Introduction

The institution of religion has significantly influenced conceptualizations of love, relationships, sexual attitudes and behaviors (Barkan 2006; Baker et al. 2015; Burdette et al. 2007; Njus and Bane 2009; Perry 2016; Soloski et al. 2013). Religion also functions to create a personal connection to a higher power (Iles et al. 2016) and serves as an important socialization agent or an individual’s values, choices, and behaviors (Barkan 2006). However, the influence of religion is likely not the same for every religious individual due to varying contextual and identity factors. The focus of this study was to investigate how emerging adults’ racial and sexual identities differentiate associations between self-identified religiosity and beliefs/behaviors related to love, relationships, and sexuality.

Differential Associations of Religion by Identity

Being religious is often associated with more conservative attitudes and behavior, but the association is not straight forward. One’s gender and racial identities, particularly when associated with impactful historical and social contexts, could contribute to religion’s influence on beliefs and behaviors. For example, religious attitudes have been more predictive of sexual attitudes for women than for men (Ahrold and Meston 2010; Owen et al. 2010). Higher intrinsic religiosity and religious fundamentalism have been associated with sexual conservativism for Asian and Euro-American female students but not for female students of other races (Ahrold and Meston 2010). Being religious has been less of a deterrent of drug use for Black teens than teens of other races (Amey et al. 1996; Taylor et al. 2011; Wallace et al. 2003). Similarly, studies have generally shown that religiosity or church attendance corresponds less with conservative sexual views and behaviors for Black adolescents than for other adolescents (Billy et al. 1994; Brewster et al. 1998; Rostosky et al. 2003).

Some researchers have attributed such racial divergence to general differences between white and black churches, the latter being more communal and forgiving (Lincoln and Mamiya 1990). Also, racially-insensitive or biased attitudes and expectations within mainstream religious contexts can lead to racial minorities feeling some disconnection with a larger, predominantly white religious community (Kim 2017). Consequently, it is possible that the general impact of more traditional, White European-influenced religious perspectives are perceived as less salient by racial minorities. Overall, the reviewed studies support the notion that associations broadly between religion and beliefs or behaviors related to love and sexuality could vary by race (and gender).

One’s sexual identity could also be relevant to the impact of religion on one’s beliefs and behavior. Religious LGB (lesbian, gay, and bisexual) individuals have reported that religion provides a source of love, support, strength, and a sense belonging to something greater than themselves (Rosenkrantz et al. 2016). However, more conservative religious beliefs are often associated with greater feelings of guilt and shame for sexual minorities (Sherry et al. 2010). This is due to the fact that religions in general have been known for promoting negative attitudes toward homosexuality (Yip 2005), and in countries like the United States, being religious is especially influential on negative beliefs people hold toward LGB issues (Adamczyk and Pitt 2009). Thus, being religious and being sexual minority may result in many sexual minorities abandoning their religion (Jeffries et al. 2008).

National data indicates that sexual minorities are indeed less likely to be religious (Sherkat 2016); those who are and remain religious are likely to have found a way to reconcile their sexual and religious identities. Specifically, LGB religious individuals are challenged to find ways to integrate their religious identity—that is often associated with anti-LGB sentiments—with their sexual identity—that is often assumed to be, and portrayed in mainstream culture, as anti-religious (Barton 2012; Sumerau et al. 2016). Some who leave and then return (or change to a different) religious community feel pressure to defend their actions to those LGB individuals who interpret such actions as hostile toward LGB identities (Sumerau 2014). Caught between seemingly incompatible subcultures, LGB individuals and communities sometimes adapt by reinventing religious rituals and beliefs and interpret scriptural passages in distinctly different ways that accommodate LGB experiences (Thumma and Gray 2004). Considered in toto, the strategies used to integrate religious and LGB identities could result in religion having a somewhat unique type of impact on LGB individuals’ attitudes and behaviors.

Some research has focused on the intersection of race and sexual identity in the context of religious faith. For example, Black adults who identified as LGB reported higher levels of internal conflict when trying to reconcile their sexual identity with their religion than individuals with a single minority identify (Walker and Longmire-Avital 2013). Black men who identified as both religious and gay had less church attendance and identified as less religious than Black, heterosexual men (Pew Research Center 2009). Hence, accounting for both race and sexual identity could help differentiate associations between being religious and other values, intentions, or behaviors.

The current study is based on the premise that religion influences attitudes and beliefs that subsequently influence intentions and behavior, consistent with Social control theory. For example, social control theory suggests that institutionalized, social norms influence whether individuals act on sexual drives (DeLamater 1981; Janowitz 1975). Research on adolescent sexual behavior found that religiosity largely influenced the internalization of conservative sexual norms that became integrated with their motivation for sexual behavior (Vasilenko et al. 2013). As suggested by the studies reviewed above, religion may intersect with other social influences idiosyncratic to specific ethnic or sexual-identity type cultures or contexts, resulting in differential associations between being religious and sexual behaviors. It is possible that for similar reasons religion is uniquely linked to various beliefs and expectations in the realm of romance and love based on identity group factors. The current study extends the existing literature by examining a broader array of relationship values and behaviors that potentially associate with being religious. Furthermore, the research illustrates how the influence of religion can differ based one’s race and sexual identity. Such findings add nuance to understanding how religion impacts the lives of contemporary, emerging adults, and demonstrates the importance of accounting for identity issues when studying the impact of religion.

Methods

Survey and Sample

Data for this study were generated from the “College Student Attitudes and Behaviors Survey” developed by the first and third authors consisting of 100 questions including demographic characteristics, attitudes toward love, relationships, and sexual values/behaviors. The sample consisted of a total of 6068 undergraduate students over a period of 8 years from two large universities in the United States–one in the Midwest and the other in the Southeast. Students in the undergraduate classes of the first and third authors were emailed a link to an anonymous online questionnaire and invited to voluntarily participate (no compensation, grade or otherwise, was provided). Colleagues of the respective researchers also sent the link to students in their introductory classes.

Of the 6068 undergraduates (mean age = 19.91) who completed the survey, 55% were female, 45% male. Over 80% (82.4%) were White, 7.9% Black, 3.9% Hispanic and 5.9% Other. Regarding sexual identity, 79% were heterosexual; 21% Other (gay/lesbian or bisexual). Over half (54%) reported being partnered. Because the specific religious denominations of the study participants are unknown, it is important to provide some context for the nature of the broad religious influences on the participants. We note that in one of the two states (Indiana) in which the two institutions are located, the population is predominantly Christian (72%), with the main subgroups breaking down as follows: 31% Evangelical Protestant, 18% Catholic, 16% Mainline Protestant, and 5% Historically Black Protestant (Pew Research Center 2014). In the other state (North Carolina), the population is likewise predominantly Christian (77%), with the main subgroups breaking down as follows: 35% Evangelical Protestant, 19% Mainline Protestant, 12% Historically Black Protestant, and 9% Catholic (Pew Research Center 2014). It is thus assumed that the religious contexts in which the individuals navigate their racial and sexual identities are largely influenced by relatively mainstream Christian beliefs systems and practices, though the denominations among the participants could have greater variation.

Measures

Dependent Variable

Respondents identified their religiosity along a continuum including very religious (5), moderately religious (4), about midway (3), moderately not religious (2) and not religious at all (1).

Background Variables

Dummy variables were created for race, divorced, parental divorce, and gender/sexual identity. Sexual identity was determined by the response to “I consider myself:” followed by a list of options (heterosexual, bisexual, gay male, gay female, other). Based on the reported gender, respondents were categorized as: heterosexual male, heterosexual female, bisexual male, bisexual female, gay male, or gay female. Other sexual identities (e.g., asexual, questioning) were not included due to the small number of respondents.

Relationship Beliefs and Intentions

Several survey items captured beliefs and intentions related to love and relationships. The following items were responded to with a 5-point agreement scale (1 = strongly disagree, 5 = strongly agree): “I would divorce my spouse if I no longer loved him or her,” and “I would end a relationship with a partner who cheated on me.” The remaining items were responded to with a yes or no: “I believe that there is only one true love that never comes again,” “I would live with a partner I was not married to,” and “I have looked for a partner on the Internet.”

Beliefs and Behaviors Related to Sexuality

Several items captured elements of sexually-related perspectives and behavior. The following items were responded to with a 5-point agreement scale (1 = strongly disagree, 5 = strongly agree): “An abortion is acceptable under certain conditions,” and “It is better to be heterosexual than homosexual.” The other items were responded to with a yes or no: “I have had sex without love,” “I have hooked up (sex upon first time meeting each other),” “I have been in a ‘friends with benefits’ relationship,” and “I have cheated on a partner I was involved with.”

Results

Correlational Findings

The first analysis focused on general associations between background variables with religiosity. Significant mean differences were found regarding religiosity and race, sexual identity, and having divorced parents. For race, a one-way analysis of variance (ANOVA) revealed a significant difference (p < .001) in religiosity with blacks having a higher mean (3.42) compared to whites (2.79), and Hispanics having a lower religiosity mean (2.53) than both blacks and whites. When sexual identity was dichotomized, heterosexuals reported a higher mean (3.00) than non-heterosexuals (2.20; (p < .001)). When compared across more precise gender/sexual identity categories (p < .001), heterosexual females were the most religious (3.12), followed by heterosexual males (2.88). Gay males were more religious than bi females (2.44 vs. 2.10, p < .05). No significant differences were found in religiosity between men and women as a whole (note that non-heterosexual females were oversampled) or age of respondent. Those whose parents were still married were more religious than those whose parents were divorced (2.88 vs. 2.76, p < .001).

The researchers also tested whether certain sexual identities were overrepresented within certain racial categories. A Chi Square analysis compared the frequencies of responses across different groups and revealed that the six gender by identity categories were evenly distributed among the four racial groups. Thus, no racial group was more disproportionately heterosexual (or other such identity) than the others.

Correlational analyses tested for patterns between being religious and endorsing certain beliefs of having had certain experiences. Being religious was significantly and negatively correlated with an unwillingness to divorce if one fell out of love (r = − .316, p < .001), revealing a connection between a high level of self-identified religiosity and a strong commitment to marriage. Religion often encourages lifetime commitment (“until death do us part”) and falling out of love is not viewed as an acceptable reason for divorce. Agreeing with the statement “I believe that there is only one true love that never comes again” corresponded with being more religious [F(1, 5835) = p < 0.001]. Religion can encourage the idea that love is destined and that one may be destined to have only one true love in a lifetime. Praying for one’s soul mate reflects a belief that there is one soul mate per person. Being more religious was also slightly correlated with less willingness to end a relationship with a cheating partner (.05, p < 0.001), perhaps reflecting the value of forgiveness.

A willingness to live with a non-marital partner was also lower for religious respondents [F(1, 5957) = 994.64, p < 0.001]. Religion encourages individuals to avoid premarital sex/cohabit with someone before marriage. In the fifties, persons who lived together before marriage were referred to as “Living in sin.” Those respondents who were religious were also less likely to have looked for a partner on the Internet [F(4, 5992) = 112.99, p < 0.001]. Religion may encourage individuals to look to divine sources for one’s partner (e.g., “I prayed to God to send me my soulmate”) rather than rely on technology which suggests one’s life partner is not “Heaven sent” or “Divinely selected” (within one’s own group at church) but waiting to be met among the profiles of Match.com. In addition, seeking a person on the Internet may be viewed by persons of high religiosity as “sinful” due to the association of finding someone to “hook up with.”

Several variables related to sexuality also differed by religiosity. Higher levels of religiosity were related to less approval for abortion (r = − .467, p < .001). There was a strong correlation between higher levels of religiosity and agreement that heterosexuality was better than homosexuality (r = .40, p < .001). With rare exception (Whitehead 2017), religion has traditionally been against being gay (such a lifestyle choice, according to the Bible, is regarded as an “abomination”).

Higher levels of religiosity were also associated with not having been in a “friends with benefits” relationship [F(1, 5949) = 99.21, p < 0.001]. Friends with benefits breaks two religious, moral barriers–having sex before marriage and having sex with someone who is not a spouse but a “friend” (hence, no love). Similarly, students scoring high on religiosity were less likely to have had sex without love [F(1, 6037), p < 0.001] and less likely to have hooked up with a stranger [F(1, 5946) = 55.121, p < 0.001]. Religion often encourages the idea that sex is sacred and should not be experienced devoid of feelings for the person. Indeed, love should be the foundation for sex. Finally, being religious was slightly correlated with not having cheated on a partner (r = .09, p < .001). Religion typically supports fidelity in relationships. Overall these results are fairly predictable when looking just at religion regardless of identity.

Regression Analysis

Associations between being religious and relational and sexual beliefs and behaviors could vary by race and gender/sexual identity; the influence of religion could depend somewhat upon cultural and other contextual factors that contribute to one’s identity. The variables above that correlated with being religious are often interrelated. Thus, separate models were used for the three largest race categories (white, black, and Hispanic) that controlled for the overlapping associations among the background characteristics and the relational and sexual beliefs and behavior variables on religiousness (Table 1). The Ordinary Least Squares (OLS) models indicated that some of the associations that were significant when analyzed in isolation were no longer significantly associated with religiosity (especially sexual behavior) once other variables were accounted for in the same model.

Table 1 Standardized coefficients of variables regressed on religiousness, by racial identity
Full size table

To statistically analyze whether associations were indeed different across the racial groups, unstandardized coefficients were compared using an online statistical calculator that compared the difference between two slopes (Soper n. d.). The following six variables were differentially related to being religious based on race: having divorced parents, willingness to divorce if no longer in love, would end a relationship if partner cheated, willingness to cohabit non-martially, and having looked for a partner online. Black individuals tended to stand out with the most distinct associations between religiosity and these variables.

Similarly, separate models were used for each of the six gender/sexual identities (Table 2). Once again, sexual behavior variables were largely unrelated to religiousness, and there was some evidence that associations with religiosity varied somewhat by sexual identity. For example, when testing for differences among unstandardized coefficients, three of the variables differentially associated with being religious: the belief in one true love, having looked for a partner online, and believing it is better to be heterosexual than gay. Lesbian women, compared to heterosexual women, differed on two variables (being religious associated with believing in a one true love for lesbian women; being religious did not associate with more positive perceptions about being heterosexual for lesbian women). Conversely, being religious was consistently, and highly, associated with negative abortion attitudes, regardless of sexual identity. Overall, the findings provide support for the premise that being religious can have some unique connotations or functions for beliefs and behaviors depending on one’s race and sexual identity.

Table 2 Standardized coefficients of variables regressed on religiousness, by gender/sexual identity
Full size table

Given the lack of significant associations with the sexual behavior variables, we investigated the intercorrelations among those variables to see if multicollinearity (very high correlations among them) could be impacting the results. The three sexual behavior variables indeed shared correlations as high as .58. The alpha reliability score for the three variables was marginal (.75). Subsequently, a composite variable was created to represent casual sex. Replacing the three separate variables with this single variable in the regressions did not change the results substantially. Heterosexual females were the only group for which the casual sex variable associated with religion, which parallels the finding that “has had sex without love” was the only sexual behavior variable associated with religion, and only for heterosexual females. For white participants only, the casual sex variable significantly associated with religion, which was different from the analyses with the separate items; however, the significant coefficient was only − .037, which is not statistically different from the − .023 coefficient for “has had sex without love” variable, both of which are small coefficients. It appears that shared variance could be playing a minor but inconsequential role in the analyses.

Discussion

Religion is considered one of the most influential social institutions impacting the daily lives of individuals. Scholars have argued that secularization among emerging adults is rapidly occurring in our society (Downey 2017). However, findings from this study revealed that religion is still relevant to some beliefs and values that emerging adults may transfer into life decisions. It is likely a means of social control that, once internalized, shapes how young people act in their romantic and sexual relationships (DeLamater 1981; Vasilenko et al. 2013). The broad array of relationship variables associated with religiosity suggests that this process of social influence or control likewise applies to various relationship beliefs and expectations and not just specific sexual attitudes and behavior. These beliefs can shape how individuals prepare for long-term adult relationships or how they interpret romantic interaction (Hall 2006; Willoughby et al. 2015). However, such religious influences potentially differ based on young adults’ race and sexual identity.

Correlational findings of this study were consistent with previous studies that examine associations between religiosity and race, sexual identity and having divorced parents. For example, Black respondents had higher levels of religiosity when compared to White and Hispanic respondents (see also Pew Research Center 2009). As found in the broader population, (Sherkat 2016), higher levels of religiosity were also found among heterosexual respondents. As noted, religious organizations often condemn being gay and promote a traditional nuclear family model (Walker and Longmire-Avital 2013). The challenges of integrating religious and LGB identities can result in becoming less religious (Sumerau et al. 2016).

The correlational analysis also illuminated the associations between self-identified religiosity and relationship beliefs and intentions. For example, being religious was significantly negatively correlated with a willingness to divorce if one were to fall out of love. Those respondents who were high on religiosity were less willing to divorce if they were to fall out of love. Being religious was also associated with marriage-solidifying beliefs such as having a one true love that never comes again and being less willingness to end a relationship with a cheating partner. Religion generally encourages both partners to be committed to their marriage for a lifetime, and religious culture emphasizes the importance that love is more than emotional feelings—it is a daily choice and falling out of love does not give one a license to divorce (e.g., Perry 2016).

Religious individuals had higher levels of disapproval of being gay, being involved in a “friends with benefits” relationship, cohabitation, hooking up, cheating, abortion, and using the Internet to find a partner. Once again, a pattern is reflected in these findings suggesting that religion impacts one’s personal views and behaviors (Barkan 2006; Baker et al. 2015; Burdette et al. 2007; Iles et al. 2016; Njus and Bane 2009; Perry 2016; Soloski et al. 2013). However, many of these associations became statistically insignificant once all variables were entered into a single regression model, suggesting that religion alone may not be a compelling source of influence on sexual decisions and behaviors.

Having more detail about the nature of respondents’ religious practices and perceived impact on their daily lives could help capture variation in how being religious might impact sexual behaviors (Baier and Wright 2001). For example, some have argued that adolescent sexual behavior is more likely influenced by religion when supported by networks of friends, families, and mentors with the same beliefs, which is rare (Regnerus 2007). For the current sample, regardless of race or sexual identity (for the most part), being religious may not in itself be sufficient to counter a broader culture with more permissive sexual mores. One might suspect that religious LGB individuals in particular, who are at risk of lacking support from broader religious and LGB communities (Sumerau 2014), would find it especially challenging to live consistently with conservative sexual values, though they may also have fewer opportunities to act on sexual desires compared to heterosexual individuals. In depth information about religious attendance, the nature or support from a religious community, and the sexual values of particular LGB congregations could help unpack the various influences that correspond with or counter sexual activities.

The findings from the regression analyses provided some evidence that religion could be differentially related to relationship behaviors and intentions by race and sexual identity. This was especially true for race and is consistent with the reviewed literature on substance use and on sexuality, showing differential associations of religion across diverse racial identities. The current study incorporated a broader array of relationship beliefs and intentions than past research, and indicated that most of the differentiation with religion across races manifested in that realm. That Black individuals appeared to have more distinct associations with religion and relationship beliefs could speak to previously discussed elements traditionally more common to Black denominations, such as less ostracization for a non-marital pregnancy and less tolerance of abortion (Brewster et al. 1998) or a heightened sense of communal worship and forgiveness (Lincoln and Mamiya 1990). Not knowing the actual denominations of the sample makes it impossible to know the relevance of this interpretation, however.

It is also possible that racial minority status contributes to navigating mainstream religious influences and culture in distinct ways, as has been shown to happen among sexual minorities (Thumma and Gray 2004). The functions of religion and nature of or need for certain religious rituals or beliefs could differ because of unique life experiences due to being a racial and/or sexual minority, resulting in somewhat differing impacts of being religious. Indeed, results of the current study also suggest that identifying as a sexual minority can likewise be relevant to religious associations with relationship variables. In this case, believing in one true love, openness to nonmarital cohabitation, looking for a partner on the internet, and believing it is better to be homosexual had some distinct correlations with being religious based on sexual identity. This pattern of findings lends some quantitative support for the premise that sexual identity is relevant for the impact of religion on relationship beliefs, intentions, and behavior.

Each regression model incorporated both types of identities, so some sense of the intersectionality of race and sexual orientation can be observed. The one variable that was differentially related to race and sexual identity was looking for a partner on the Internet. While religious Black respondents were less likely to have done so, religious gay males were more likely to have done so. Being Black was also especially associated with being religious among gay males. Together this may suggest that religious non-Black males were accounting for a propensity to look for a partner online. This serves as an example of how distinct identity combinations could be relevant for religious influence. Scholarship that focuses exclusively on the impact of religion without sensitivity to identity variation may be at risk for overgeneralizations and culturally biased conclusions.

Limitations

There are several limitations of this research. First, the study is cross-sectional; respondents were asked to report on their level of religiosity and views of love, relationships and sexual values at the time they took the survey so we were unable to track fluctuations in their level of religiosity/views over time that could account for the direction of influence between religion and the other variables. The sample was also exclusively college students ages 18 to 22, narrowing the potential generalizability of the findings. The measure or religiosity was very basic, limiting the ability to explore nuances in the nature of religiosity that could account for the identified associations.