Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2021 Aug 30.
Published in final edited form as: Psychol Addict Behav. 2020 Feb 20;34(4):512–520. doi: 10.1037/adb0000559

Relations among Religiosity, Age of Self-Identification as Gay, Lesbian, or Bisexual, and Alcohol Use among College Students

William R Corbin 1, Thai Q Ong 2, Charlie Champion 3, Kim Fromme 4
PMCID: PMC8405059  NIHMSID: NIHMS1068238  PMID: 32077708

Abstract

Research indicates that Lesbian, Gay, and Bisexual (LGB) youth are at increased risk for heavy drinking relative to their heterosexual peers. One possible contributing factor is that religiosity fails to provide the significant protection for LGB youth that it provides in general population samples. Although prior studies provide some support for this hypothesis, there is little research on the reasons that religiosity may fail to protect against heavy drinking among LGB youth. The current study addressed this question by examining relations among religiosity, age of self-identification as LGB, and alcohol use in a sample of 162 young adults self-identified as lesbian, gay, bisexual, or questioning. Age of self-identification was conceptualized as an indicator of stress or internal conflict between religious beliefs and self-identification. We hypothesized that higher levels of religiosity would be associated with a later age of identification as LGB which, in turn, would contribute to heavier drinking. Path analyses conducted in Mplus provided support for this hypothesis, with higher levels of religiosity indirectly contributing to increased alcohol use through later age of self-identification. Whereas additional studies directly assessing stress associated with the self-identification process are needed, the results of this study suggest that religiosity may serve as a risk, rather than protective, factor among LGB youth. The development of religious support groups specifically for LGB youth may help these individuals reconcile their religious beliefs with their emerging sexual identities, thereby allowing them to derive some of the same benefits that heterosexual youth derive from religious beliefs and practices.

Keywords: Religiosity, Sexual Minority, Self-Identification, Alcohol, College

Although college students as a group are at increased risk for heavy drinking (Hatzenbuehler et al., 2008; Marshal et al., 2008, 2009; Stall et al., 2001), there is considerable variability across subpopulations. Young adults who identify as lesbian, gay, or bisexual (also referred to in the literature as “LGB” or “sexual minorities”) may be at particularly high risk. For example, Marshal et al. (2009) found that rates of substance use for lesbian, gay, and bisexual young adults were 190% higher than for young adults who identified with a majority sexual orientation. Young adults who identify with a minority sexual orientation also report having a younger age of drinking onset relative to their heterosexual counterparts (Corliss et al., 2008).

Initial studies documenting increased risk for heavy drinking and related problems among LGB youth relied on cross-sectional data, but longitudinal studies have now substantiated the results of cross-sectional studies. For example, Marshal, Friedman, Stall, & Thompson (2009) assessed the relation between sexual orientation and both initial levels (intercepts) of alcohol use in adolescence (average age of 15.8 years), and changes in drinking over the following six years (slopes). The results revealed that youth who self-identified as LGB reported both higher initial rates of alcohol use and more rapid increases over time when compared to their heterosexual peers. Same-sex romantic attraction was related to alcohol use at wave 1 (intercept) but not to changes over time (slope), whereas same-sex sexual behavior was not related to alcohol use at wave 1 but was related to changes in alcohol use over time. These results suggest that the effects of identifying as LGB on drinking behavior are not static and that processes related to self-identification may contribute to changes in drinking over time. Although this study made an important contribution by demonstrating effects of sexual identity on changes in drinking over time, the findings are limited by the fact that sexual orientation was not assessed until the final wave of data collection in early adulthood.

In another longitudinal study, Talley, Sher, & Littlefield (2010) examined developmental changes in substance use behaviors based upon sexual orientation in a sample of 3,720 first-time college students at a large public university. Alcohol use was assessed from the summer prior to college through the fourth year of college, and sexual orientation was assessed in the second semester of both the first and fourth years of college. The authors found that individuals who reported any level of sexual minority self-identification at the beginning of emerging adulthood reported higher initial levels of binge drinking along with higher frequencies of alcohol and cannabis use. Findings for same-sex attraction and same-sex behavior were less consistent than for self-identification. Importantly, the timing of the assessment of sexual orientation impacted the findings. For those who self-identified as LGB in the first year of college, cannabis use and binge drinking declined in the early college years and accelerated in the later years of college, whereas those who self-identified at a later age had more stably elevated rates of binge drinking and cannabis use across the four years (Talley et al. 2010). These findings highlight the idea that timing of self-identification may play a major role in use of certain substances and therefore warrants further investigation.

Although there is some evidence that LGB youth are at increased risk for both earlier heavy drinking and greater increases in drinking across time, results of a study by Hatzenbuehler, Corbin, & Fromme (2008) suggest that there may be important gender differences in these relations. In a study examining trajectories of alcohol use among 111 LGB young adults and a comparison group of heterosexual young adults, at baseline (senior year of high school), lesbian and bisexual females reported heavier drinking than heterosexual females whereas no significant differences were found between heterosexual and sexual minority males. In contrast, when looking at changes in alcohol use over time, gay and bisexual males increased their alcohol use at a more rapid rate than heterosexual males. Newcomb et al. (2012) also found that gay and bisexual males increased their drinking at a faster rate from age 16 to 20 than did lesbian and bisexual females. However, a recent study by Dermody et al. (2019) suggests that gender differences in the effects of sexual minority self-identification on drinking trajectories may be age dependent. Consistent with prior studies, they found that self-identification as sexual minority at any point between age 14 and 20 was associated with higher overall levels of alcohol use. With respect to changes over time, sexual minority females showed more rapid increases in early adolescence (ages 13–15) and less rapid increases in late adolescence and early adulthood (17–20), relative to heterosexual females. Collectively, these findings suggest that effects of identifying as LGB on alcohol use may differ by gender and change across development.

The minority stress model (Meyer, 2003) offers a theoretical framework for understanding why LGB youth are at heightened risk for alcohol use and related problems. The minority stress model proposes that the stigmatization, discrimination, and internalized biases experienced by minorities create unique and chronic stress which leads to increased negative health and psychosocial outcomes (Meyer, 2003). Consistent with this model, a study using data from the 2nd wave of data of the National Epidemiologic Survey on Alcohol and Relation Conditions (NESARC) found that the odds of having a past-year substance use disorder were 1.72 times greater for LGB adults who reported only discrimination related to their sexual minority status, 2.24 greater for those who experienced discrimination related to both biological sex and sexual orientation, and 4.12 greater for those who experienced discrimination related to race, biological sex, and sexual orientation (McCabe, Bostwick, Hughes, West, & Boyd, 2010). In addition, a meta-analysis found that both general and gay-related stress and victimization experiences were related to greater substance use (Goldbach, Tanner-Smith, Bagwell, & Dunlap, 2014).

Although research on risk factors has helped establish a link between sexual orientation and drinking outcomes, little attention has been given to potential protective factors. One well-established protective factor against negative drinking outcomes is religiosity. For example, a recent study using data from Monitoring the Future showed that higher religiosity consistently protected against an early age of onset, delaying onset by an average of two years (Barry, Valdez, & Russell, 2019). In addition, a systematic review of 43 studies on the relationship between adolescent religiosity/spirituality and health outcomes, over a fourth of which included national samples, demonstrated a reliable association between the two with higher levels of religiosity/spirituality resulting in better health outcomes (Rew & Wong, 2006). Relative to religiosity/spirituality, religious denomination showed less consistent associations with adolescent health outcomes. Further, a longitudinal study in New York public schools showed that religion can act as a buffer against use of a wide range of substance use outcomes including cigarette smoking, heavy drinking, and marijuana use (Wills, Yaeger, & Sandy, 2003).

However, research suggests that religiosity might operate differently for sexual minority relative to sexual majority youth. Unfortunately, the few studies that have focused specifically on substance use among LGB youth have failed to find consistent evidence for a protective role of religiosity. For example, a longitudinal study by Rostosky, Danner, & Riggle (2007) using the Add Health data set demonstrated that religiosity in high school was not protective against substance use in early adulthood for self-identified sexual minorities. In a separate paper using the Add Health data set, Rostosky, Danner, & Riggle (2010) found that religiosity protected against heavy episodic drinking (HED) among heterosexual women, was not protective for lesbian women, and was associated with greater HED for bisexual women. For men, religiosity was protective and the effects did not differ by sexual identity. In another study that looked at interactions between sexual identity, sex/gender, and religiosity in the prediction of substance use in college students, Eliason, Burke, Olphen, & Howell (2011) found a three-way interaction, such that being religious was associated with reduced risk for heavy drinking among LGB women but increased risk among men. A study of women in later adulthood using nationally representative samples also demonstrated protective effects of religiosity among LGB women (Drabble, Trocki, & Klinger, 2016). The disparate findings across these studies may be a result of differences in the measurement of religiosity and/or differences in the age group under study (high school vs. college students and older adults). Regardless, the findings highlight the importance of examining potential sex differences in relations between religiosity and alcohol use in sexual minority populations.

Having demonstrated that religiosity does not consistently protect against heavy drinking in sexual minority samples, it is important to begin to identify potential explanations for this lack of protective effect. One possibility is that, because orthodox religions tend to be heterosexist and often opposed to the lifestyles of the LGB community (Clark, Brown, & Hochstein, 1990; Valera & Taylor, 2011), religious identification could be in conflict with sexual minority identification. As a result, sexual minority youth may distance themselves from their religion, suppress their sexual minority identity, or continue to struggle with maintaining what may seem like incompatible identities. Consistent with these possibilities, a study using a sample of 11,699 adolescents found that youth who identified as sexual minorities were less likely to report a religious affiliation compared to their heterosexual peers. Sexual minority individuals were also found to have larger declines in religiosity from adolescence to adulthood (Rostosky, Danner, & Riggle, 2008). These findings suggest that LGB youth may resolve the discrepancy by reducing their levels of religiosity. However, because sexual identity was assessed only in adulthood, it is possible that the results reflect greater difficulty with self-identification for religious youth, and a greater acceptance of sexual minority status among those who become less religious from adolescence to adulthood. Consistent with this possibility, Ross & Rosser (1996) found that having moral or religious concerns about homosexuality was associated with a lower likelihood of being openly gay. There is also evidence that conflict between religious beliefs and sexual identity may have important behavioral consequences. For example, Eliason et al. (2011) found that college men who were high in religiosity and/or were not comfortable labeling their sexual identity were at greatest risk for negative drinking outcomes.

Although prior studies have not specifically examined the role of religion as a potential stressor in the coming out process, the process of identifying as LGB may serve as a stressor due to the prevalence of homophobia and heterosexist ideals in society more broadly (Charbonnier & Graziani, 2016; D’Augelli, 2002; Meyer, 2003). Unfortunately, this may lead to a delay in acceptance of one’s emerging sexual identity, and such delays may have negative implications for health and well-being Although coming out as LGB is associated with significant discrimination experiences and coming out at a younger age may increase exposure to discrimination and victimization (Pilkington & D’Augelli, 1995), hiding one’s sexual identity also has a negative impact (Harrison, 2003). Although some have suggested that hiding one’s stigmatized identity may act as a protective strategy, concealment is related to decreased feelings of acceptance, increased feelings of loneliness and isolation, and worse mental health outcomes (Camacho, Reinka, & Quinn, 2019). For example, Mohr and colleagues (2019) found that, within a workplace setting, individuals who concealed their sexual minority identity felt decreased positive affect and increased negative affect that same day, with some effects spilling over into the next day. Further, although coming out is associated with acute increases in distress (D’Augelli, 2006), it is also associated with growth and more positive long-term outcomes (Solomon, McAbee, Asberg, & McGee, 2015; Halpin & Allen, 2004). For example, Halpin & Allen (2004) found evidence that health and well-being improved as individuals moved through the latter stages of sexual minority identity formation. Although well-being was lowest during the Tolerance stage of identity development when self-identification is likely to occur, well-being increased relatively linearly through the latter stages of identity development. Prior research has also demonstrated that individuals who are out as LGB report less avoidance and greater parental support (Mohr & Fassinger, 2003), and perceive that coming out ultimately improves quality of life (Monroe, 2001), and relationships with their partners (Lasala, 2000). Thus, increases in self-acceptance, access to positive role models, and social support (at least from some individuals) may help offset the stress associated with potential negative parental reactions and greater discrimination experiences that often occur when a sexual minority status is disclosed (Harrison, 2003).

Studies have also linked the process of self-identification as LGB to substance use. For example, Savin-Williams & Ream (2003) found that a longer delay between awareness and acceptance of one’s sexual minority status was associated with increased risk of both substance use and suicide attempts. Although less directly tied to the self-identification process, Weber (2008) found that individuals with substance use disorders reported experiencing greater heterosexism and internalized homophobia, the latter of which is associated with lower levels of disclosure of sexual orientation (Ross & Rosser, 1996). Although these studies suggest that age of identification could serve as a mechanism of risk among LGB youth who are high in religiosity, we are aware of no prior studies directly addressing this question.

Thus, the current study sought to build upon prior research by clarifying relations among religiosity, the process of identifying as LGB, and changes in drinking behavior during the period of peak risk for hazardous drinking. Based on prior research, we hypothesized that religion would act as both a risk and protective factor for LGB young adults. Consistent with the findings of prior research, direct protective effects were expected when controlling for indirect risk effects. Risk effects were expected to operate through a later identification as LGB which we conceptualized as representing conflict between religion and emerging sexual identity. Consistent with this conceptualization, prior research suggests that religious and moral concerns about homosexuality are associated with reduced disclosure of sexual identity (Ross & Rosser, 2004), and that a longer delay in accepting one’s sexual identity is associated with negative drinking outcomes (Savin-Williams & Ream, 2003). Thus, we anticipated that LGB youth who were higher in religiosity would take longer to self-identify as LGB, and that this later self-identification would be associated with increased risk for heavy drinking and related problems. We also examined interactions between gender and sexual orientation in the prediction of both time to self-identification and changes in drinking and problems over time. Given the inconsistent nature of prior findings regarding gender differences, we made no a priori hypotheses about gender differences.

Method

Participants and Procedures

The final sample comprised only those who identified as sexual minorities at some point across the 10 waves of the study. First year students between the ages of 17–19 who were unmarried and had not previously attended college or university were eligible to participate in the study. A total of 3,046 participants were invited to complete a baseline survey (wave 1) in the summer prior to college matriculation, and additional surveys in the fall and spring semesters of the first three years of college (waves 2–7), the fall semester of year four of college (wave 8) and the fall of the two years following college (waves 9–10). Students accessed the baseline and all follow-up surveys though a secure web site (DatStat, Seattle, WA), where they were able to give consent and complete the surveys. Of the 3,046 participants invited to participate in this longitudinal study, 2,245 (73.7%) consented to the study and completed the high school survey. Of the 2,245 who initiated the study, 162 (7.2%) reported identifying as a sexual minority at one or more time points across the 10 waves of data collection. This is likely an underestimate of the percentage of students who are LGB as some participants who dropped out of the study before year 4 may have identified as LGB at later waves had they continued in the study. Within the LGB sample (n = 162, 51.9 % female), the majority were Caucasian (54.9 %), with 18.5% Hispanic/Latino, 11.7% of mixed race/ethnicity, 9.3% Asian American, 4.3% African American, and 1.2% who declined to report their ethnicity. This racial/ethnic distribution is representative of the university population at the time of recruitment.

Measures

Demographics.

Participants provided basic demographics including age, sex (n = 78 male; n = 84 female), and race/ethnicity. Since prior research has provided evidence for gender variability in drinking behavior among LGB youth, gender was used as a covariate in the current study and also as a potential moderator of relations among religiosity, age of self-identification as LGB, and alcohol use.

Sexual Identity Classification.

Sexual identity was measured by a single-item: “What is your sexual orientation?” Participants were given five possibilities: heterosexual/straight, bisexual, gay, lesbian, or questioning. The time to self-identify was measured and calculated based on the survey during which participants first identified as gay, lesbian, bisexual, or questioning (e.g. wave 1=1, wave 2=2, etc.). Forty-two percent of the sample identified as LGB at wave 1. Reflecting ongoing sexual minority identity development, there was considerable variability in reports across waves. A total of 90 participants (55.6%) consistently identified as LGB after first identifying as such, but the remaining 77 participants (44.4%) fluctuated between identifying as heterosexual and LGB across waves. We considered participants sexual minorities if they identified as sexual minority at any time-point.

Drinking Outcomes.

Binge drinking was measured by a single item: “During the past 3 months, how many times did you have four/five (for women/men) or more drinks at a sitting?” (Wechsler, Davenport, Dowdall, Moeykens, & Castillo, 1994). Frequency of intoxication was measured with a single item: “During the past 3 months, how many times did you get drunk (not just a little high) on alcohol?” (Jackson, Sher, Gotham, & Wood, 2001). Scores on both variables were standardized and the mean of the variables was used as a composite measure. The mean was log-transformed due to positive skew which resulted in a normal distribution.

Religiosity.

Religiosity was measured on two-dimensions, religious activities and religious values. On the first item, participants were asked, “During an average week of the last 3 months how many hours did you spend in religious activities?” Participants responded on a five-point scale ranging from 1 = 0 to 5 = more than 20 hours. On the second item, participants were asked, “For you, how important is it to observe religious beliefs and/or traditions?” using a 5-point Likert scale (1 = not at all important to 5 = very important to me). The two items were strongly correlated (r = .41) so they were standardized and the mean of the two items was used as a composite measure of religiosity.

Data Analytic Plan

Preliminary Analyses

Prior to conducting the primary analyses, all variables were examined to evaluate assumptions of normality. Any variables that were not normally distributed were log-transformed. See Table 1 for means and standard deviations of all study variables.

Table 1.

Descriptive Statistics by Gender

Wave 1 Wave 10
Women (n = 84) Men (n = 78) Women Men
M SD M SD M SD M SD
Religious Attendance 0.54 0.81 0.47 0.64
Religious Values 2.65 1.53 2.71 1.42
Age of Self-Identification 19.85 1.65 19.34 1.52
Drinking to Intoxication 1.95 3.30 1.79 4.17 3.15 4.91 5.86 12.26
Frequency of Binge Drinking 2.30 4.15 1.77 3.99 4.54 7.42 5.35 12.29
Open in a new tab

Note. N = 162. Time of self-identification was used in the indirect effects model, but for ease of interpretation, age of self-identification is presented here for descriptive purposes. Religious attendance was measured by the number of hours the participant spent on religious activities and religious values was measured by the importance of observed religious beliefs and/or traditions using a Likert scale (1=not important at all and 5=very important to me).

Primary Analyses

The primary analyses were tested using Mplus Version 7.11 (Muthén & Muthén, 2012). Missing data was handled using the full information maximum likelihood (FIML) method. Bootstrapping (K = 5,000) was used to address remaining issues of non-normality in the data following transformations. Gender and Wave 1 drinking were both used as covariates in the analyses. The primary analyses included gender, Wave 1 alcohol use, and Wave 1 religiosity as predictors of both age of self-identification and Wave 10 alcohol use, and age of self-identification as a predictor of Wave 10 alcohol use. We chose to use Wave 10 drinking as the outcome to allow for prospective prediction from age of self-identification, as students could identify as sexual minority at any time up through Wave 10.

To test the primary study hypotheses, we examined both direct effects and indirect effects of religiosity on alcohol use. The test of the indirect effect of religiosity on alcohol use operating through age of self-identification used the bias-corrected bootstrap technique. This approach uses asymmetric confidence intervals to examine indirect effects because the product of two random variables is often not normally distributed. If the confidence interval does not contain the value of zero, the indirect effect is considered statistically significant at the specified alpha level (.05 in this case). Prior studies have shown that confidence intervals calculated based on the product distribution have more power and more precise type 1 error rates than alternate approaches (MacKinnon et al., 2002; Mackinnon, Lockwood, & Williams, 2004).

Moderation by Gender

To examine potential moderation of the direct and indirect effects by gender, we conducted multi-group models. The models were identical to those outlined above with the exception that gender was removed from the models as a predictor and instead included as a grouping variable. We tested two separate multi-group models. The first model was run simultaneously for men and women but the relations among study variables were allowed to freely vary by gender (no constraints). In the second model, the four structural paths (Wave 1 drinking to Wave 10 drinking, Wave 1 religiosity to Wave 10 drinking, Wave 1 religiosity to age of self-identification, and age of self-identification to Wave 10 drinking) were constrained to equality for men and women. We then compared the fit of the two models (change in chi-square) to determine if constraining the parameters by gender resulted in a decrement in model fit. A significant decrement in model fit would suggest gender moderation. In the event that the difference in model fit was statistically significant, we planned follow-up single degree of freedom tests to determine the specific path(s) that differed by gender.

Results

The hypothesized model testing direct and indirect effects of religiosity on alcohol use in the full sample provided adequate fit to the data, χ2 (2 df) = 3.548, p = .17; RMSEA = 0.07; CFI = .92; SRMR = .03. With respect to the covariates, gender did not significantly predict wave 10 drinking (b = .09, SE = .11, p = .41) but heavier drinking at wave 1 was associated with heavier drinking at wave 10 (b = .37, SE = .14, p = .006). The direct effect of religiosity on wave 10 drinking was not statistically significant (b = −.03, SE = .07, p = .70). As depicted in figure 1 and of central interest to the study, religiosity significantly predicted age of self-identification (b = .82, SE = .31, p = .010), and age of self-identification significantly predicted wave 10 drinking (b = .04, SE = .02, p = .02). Individuals who were higher in religiosity reported a later age of self-identification as LGB which, in turn, was associated with heavier drinking. The formal test of the indirect effect of religiosity on wave 10 drinking through age of self-identification was statistically significant based on the bias-corrected bootstrapped confidence interval, which did not contain the value of zero (CI [.005, .088]).1

Figure 1.

Figure 1.

Open in a new tab

Path model demonstrating indirect effects of religiosity on alcohol use operating through age of self-identification. Regression coefficients are standardized. * p < .05.

Moderation by Gender

When the model was tested simultaneously for men and women without any constraints on the parameters (allowing paths to freely vary by gender), model fit was good, χ2 (4 df) = 4.74, p = .31; RMSEA = 0.05; CFI = .96; SRMR = .05. When the structural paths were constrained to equality by gender, the model fit was excellent, χ2 (8 df) = 8.01, p = .43; RMSEA = 0.01; CFI = 1.00; SRMR = .06. The test of the difference in Chi-Square between the two models was not statistically significant, Δχ2 (df = 4) = 3.27, p = .51, suggesting that relations among the study variables did not differ significantly by gender. Given the relatively small sample size for testing gender moderation, we conducted post-hoc single degree of freedom tests for each of the primary paths of interest (all paths other than the autoregressive effect of Wave 1 drinking on Wave 10 drinking). No significant gender differences were found for any of these paths (all p values > .16), confirming the lack of gender moderation.

Discussion

The purpose of this study was to examine relations among religiosity, age of identification, and alcohol use in a sample of LGB college students. We conceptualized later age of self-identification as a marker of stress related to the conflict between religiosity and identification as LGB or stress associated with the LGB self-identification process. With this conceptualization, we proposed that religiosity would indirectly impact later alcohol use through a later age of self-identification. More specifically, we hypothesized that students who were high in religiosity would report a later age of identification as LGB. This delay in age of self-identification was expected to predict heavier drinking during the college years.

Consistent with this hypothesis, higher levels of religiosity were associated with a later age of self-identification which, in turn was associated with greater increases in alcohol use from high school through the college years. This pattern of results is consistent with the notion that detrimental effects of religiosity in LGB youth may be related to struggles associated with self-identification. We found no evidence that relations among the study variables differed by gender in the multi-group model, despite prior evidence in this data set for greater baseline drinking among LGB women and greater increases in drinking during college among LGB men relative to heterosexual peers. This suggests that factors other than age of identification and religiosity may account for gender differences in patterns of heavy drinking in this sample. However, future studies with larger samples of LGB youth will be important given the relatively small sample size in the current study, and prior evidence that differences in substance use between sexual minority and heterosexual individuals are considerably larger for women than for men (Hughes, Wilsnack, & Kantor, 2016; Marshall et al., 2008).

Although prior research has shown that religiosity is a protective factor in the general population, our findings suggest that LGB youth may fail to receive the same benefits that others experience. It may be that some LGB youth are struggling to reconcile religious beliefs with their emerging sexual identity due to attitudes within some religious organizations that are unsupportive and at times actively opposed to homosexuality (Mahaffy, 1996; Thumma, 1991; Perez et al., 2000). This could be a foundation for internalized homophobia and shame (Allen, 1999). For example, a study examining mental health outcomes in same-sex attracted youth as a function of views about homosexuality within their religious organizations (accepting versus non-accepting) discovered that those whose religion made it difficult to accept their sexual identity reported higher internalized homophobia. In addition, religious conflicts explained a large percentage of the variance (32%) in internalized homophobia (Ream & Savin-Williams, 2005).

Evidence for an indirect effect of religiosity on alcohol use through age of self-identification may be informative in the development of efforts to help address the struggles of LGB youth related to conflict between their emerging sexual identity and their religious beliefs and experiences. One aspect of this might include having faith-based groups that specifically focus on LGB youth and their emerging development of sexual identity. Such programs might help LGB youth to manage the stress and perceived shame associated with their sexual identity while also maintaining a healthy connection with religious organizations (Rodriguez & Ouellette, 2010; Wolkomir, 2001). Having faith-based groups that are primarily for LGB youth in the community may also decrease the negative stigma that is embedded in the LGB label within some religious organizations. Although faith-based groups would have a number of advantages, this may not currently be realistic within some religious organizations. Because some institutions may not sanction LGB youth groups or because youth may not feel supported and safe within those spaces, interventions may need to target parents within religious organizations and religious organizations themselves to address negative attitudes towards the LGB community. These interventions may take the form of psychoeducation in which parents and religious leaders learn about the experiences of LGB youth and how their environment impacts emotional and psychological well-being. In this way, youth may experience structural shifts in how they are perceived and treated by their religious communities and therefore have more opportunity to benefit from some of the well-studied protective factors that religion and its organizations can provide. In addition, LGB youth might form their own groups outside of their religious organizations to help them cope with their emotions and the process of coming out in the context of their religious beliefs. Such groups would give LGB youth a sense of support and community among at least a subset of members within their religious organizations.

Although our conceptual model considers religiosity as a predictor of age of self-identification, and our analyses demonstrated this temporal ordering, it is not possible to make claims about causation given the correlational nature of the data. It is possible that youth who are highly religious hold more heteronormative notions of sexuality and/or have little to no exposure to the LGBTQIA community until they are more autonomous. In this case, some highly religious youth may not feel able to explore their sexual identity, or even have the understanding of a sexual minority orientation to start considering this possibility. This may lead to a later self-identification relative to non-religious peers in the absence of substantial stress associated with this process. Future studies might include questions about when youth began wondering about their sexual identity in addition to when they actually identified in order to create a more nuanced picture of religious youths’ experiences of self-identifying as LGB. Future studies might also compare these questions among religious versus non-religious youth. It is also possible that youth coming from more religious backgrounds and home environments may enter into risky drinking during college because they have been relatively sheltered from these experiences and are experiencing increased independence and access to alcohol upon college entry. In summary, although the findings are consistent with our conceptualization of stress associated with the self-identification process as a critical mechanism linking religiosity with alcohol use, and there is strong support for minority stress as a risk factor for substance use (Goldbach et al., 2014; McCabe et al., 2010), several alternative explanations for the lack of protective effects of religiosity among LGB youth are plausible.

The results of the current study must also be considered in light of several methodological limitations. First, there was a relatively small sample of LGB youth despite the fact that the initial data collection involved 2245 incoming college students. A total of 162 participants identified as LGB at some point during the study. This provided modest power for detecting study effects and also calls into question the extent to which the sample was representative of all LGB youth. The fact that the sample comprised primarily white college students from a single campus also limits generalizability to LGB college students more broadly. Studies with larger and more diverse samples are critical, as prior studies have demonstrated the importance of examining racial/ethnic group differences in relations between religiosity/spirituality and substance use (Drabble et al., 2018). Although the sample included relatively equal numbers of men and women, the sample was not large enough to examine differences across unique sexual orientations (e.g., gay men, lesbian women, bisexual men and women). This is an important limitation as bisexual youth have been identified as particularly high risk for substance misuse (Marshall et al., 2008). Thus, relations among religiosity, age of identification, and heavy drinking may differ across distinct LGB groups.

Our exclusive focus on self-identity is another potential limitation as other aspects of sexual orientation (same-sex attraction and same-sex behavior) may be important in understanding relations between religiosity and substance use. Although prior work suggests more consistent links between sexual identity and substance use relative to other indicators of sexual orientation (Marshal et al., 2009; Talley et al., 20010), there is also evidence that youth whose same-sex attractions are in strong conflict with their religion report high levels of internalized homophobia (Ream & Savin-Williams, 2005). Thus, future studies using a broader assessment of sexual orientation are needed. It is also important to note that students who had already self-identified as LGB at the time of the first assessment (42%) were given the same value. Thus, we could not distinguish between a student who self-identified at 13, for example, and one who self-identified at age 18 (Wave 1). Such different ages of self-identification may be associated with important differences in family and peer contexts during adolescence. Also, for those that identified as LGB at the final wave (Wave 10), the relation between self-identification and alcohol use at Wave 10 was cross-sectional. Further, because we assessed age of self-identification up through the final wave of data collection, it was not possible to examine whether self-identification was associated with changes in religiosity, which might be related to changes in drinking behavior. This is an important question to address in future research.

It is also important to recognize that religiosity was measured using two single items that could not fully capture the multi-dimensional nature of the construct. This was a function of the broad focus of the study on predictors of a range of risky behaviors during the transition to college, which precluded comprehensive assessment of all constructs of interest. A more comprehensive measure of religiosity (e.g., the Brief Multidimensional Measure of Religiousness/Spirituality among Adolescents) would have provided better insight into potentially complex relations between religiosity and alcohol use among LGB youth (Harris et al., 2008). For example, Allport & Ross (1967) proposed two main types of religious orientations, intrinsic and extrinsic religiosity. Intrinsic religiosity is a way of living in one’s religion and affects more areas of life than just the religious aspect. In contrast, extrinsic religiosity is a way of using religiosity to gain social status, group protection, and consolation. In theory, a person with high intrinsic religiosity can be described as wholly committed to their religion in every aspect of their life including views about sexual orientation within their religion. Therefore, a LGB person who is high in intrinsic religiosity may be more likely to come out at a later age due to the internalization of religious beliefs related to homosexuality. In contrast, individuals high in extrinsic religiosity may be less likely to live by their religion and may be more likely to come out at an earlier age relative to people who are high in intrinsic religiosity. However, it is also possible that those high in extrinsic religiosity have other motives for delaying age of identification, such as fearing loss of social status or familial reprimands. Additional studies with measures of intrinsic and extrinsic religiosity are needed to explore these more these nuanced questions about relations between religiosity and alcohol use in LGB samples.

As discussed previously, age of self-identification was conceptualized in this study as a proxy for stress associated with conflict between sexual identity and religious beliefs. Although the results are consistent with this interpretation, we did not have the ability to confirm an association between age of self-identification and stress associated with the self-identification process. It is possible that age of self-identification is not the most accurate proxy, particularly in assessing risk for heavy alcohol use. Future studies should attempt to specifically assess stress that is directly associated with the process of self-identification. This could be accomplished by asking about stress the individual associates with self-identification in the context of religious beliefs or by developing a new measure that specifically assesses stressors associated with coming out in the context of religiosity.

Finally, it should be acknowledged that, at the time of data collection, our terminology for gender and sexual minorities was limited and we did not have comprehensive measures of multiple identities. Therefore, our results are not generalizable beyond individuals who identify as lesbian, gay, bisexual, or questioning. Guidelines set forth by the American Psychological Association such as Guidelines for Psychological Practice with Lesbian, Gay, and Bisexual Clients and Guidelines for Psychological Practice with Transgender and Gender Nonconforming People and (APA, 2011, 2015) and the World Professional Organization for Transgender Health’s (WPATH) Standards of Care (Version 7, 2012) suggest that demographic information gathered for research purposes capture an inclusive range of gender and sexual identities. Today, our measures would have included those who identify as Lesbian, Gay, Bisexual, Transgender, Queer, Intersex, and Asexual (LGBTQIA). Although gender and sexual identities are distinct constructs, it is important that future work strive to understand the impact of these intersecting identities, as well as how they may intersect with other aspects of identity such as race, ethnicity and socioeconomic status.

Future studies should also attempt to examine moderators of the relationships observed in the current study. It may be that some LGB youth benefit from religion in the same ways as heterosexual youth. If so, future research could identify characteristics of these individuals that serve to protect them, thereby identifying potential targets of intervention for other LGB youth. In addition, samples of adolescents should be included in future studies so that the impact of variability in age of self-identification earlier in development can be examined in relation to religiosity and drinking behavior.

Despite the limitations discussed above, we found that self-identification was an important intervening variable between religiosity and changes in drinking behavior over time within a sample of LGB youth. The fact that these findings emerged even with a relatively crude measure of stress associated with self-identification and a modest sample size suggests that this may be a robust phenomenon and that, with more specific measures and larger samples, even stronger effects may emerge. Nonetheless, this study is merely a first step in understanding how religiosity may affect drinking behavior among LGB youth. In a broader context, more comparative work on sexual minority and heterosexual youth would be helpful in improving our understanding of the impact of sexual orientation on relations between religiosity and drinking behavior. Examination of additional factors that we did not test in the current study (race, class, guilt, shame) will also serve to advance this important area of research.

Acknowledgments

This research was supported by grants 1 R01 AA013967 and 1 R01 AA020637 from the National Institute on Alcohol Abuse and Alcoholism. The NIAAA played no role in manuscript writing, preparation, or interpretation of the results. The content is solely the responsibility of the authors and does not necessarily represent the views of the NIAAA or the National Institutes of Health.

Results of this study were previously presented as a poster at the 37th Annual Meeting of the Research Society on Alcoholism in Bellevue Washington.

Footnotes

We ran a sensitivity analysis using the two items comprising the religiosity composite. The results were similar though the effect of religious beliefs on age of self-identification (standardized coefficient = .23, p = .004) was stronger than the effect of religious attendance (standardized coefficient = .11, p = .16). The magnitude of the path from age of self-identification to wave 10 drinking was identical and statistically significant across the two models (standardized coefficients = .21).

Contributor Information

William R. Corbin, Arizona State University

Thai Q. Ong, Arizona State University

Charlie Champion, Arizona State University.

Kim Fromme, The University of Texas at Austin.

References

  1. Allen DJ (1999). Shame and internalized homophobia in gay men. Journal of Homosexuality, 37, 33–43. [DOI] [PubMed] [Google Scholar]
  2. Allport GW, & Ross JM (1967). Personal religious orientation and prejudice. Journal of Personality and Social Psychology, 5, 432–443. [DOI] [PubMed] [Google Scholar]
  3. Baron RM, & Kenny DA (1986). The moderator-mediator variable distinction in social psychological research: Conceptual, strategic and statistical considerations. Journal of Personality and Social Psychology, 51, 1173–1182. [DOI] [PubMed] [Google Scholar]
  4. Barry AE, Valdez D, & Russell AM (2019) Does religiosity delay adolescent alcohol initiation? A long-term analysis (2008–2015) of nationally representative samples of 12th Graders. Substance Use & Misuse, DOI: 10.1080/10826084.2019.1686024 [DOI] [PubMed] [Google Scholar]
  5. Camacho G, Reinka MA, & Quinn DM (2019). Disclosure and Concealment of Stigmatized Identities. Current Opinion in Psychology, 1–23 [DOI] [PubMed] [Google Scholar]
  6. Charbonnier E, & Graziani P. (2016). The stress associated with the coming out process in the young adult population. Journal of Gay & Lesbian Mental Health, 20, 319–328. [Google Scholar]
  7. Clark JM, Brown JC, & Hochstein LM (1990). Institutional religion and gay/ lesbian oppression. Marriage and Family Review, 14, 265–284. [Google Scholar]
  8. Corliss HL, Rosario M, & Wypij D, Fisher LB, & Austin B. (2008). Sexual orientation disparities in longitudinal alcohol use patterns among adolescents: Findings from the Growing Up Today study. Archives of Pediatric and Adolescent Medicine, 162, 1071–1078. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. D’Augelli AR (2002). Mental health problems among lesbian, gay, and bisexual youths ages 14 to 21. Clinical Child Psychology and Psychiatry, 7, 433–456. [Google Scholar]
  10. D’Augelli AR (2006). Developmental and contextual factors and mental health among lesbian, gay, and bisexual youths. In Omoto A. & Kurtzman HS (Eds.), Sexual orientation and mental health: Examining identity and development in lesbian, gay, and bisexual people. Contemporary Perspectives on Lesbian, Gay, and Bisexual Psychology (pp. 37–53). Washington: American Psychological Association. [Google Scholar]
  11. Dermody SS, McGinley J, Eckstrand K, & Marshal MP (2019). Sexual minority female youth and substance use disparities across development. Journal of LGBT Youth. DOI: 10.1080/19361653.2019.1598313 [DOI] [Google Scholar]
  12. Drabble L, Trocki KF, & Klinger JL (2016). Religiosity as a protective factor for hazardous drinking and drug use among sexual minority and heterosexual women: Findings from the National Alcohol Survey. Drug and alcohol dependence, 161, 127–134. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Drabble L, Veldhuis CB, Riley BB, Rostosky S, & Hughes TL (2018). Relationship of religiosity and spirituality to hazardous drinking, drug use, and depression among sexual minority women. Journal of Homosexuality, 65, 1734–1757. 10.1080/00918369.2017.1383116 [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Eliason MJ, Burke A, van Olphen J. and Howell R. (2011). Complex interactions of sexual identity, sex/gender, and religious/spiritual beliefs on college student substance use. Sexuality Research and Social Policy, 8, 117–135. [Google Scholar]
  15. Ream Geoffrey L. & Savin-Williams Ritch C. (2005). Reconciling Christianity and positive non-heterosexual identity in adolescence, with implications for psychological well-being, Journal of Gay & Lesbian Issues in Education, 2, 19–36. [Google Scholar]
  16. Goldbach JT, Tanner-Smith EE, Bagwell M, & Dunlap S. (2014). Minority stress and substance use in sexual minority adolescents: A meta-analysis. Prevention Science, 15, 350–363. [DOI] [PubMed] [Google Scholar]
  17. Halpin SA, & Allen HMW (2004). Changes in psychological well-being during stages of gay identity development. Journal of Homosexuality, 47, 109–126. [DOI] [PubMed] [Google Scholar]
  18. Harris SK, Sherritt LR, Holder DW, Kulig J, Shrier LA, & Knight JR (2008). Reliability and validity of the brief multidimensional measure of Religiousness/Spirituality among adolescents. Journal of Religion and Health, 47, 438–457. [DOI] [PubMed] [Google Scholar]
  19. Harrison TW (2003). Adolescent homosexuality and concerns regarding disclosure. Journal of School Health, 73, 107–112. [DOI] [PubMed] [Google Scholar]
  20. Hatzenbuehler ML, Corbin WR, & Fromme K. (2008). Trajectories and determinants of alcohol use among LGB young adults and their heterosexual peers: Results from a prospective study. Developmental Psychology, 44, 81–90. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Hughes TL, Wilsnack SC, & Kantor LW (2016). The influence of gender and sexual orientation on alcohol use and alcohol-related problems: Toward a global perspective. Alcohol Research: Current Reviews, 38, 121–132. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Jackson KM, Sher KJ, Gotham HJ, & Wood PK (2001). Transitioning into and out of large-effect drinking in young adulthood. Journal of Abnormal Psychology, 110, 449–461. [DOI] [PubMed] [Google Scholar]
  23. Lasala MC (2000). Gay male couples: The importance of coming out and being out to parents. Journal of Homosexuality, 39, 47–71. [DOI] [PubMed] [Google Scholar]
  24. MacKinnon DP, Lockwood CM, Hoffman JM, West SG, & Sheets VA (2002). Comparison of methods to test the significance of mediation and other intervening variable effects. Psychological Methods, 7, 83–104. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. MacKinnon DP, Lockwood CM, & Williams J. (2004). Confidence limits for the indirect effect: Distribution of the product and resampling methods. Multivariate Behavioral Research, 39, 99–128. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. MacKinnon DP, Fritz MS, Williams J, & Lockwood CM (2007). Distribution of the product confidence limits for the indirect effect: Program PRODLIN. Behavior Research Methods, 39, 384–389. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Mahaffy KA (1996). Cognitive dissonance and its resolution: A study of lesbian Christians. Journal for the Scientific Study of Religion, 35, 392–402. [Google Scholar]
  28. Marshal MP, Friedman MS, Stall R, King KM, Miles J, Gold MA et al. (2008). Sexual orientation and adolescent substances use: a meta-analysis and methodological review. Addiction, 103, 46–56. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Marshal MP, Friedman MS, Stall R, & Thompson AL (2009). Individual trajectories of substance use in lesbian, gay and bisexual youth and heterosexual youth. Addiction, 104, 974–981. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. McCabe SE, Bostwick WB, Hughes TL, West BT, & Boyd CJ (2010). The relationship between discrimination and substance use disorders among lesbian, gay, and bisexual adults in the United States. American Journal of Public Health, 100, 1946–1952. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Meyer IH (2003). Prejudice, social stress, and mental health in lesbian, gay, and bisexual populations: conceptual issues and research evidence. Psychological Bulletin, 129, 674–697. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Mohr JJ, Markell HM, King EB, Jones KP, Peddie CI, & Kendra MS (2019). Affective antecedents and consequences of revealing and concealing a lesbian, gay, or bisexual identity. Journal of Applied Psychology, 1–18. [DOI] [PubMed] [Google Scholar]
  33. Monroe EJ (2001). Drawing upon the experiences of those who are out: A qualitative study of the coming-out process of gays and lesbians. Dissertation Abstracts International: Section B: The Sciences and Engineering, 61(12-B), 6739. [Google Scholar]
  34. Mohr JJ, & Fassinger RE (2003). Self-acceptance and self-disclosure of sexual orientation in lesbian, gay, and bisexual adults: An attachment perspective. Journal of Counseling Psychology, 50, 482–495. [Google Scholar]
  35. Muthén LK, & Muthén BO (1998–2013). Mplus user’s guide (6th Ed.). Los Angeles, CA: Author. [Google Scholar]
  36. Newcomb ME, Heinz AJ, & Mustanski B. (2012). Examining risk and protective factors for alcohol use in lesbian, gay, bisexual, and transgender youth: A longitudinal multilevel analysis. Journal of Studies on Alcohol and Drugs, 73, 783–793 [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Perez RM (Ed); DeBord KA (Ed); Bieschke KJ (Ed), (2000). Handbook of Counseling and Psychotherapy with Lesbian, Gay, and Bisexual Clients, (pp. 409–433). Washington, DC, US: American Psychological Association, xvi, 484 pp. [Google Scholar]
  38. Pilkington NW, & D’Augelli AR (1995). Victimization of lesbian, gay, and bisexual young in community settings. Journal of Community Psychology, 23, 34–56. [Google Scholar]
  39. Rew L, & Wong YJ (2006). A systematic review of associations among religiosity/spirituality and adolescent health attitudes and behaviors. Journal of Adolescent Health, 38, 433–442. [DOI] [PubMed] [Google Scholar]
  40. Rodriguez EM and Ouellette SC (2000), Gay and Lesbian Christians: Homosexual and Religious Identity Integration in the Members and Participants of a Gay-Positive Church. Journal for the Scientific Study of Religion, 39, 333–347. [Google Scholar]
  41. Ross MW, & Rosser BRS (1996). Measurement and correlates of internalized homophobia: A factor analytic study. Journal of Clinical Psychology, 52, 15–21 [DOI] [PubMed] [Google Scholar]
  42. Rostosky SS, Danner F, & Riggle EDB (2007). Is religiosity a protective factor against substance use in young adulthood? Only if you’re straight! Journal of Adolescent Health, 40, 440–447. [DOI] [PubMed] [Google Scholar]
  43. Rostosky SS, Danner F, & Riggle EDB (2010). Religiosity as a protective factor against heavy episodic drinking (HED) in heterosexual, bisexual, gay, and lesbian young adults. Journal of Homosexuality, 57, 1039–1050. [DOI] [PubMed] [Google Scholar]
  44. Rostosky SS, Danner F, & Riggle EDB (2008). Religiosity and alcohol use in sexual minority and heterosexual youth and young adults. Journal of Youth and Adolescence, 37, 552–563. [Google Scholar]
  45. Savin-Williams RC, & Ream GL (2003). Suicide attempts among sexual-minority male youth. Journal of Clinical Child and Adolescent Psychology, 32, 509–522. [DOI] [PubMed] [Google Scholar]
  46. Solomon D, McAbee J, Asberg K, & McGee A. (2015). Coming out and the potential for growth in sexual minorities: The role of social reactions and internalized homonegativity. Journal of Homosexuality, 62, 1512–1538. [DOI] [PubMed] [Google Scholar]
  47. Stall R, Paul JP, Greenwood G, Pollack LM, Bein E, Crosby GM, Mills TC, Binson D, Coates TJ, & Catania JA (2001). Alcohol use, drug use and alcohol-related problems among men who have sex with men: The Urban Men’s Health Study. Addiction, 96, 1586–1601. [DOI] [PubMed] [Google Scholar]
  48. Talley AE, Sher KJ, & Littlefield AK (2010). Sexual orientation and substance use trajectories in emerging adulthood. Addiction, 105, 1235–1245. [DOI] [PMC free article] [PubMed] [Google Scholar]
  49. Thumma S. (1991). Negotiating a religious identity: The case of the gay evangelical. Sociological Analysis, 52, 333–347. [Google Scholar]
  50. Valera P, & Taylor T. (2011). “Hating the sin but not the sinner”: A study about heterosexism and religious experiences among Black men. Journal of Black Studies, 42, 106–122. [DOI] [PubMed] [Google Scholar]
  51. Wallace JM Jr., Yamaguchi R, Bachman JG, O’Malley PM, Schulenberg JE, & Johnston LD (2007). Religiosity and adolescent substance use: The role of individual and contextual influences. Social Problems, 54, 308–327. [Google Scholar]
  52. Weber GN (2008). Using to numb the pain: Substance use and abuse among lesbian, gay and bisexual individuals. Journal of Mental Health Counseling, 30, 31–48. [Google Scholar]
  53. Wechsler H, Davenport A, Dowdall G, Moeykens B, & Castillo S. (1994). Health and behavioral consequences of binge drinking in college. A national survey of students at 140 campuses. Journal of the American Medical Association, 272, 1672–1677. [PubMed] [Google Scholar]
  54. Wills TA, Yaeger AM, & Sandy JM (2003). Buffering effect of religiosity for adolescent substance use. Psychology of Addictive Behaviors, 17, 24–31. [DOI] [PubMed] [Google Scholar]
  55. Wolkomir Michelle. 2001. Emotion work, commitment, and the authentication of the self: The case of gay and ex-gay support groups. Journal of Contemporary Ethnography 30, 305–34. [Google Scholar]

RESOURCES