs. 24.4%). Finally, those who stayed Christian were older (M = 35.0, SD = 15.3) than those who left Christianity (M = 30.5, SD = 12.9; RRR = 1.02, 95% CI: [1.01, 1.03]) or were never Christian (M = 27.7, SD = 10.1; RRR = 1.04, 95% CI: [1.03, 1.06]) groups. Comparison of percentages and confidence intervals failed to identify differences in education between individuals who stayed Christian into adulthood and those who left Christianity.

Table 2.

Demographic Correlates of Religious Identification

	Stayed Christian (N = 455)	Left Christianity (N = 778)	Never Christian (N = 296)
	% [95% CI]	% [95% CI]	% [95% CI]	F
Total sample (n = 1,529)	27.3 [24.7, 26.1]	49.5 [46.4, 52.7]	23.2 [20.5, 26.1]
Race/ethnicity				7.44*
White†, (n = 968)	22.3 [19.2, 25.7]	52.4 [48.3, 56.4]	25.4 [21.3, 29.3]
Black (n = 226)	46.5+ [38.9, 54.3]	40.5+ [33.2, 48.3]	13.0+ [8.7, 18.9]
Latinx (n = 265)	29.0 [23.0, 35.8]	49.8 [42.6, 57.0]	21.2 [15.7, 28.1]
Other (n = 69)	18.4 [9.7, 32.2]	41.9 [27.9, 57.4]	39.7 [26.6, 54.5]
Gender Identity				6.29*
Cisgender man† (n = 612)	34.2 [29.8, 39.0]	51.4 [46.5, 56.2]	14.4 [11.2, 18.2]
Cisgender woman (n = 651)	25.9 [22.1, 30.2]	48.4 [43.6, 53.2]	25.6+ [21.5, 30.3]
Transgender woman/ man (n = 128)	29.0 [19.7, 40.5]	38.9 [29.0, 49.7]	32.1+ [22.8, 43.2]
Gender nonbinary person (n = 138)	9.6+ [4.9, 17.8]	57.7 [47.1, 67.6]	32.7+ [23.7, 43.3]
Sexual Orientation				7.18*
Gay/lesbian† (n = 822)	36.4 [32.3, 40.6]	47.4 [43.1, 51.7]	16.3 [13.2, 19.9]
Bisexual, pansexual, queer (n = 617)	19.2† [15.7, 23.2]	53.0 [48.2, 57.7]	27.9† [23.7, 32.5]
Other, asexual, anti-label (n = 44)	23.0† [24.6, 30.1]	50.2 [46.5, 52.9]	26.9† [20.4, 25.9]
Straight (n = 43)	28.4 [14.2, 48.7]	32.2 [18.3, 50.0]	39.5† [22.4, 59.6]
Education				3.70*
High school or less† (n = 313)	27.5 [22.7, 32.9]	45.1 [39.3, 51.1]	27.4 [22.4, 33.0]
More than high school (n = 1,214)	27.3 [24.4, 30.3]	52.8 [49.5, 56.1]	19.9† [17.3, 22.8]
Region				4.98*
All others† (n = 1, 019)	24.4 [21.3, 27.8]	50.1 [46.2, 54.0]	25.5 [22.2, 29.2]
American South (n = 508)	33.1† [28.2, 38.3]	48.4 [43.0, 53.9]	18.6 [14.4, 23.5]
	M (SD)	M (SD)	M (SD)	F
Age (18 – 60)	35.0 (15.3)a	30.5 (12.9)b	27.7 (10.1)c	22.81*

Note: Other race/ethnicity = Asian/Native Hawaiian/Pacific Islander, Middle Eastern/North African, and American Indian/Alaska Native; Straight consists of transgender respondents in addition to cisgender respondents who identify as heterosexual but also report varying levels of different-gender attraction;

⁺indicates that percentage of individuals in the given group differs significantly from the percentage of individuals in the reference group (denoted with ^†);

F value derived from Multinomial Logistic Regression; Estimates with differing subscripts within rows differ significantly at p < .05 based on multinomial logistic regression results;

^*p < .05.

Childhood Correlates of Religious Experience

We next examined childhood correlates of religious identification through six weighted, multinomial regressions (see Table 3). The mean age at which sexual minority respondents realized they were LGB did not vary substantially by religious affiliation. However, the average age when respondents told a family member about their sexual orientation varied, with never Christian people telling family members at younger ages than those who left Christianity (RRR = 0.95, 95% CI: [0.95, 0.99]) or stayed Christian (RRR = 0.92, 95% CI: [0.89, 0.96]) groups. People who were never Christian were more out about their sexual identity in high school than those who were raised in Christianity (Compared to Stayed: RRR = 1.23; 95% CI: [1.07, 1.41]; Compared to Left (RRR = 1.15; 95% CI: [1.02, 1.30]). People who left Christianity reported more adverse childhood experiences, than people who stayed Christian and those who were never Christian (Compared to Stayed: RRR = 1.12, 95% CI: [1.04, 1.20]; Compared to Never: RRR = 1.13; 95% CI: [1.05, 1.22]). People who left Christianity were bullied in childhood more frequently than people who stayed Christian (RRR = 1.32; 95% CI: [1.15, 1.52]). For transgender people, this was true for those who left Christianity only when compared to those who were never Christian (RRR = 2.00; 95% CI: [1.26, 3.17]; see Supplemental Table 2).

Table 3.

Childhood Correlates of Religious Experience

	Stayed Christian (N = 455)	Left Christianity (N = 778)	Never Christian (N = 296)
	M (SD)	M (SD)	M (SD)	F
Age knew LGB (1-56)#	15.4 (6.1)	15.3 (5.8)	14.7 (4.7)	0.94
Age told family (1-57)#	21.5 (7.9)a	19.8 (6.4)b	18.4 (4.9)c	11.05*
Outness in high school (1-5)#	2.0 (1.5)a	2.2 (1.5)a	2.5 (1.5)b	4.55*
Childhood bullying (1-4)	2.8 (1.1)a	3.1 (1.0)b	2.9 (1.0)a	7.56*
Average # of ACEs (0-8)	3.1 (2.2)a	3.6 (2.2)b	3.1 (1.9)a	7.35*

Note. LGB = Lesbian, gay, or bisexual; ACE = Adverse Childhood Experiences; F value derived from Multinomial Logistic Regression; Estimates with differing subscripts within rows differ significantly at p < .05 based on multinomial logistic regression results;

^#these variables were only asked of sexual minority respondents in the Generations study (n = 1,342, Stayed Christian n = 420, Left Christianity n = 675, Never Christian n = 247);

^*p < .05

Adult Minority Stress and Health Correlates of Religious Experience

Finally, we examined adult correlates of religious experience through 13 weighted, multinomial regressions (see Table 4). Respondents in the three groups differed in most of their experiences of minority stressors. In particular, sexual minority people who were never Christian reported less current felt stigma than those who were raised Christian (Compared to Stayed: RRR = 0.74, 95% CI: [0.61, 0.91]; Compared to Left: RRR = 0.73; 95% CI: [0.60, 0.89]). Further, people who stayed Christian reported more current internalized homophobia and transphobia than those who left Christianity (RRR = 1.32, 95% CI: [1.12, 1.54]), though this relationship was not significant for transgender people separately (See Supplemental Table 3). More people who stayed Christian had a history of conversion therapy than those who left Christianity (RRR = 1.77, 95% CI: [1.07, 2.92]) and those who were never Christian (RRR = 2.85, 95% CI: [1.41, 5.74]). Although the groups did not differ in their experiences with discrimination when looking at LGBT people together, transgender people who were never Christian experienced discrimination less frequently than transgender people who stayed or left Christianity (Compared to Stayed: RRR = 0.84; 95% CI: [0.73, 0.97]; Compared to Left: RRR = 0.83; 95% CI: [0.69, 0.99]; see Supplemental Table 3).

Table 4.

Adulthood Correlates of Religious Experience

	Stayed Christian (N = 455)	Left Christianity (N = 778)	Never Christian (N = 296)
Resilience	M (SD)	M (SD)	M (SD)	F
Current outness (1-4)#
Family	2.9 (1.1)	2.8 (1.1)	2.8 (1.0)	1.16
Friends	3.2 (1.0)	3.3 (0.9)	3.2 (0.8)	0.01
LGBT community connectedness (1-4)	3.0 (0.7)	3.1 (0.6)	3.0 (0.6)	1.05
Social support (1-7)	5.2 (1.4)	5.1 (1.2)	5.2 (1.2)	0.80
Health Indicators
Psychological distress (0-24)	8.3 (6.0)	8.9 (5.4)	9.5 (5.2)	2.55
Social well-being (1-7)	4.5 (0.9)	4.6 (0.9)	4.4 (0.9)	1.64
Minority Stress
Felt stigma (1-5)#	2.8 (0.9)a	2.8 (1.0)a	2.5 (0.9)b	5.22*
Internalized homophobia and transphobia	1.9 (0.9)a	1.7 (0.8)b	1.8 (0.8)a, b	5.85*
Discrimination (0-9)	3.0 (3.0)	3.2 (2.9)	2.7 (2.5)	2.03
	% [95% CI]	% [95% CI]	% [95% CI]	F
Experienced SOGICE	13.9 [10.0, 19.0]	8.3 [6.1, 11.2]	5.3 [3.0, 9.3]	497 **

Note. SOGICE = Sexual Orientation Gender Identity Change Effort. F value derived from Multinomial Logistic Regression; Means with differing subscripts within rows differ significantly at p < .05 based on multinomial logistic regression results;

^#these variables were only asked of sexual minority respondents in the Generations study (n = 1,342, Stayed Christian n = 420, Left Christianity n = 675, Never Christian n = 247);

^*p < .05

The groups did not differ from one another in health indicators, including psychological distress, F(2, 1527) = 2.55, p = .078, or social well-being F(2, 1527) = 1.64, p = .194, though when LGB people were assessed separately from transgender people, LGB people who were never Christian experienced higher psychological distress compared to LGB people who stayed Christian (RRR = 1.05, 95% CI: [1.01, 1.09]; see Supplemental Table 3)The groups also did not differ in their experience of resilience factors, including outness to family, F(2, 1311) = 1.16, p = .313, outness to friends, F(2, 1315) = 0.01, p = .988, LGBT community connectedness, F(2, 1527) = 1.05, p = .350, and social support, F(2, 1527) = 0.80, p = .451).

Discussion

Using a national probability sample, we examined the demographic and developmental correlates of religious de-identification. We found that SGMs who were older, were Black, were cisgender men, and/or who lived in the American South were more likely to identify as Christian as adults. SGMs who were raised Christian experienced more adverse childhood experiences and more bullying than other SGMs, and sexual minorities who were raised Christian were less likely to be out than other sexual minorities. Finally, sexual minorities who identified as Christian as adults reported more stigma, and SGMs who identified as Christian as adults reported more internalized homophobia/transphobia but similar degrees of health and resilience as other SGMs.

Although others have examined the link between religiousness and stigma or stress, few have done so using national probability samples (Lefevor, Davis et al., 2022). Further, few studies have sought to examine the separate effects of childhood and adulthood religious affiliation on religious identification, stress, and health. Our study contributes to research on religious identification and de-identification by remedying each of these oversights.

Why are Some Groups of SGMs More Likely to Religiously Identify Than Others?

Religiousness may provide a framework of belief, the experience of bonding, a sense of belonging, and standards for behaving for those who identify religiously (Saraglou et al., 2020). In contrast, de-identifying religiously is a costly process marked by disbelief, disengagement, discontinuance, and disaffiliation (van Tongeren & DeWall, 2021). For SGMs who are raised religious, religious de-identification may often involve a sense of loss that affects them personally (e.g., loss of a source of meaning in life), interpersonally (e.g., loss of connections to family and friends), and structurally (e.g., loss of a structure to govern behavior; Fenelon & Danielson, 2016). Given the costs of religious de-identification, we posit that individuals raised religious are often naturally inclined to identify religiously until the costs of identifying religiously outweigh the costs of religiously de-identifying (Lefevor, McGraw et al., 2022).

This lens is useful in explaining the trends in religious de-identification observed in response to our first hypothesis (that SGMs who are older, people of color, and women will be more likely to religiously identify). One reason why some individuals may have remained religious while others religiously de-identified may be that those who remained religious did so because it allowed them to retain access to systems of power and privilege. Historically, Christianity has privileged the experiences of heterosexual, cisgender men in a variety of ways, ranging from providing them opportunities for priesthood and leadership to providing religious doctrine to justify their gender expression. Because these experiences are privileged, those who hold heterosexual, cisgender, and male identities often experience religious traditions and communities as accommodating more of their needs and experiences. In contrast, those who hold bi/plurisexual, transgender, female, and nonbinary identities are less likely to feel that religious traditions and communities accommodate their unique needs and experiences. Consequently, those who hold more relatively privileged identities (e.g., cisgender, male) may be more likely to continue to identify religiously than those with less relatively privileged identities (e.g., bi/plurisexual, transgender, female, nonbinary; Miller, 2013).

Another reason that some individuals may continue to identify religiously where others religiously de-identify is that religiousness may provide some individuals with access to greater sociocultural resources than it does for other individuals. For example, SGMs of color or SGMs raised in more rural areas are more likely to exist in a more religiously infused milieu than other SGMs (Lefevor, Smack et al., 2020). For these individuals, identifying religiously may provide access to support from family or friends. In contrast, White SGMs or SGMs in more urban areas may be less likely to have religiously based familial and social networks, affording them greater social flexibility (Peach, 2003). Further, older adults may be more likely to have more highly religious familial and social networks than younger adults because of the historical importance of religion. Consequently, remaining religiously engaged may help older adults connect with their peers in a way that may be less helpful for younger adults (Peach, 2003).

A final reason that some individuals may religiously identify where others religiously de-identify is that religiousness may result in a more stressful experience for some individuals than others. At the institutional level, different Christian denominations vary in the degree to which they affirm SGM experiences, with some denominations advocating for SGM rights and allowing all members to fully participate in religious rites and others restricting the involvement opportunities for SGMs (Cravens, 2018). Although we were not able to assess the degree to which religious traditions discourage same-sex sexual behaviors or gender-expansive expression, we suspect that participants who were part of traditions that discourage these things would have been more likely to de-identify religiously than other SGMs.

How Does Christianity Relate to SGMs’ Experiences in Childhood?

One reason that SGMs may religious de-identify is because they experience distal and proximal stressors related to identifying religiously. Our second (i.e., that SGMs who were not raised Christian will report an earlier awareness and acceptance of their SGM experience) and third (i.e., SGMs raised Christian will report more childhood bullying and adverse childhood events than SGMs not raised Christian) hypotheses investigated these childhood stressors. Confirming our hypotheses, we found that SGMs who left Christianity experienced more adverse childhood experiences and more bullying than SGMs who were never Christian or who continued to identify as Christian. Further, we found that sexual minorities who were raised Christian—regardless of their ultimate religious identification—reported being less out in high school and telling their family about their sexuality later than sexual minorities who were never Christian. These findings support research that shows that religiously identified SGMs report experiencing greater distal (e.g., discrimination, rejection; Lefevor, Huffman et al., 2021; Pew Research Center, 2013) and proximal (e.g., internalized homophobia, concealment; Barnes & Meyer, 2012; Wilcox, 2006) stressors. Likely, SGMs who were raised Christian experienced greater stress at least in part because of the cis/hetero-normative religious ideologies they were raised in (Etengoff & Lefevor, 2021).

Although not a research question or hypothesis, we noted that although more than three-quarters of our sample was raised Christian, two-thirds of these individuals religiously de-identified as adults. This overwhelming movement toward religious de-identification—particularly in light of the costs associated with de-identification (Fenelon & Danielson, 2016)—likely suggests that these SGMs found their religious identification stressful. This trend follows national trends that suggest that sexual minorities de-identify religiously at higher rates than their heterosexual peers (Woodell & Schwadel, 2020). Nonetheless, longitudinal research that tracks current experiences of minority stressors with ultimate religious de-identification is needed to verify that stress is the reason for this movement.

Religious Disaffiliation, Minority Stress, and Health in Adulthood

Some have suggested that childhood minority stress processes result in enduring vulnerabilities in adults (Wight et al., 2015). In the context of religion, this would mean that, because SGMs have been exposed to a higher degree of religiously based stigma and discrimination as children, they would show enduring health disparities relative to SGMs who were exposed to lower degrees of religiously based stigma and discrimination. We investigated these possibilities through our fourth (i.e., that adult SGMs who stayed Christian would report greater degrees of minority stress and distress than SGMs who left Christianity or who were not raised Christian) and fifth (i.e., that adult SGMs who stayed Christian would report less connectedness to SGM communities and resources than SGMs who left Christianity or who were not raised Christian) hypotheses.

In general, we found evidence for the concurrent relationship between identifying as Christian and minority stressors. Those who currently identified as Christian reported greater internalized homophobia/transphobia than any other group and greater felt stigma than those who never identified as Christian. Given that sexual minorities who remain Christian also report more internalized homophobia and less outness (e.g., Lefevor, Skidmore et al., 2021), we suggest that the benefits of continued engagement in Christianity may counteract the adverse effects of the increased internalized homophobia/transphobia SGMs experience because of their engagement with Christianity (e.g., Barnes & Meyer, 2012; Dodge et al., 2012). In contrast, it is also possible that SGMs who had the most difficult time with their Christian identity may choose to de-identify, thus creating a healthy worker effect where SGMs who continue to identify as Christian are those who were not as adversely affected by their identity in the first place.

We did not find evidence to support the idea that childhood identification as Christian or current identification as Christian were related to psychological distress or social well-being. Many SGMs experience the support and benefits of Christianity that are often reported by the general population (e.g., Koenig et al., 2012; Pargament et al., 2004; Villani et al., 2019), which may hold true for some SGMs even when childhood and adulthood stressors are present (e.g., Dodge et al., 2012). This finding coincides with trends suggesting that SGMs in conservative religions largely experience similar mental health outcomes as their nonreligious counterparts, even though these SGMs report experiencing more minority stressors, likely because they experience more religiously related support (e.g., Lefevor, McGraw et al., 2022; Lefevor, Sorrell et al., 2021; Mizock & Mueser, 2014).

Despite the lack of relationship between religious identification and distress, many SGMs religiously de-identify. This trend may suggest that the stressors and stigma associated with their religious affiliation may begin to contribute more to stigma than to support, thus leading Christian SGMs to distance from religious teachings, feel disillusioned with religious people, and feel a lack of support or belonging from their religious communities (van Tongeren & DeWall, 2021). SGMs may thus leave Christianity for social reasons (e.g., stigma and social struggles), and also remain Christian for social reasons (e.g., affiliation fulfills their social needs and gives some SGMs continued access to sociocultural benefits, such as resources and opportunities).

Limitations

The present study was limited by several factors. Firstly, our study screened in participants who consider themselves to be LGBT. We note that some individuals who identify as queer or asexual may have been excluded from the study and therefore we can only generalize our results to LGBT individuals. We relied on respondents’ reports of their childhood and adult religious affiliation. Thus, although our research questions were inherently longitudinal, we relied on cross-sectional recollected information to evaluate our hypotheses. Like all self reported information, this data is potentially subject to various recall biases. Still we have no significant reasons to think that respondents biased responses to these questions. For example, as they don’t seem to present demand characteristics or social desirability biases. These are likely in items where respondents assume the researchers purpose in the study or want to present themselves in a favorable light (Krumpal, 2013). Nonetheless, future research should examine longitudinal changes in religiousness rather than relying on retrospective reports.

We also did not ask specific questions about the motivation for and context of religious disaffiliation. Further, we were not able to examine whether respondents changed between Christian congregations, for example, from a congregation with strict rules prohibiting same-sex sexual behavior to one that supports SGMs. Although we found changes in religious identification, we did not have measures of congregation- or denomination-specific variables. As such, we could not examine the LGB-affirmativeness of respondents’ religious milieus, the degree to which minority stressors occurred in religious settings, or the degree to which individuals held their religious beliefs to be central to their identities. Future research should enquire about (a) the explicit policies/procedures of a congregation/denomination about SGMs, (b) SGMs’ perceptions of the affirmativeness of their congregation/denomination, and (c) the degree to which SGMs’ religious identities are salient to them. Our categorization of participants into only three categories of “Stayed Christian”, “Left Christianity”, and “Never Christian” is quite broad and assumes that all experiences of Christianity are the same, which they are assuredly not. We recognize that each of these variables could explain the relationship of religiousness with minority stressors and health but continue to assert that at least some new knowledge was gained from the broad analyses we employed. Thus we both recognize that their omission constitutes a limitation to our results and believe that our findings continue to carry important implications about religiousness, minority stressors, and health.

Implications

These findings may be particularly helpful to therapists who work with SGMs, regardless of SGMs’ current religious affiliation. Our findings first highlight that SGMs are likely to have been raised in a religious background, regardless of their current identification. As such, therapists would do well to enquire about the relevance of SGMs’ religious background (or lack thereof) for them at present. Secondly, our findings highlight that many SGMs ultimately religious disaffiliate. In our introduction/discussion, we have introduced framework around religious affiliation/disaffiliation that may help therapists discuss with clients the particular ways in which religion may have influenced clients (i.e., believing, bonding, behaving, belonging). This framework may be particularly helpful for clients who are currently navigating religious disaffiliation because it may help clients articulate the pieces they found/find valuable about religion and either grieve or strive to maintain these aspects.

Conclusion

Using a national probability sample of SGMs, we examined the rates of Christian religious identification and de-identification among SGMs, as well as its correlates with adulthood psychological distress. By examining changes in religious affiliation over time, minority stressors, and health, we found that certain groups of SGMs are much more likely than others to retain affiliation with Christianity, particularly SGMs who hold privileged identities or those for whom Christianity may be a larger part of their cultural milieu. We also found that most SGMs raised Christian ultimately leave Christianity, and found tentative evidence to suggest that SGMs may leave Christianity due to the heightened experience of minority stressors within Christian environments. Further, although the process of disaffiliation may be arduous, we did not find lasting harmful effects from Christian religious disaffiliation. Finally, we found that at least some SGMs maintained connection with Christianity, which may have ultimately been protective for their mental health. Taken together, our findings show multiple ways that religious identification/de-identification affects resilience and distress for SGMs raised Christian.

Public Significance Statement:

Using a national, representative sample of sexual and gender minorities (SGMs), we found that SGMs who were raised Christian reported more minority stressors but similar health and well-being to those who were not raised Christian.