Religiosity, sexual behaviors, and sexual attitudes during emerging adulthood.

Recent work suggests that individuals engage in their most extensive identity exploration during emerging adulthood--the period from approximately 18 to 25 years of age--rather than early adolescence (Arnett, 2000). For individuals attending residential colleges, this period of exploration may be intensified by the experiences of living away from home for the first time, and living in an environment surrounded by many social, political, religious, and interest-related activities. Two areas of development in particular--religiosity and sexuality--may be especially impressionable during this period of maturation (Arnett, 1992; CDC, 2000; Hoge, Johnson, & Luidens, 1993; Yankelovich, 1974).

As individuals transition from adolescence to adulthood, they become more committed to religion and their religious beliefs become more intrinsic; in contrast, their attendance at religious services actually decreases (De Haan & Schulenberg, 1997; Johnston, Bachman, & O'Malley, 1995; Kotesky, Walker, & Johnson, 1990; Stolzenberg, Blair-Loy, & Waite, 1995). Whereas religious practices may have been prescribed by family while living at home, once away at college individuals may have more opportunities to examine different religions and beliefs (Arnett, 2000).

In terms of sexuality, individuals are more likely to be sexually active in emerging adulthood compared to high school, but they are also less likely than younger individuals to use condoms (Arnett, 1996; CDC, 2000; Miller & Moore, 1990). In addition, college students feel more accepting of casual sex and feel less guilt about sex than do younger individuals (Chara & Kuennen, 1994; Miller & Moore, 1990). Love and dating are qualitatively different during emerging adulthood than they are during adolescence, with more focus on sexuality as it relates to developing one's own identity (Arnett, 2000).

Religiosity and sexuality are also closely linked to each other, in that religion potentially influences a range of decisions about sex-related issues such as abstinence, birth control, and abortion. The current study examines associations between religiosity and sexual behaviors and attitudes among emerging adults, using multidimensional assessments of both religiosity and sexuality.

Reference group theory has been used to explain associations between religiosity and sexual behaviors and attitudes (Bock, Beeghley, & Mixon, 1983; Mirande, 1968; Studer & Thornton, 1987; Zaleski & Schiaffino, 2000). This theory suggests that individuals will follow the teachings of their religion when determining their own sexual behaviors and attitudes. Differences in religiosity may help to determine differences in sexual behaviors and attitudes.

A number of studies have examined associations between religiosity and sexuality (e.g., Crockett, Bingham, Chopak, & Vicary, 1996; Sheeran, Abrams, Abraham, & Spears, 1993; Studer & Thornton, 1987). However, most studies have been limited in their measurement of religiosity, sexuality, or both. Many studies use one-item measures of religiosity, such as frequency of religious service attendance, or one-item self assessments of the importance of religion. As Rostosky, Wilcox, Wright, and Randall (in press) have concluded, most studies fail to capture the multidimensionality of religiosity, or to establish the ways in which specific aspects of religiosity may be associated with sexuality in different ways. For instance, frequency of attendance at religious services is routinely used to assess religiosity, but this measure fails to capture the particular religion's views toward sexual behaviors. The current study focuses on five specific aspects of religiosity: identity (religious group affiliation), behavior (frequency of attendance at religious services), attitudes (importance of religion in daily life), perception (religion's negative sanctions against sexual behaviors), and practice (extent to which individuals adhere to their religion's sanctions against sexual behavior). These five constructs assess several dimensions of religiosity, in that they include both behavior and attitudes, as well as both religion's role in daily life and religion's role in sexual behaviors. Two of these measures--group affiliation and frequency of attendance at services--are those most commonly used in studies of religiosity and sexual behaviors and attitudes. We were interested not only in our participants' religious behaviors and attitudes, but also in their perceptions of their religion's negative sanctions about sex. Two Catholic individuals may have very different perceptions of the church's position on specific sexual behaviors such as birth control, and thus it is important to assess each person's perceptions of these rules.

In the current study, we included measures of both sexual behaviors and sexual attitudes. Most studies of religiosity and sexuality have focused on sexual behaviors rather than sexual attitudes (e.g., Jessor & Jessor, 1975; Samuels, 1997; Studer & Thornton, 1987). These studies provide important information about how religiosity may be protective for individuals. However, fewer studies have examined the ways in which religiosity relates to sexual attitudes (but see for exceptions Miller & Olson, 1988; Sheeran et al., 1993; Thornton & Camburn, 1989). Most studies of religiosity that have included sexual attitudes have focused on general conservatism or permissiveness toward sex, but have not examined other attitudes that may place emerging adults at risk for negative outcomes, such as attitudes about HIV or condoms. These associations also may be particularly important for understanding more religiously conservative individuals, who may not currently be engaging in sexual activities. During emerging adulthood, attitudes about sexuality may develop and solidify, and thus religiosity has important implications not only for sexual behaviors, but also for sexual attitudes. Individuals who are not currently sexually active may develop attitudes during emerging adulthood about sex, HIV, and condoms that have the potential to influence their future sexual and contraceptive behaviors, whether marital or non-marital (Morrill, Ickovics, Golubchikov, Beren, & Rodin, 1996; Toon & Semin, 1999).

Studies of religious affiliation and sexual behavior have yielded mixed results, with some studies failing to find differences in sexual behaviors based on affiliation with specific religious groups (Studer & Thornton, 1987; Thornton & Camburn, 1989) and others finding that certain Christian groups, including Catholics and conservative or fundamentalist Protestants, were less likely to be sexually active than members of other religious groups (Bearman & Bruckner, 1999; Beck, Cole, & Hammond, 1991; Brewster, Cooksey, Guilkey, & Rindfuss, 1998). In contrast, research on sexual attitudes has more consistently revealed that Catholics and Protestants were more conservative than other groups (Sheeran et al., 1993; Thornton & Camburn, 1989).

In line with reference group theory, Thornton and Camburn (1989) argued that because most religions disapprove of premarital sex, religious behaviors may be a better predictor of sexual behaviors and attitudes than religious group membership. That is, reference group views about sexuality may not differ much across religions, but time spent with members of the reference group does vary. Research generally has found that individuals who attend religious services more frequently are less likely to be sexually active, have fewer sexual partners and less frequent sexual intercourse (Crockett et al., 1996; Jessor & Jessor, 1975; Miller & Olson, 1988; Sheeran et al., 1993; Studer & Thornton, 1987; Thornton & Camburn, 1989), and if sexually active, are less consistent contraception users (Studer & Thornton, 1987). Likewise, studies of sexual attitudes have revealed that individuals who attended religious services more frequently had more conservative attitudes about sex (Miller & Olson, 1988; Sheeran et al., 1993; Thornton & Camburn, 1989).

In addition to research on identity and behavior, another relatively frequently examined area is religious attitudes or the importance of religion. Religious attitudes may be particularly important for emerging adults because these individuals are in the process of exploring new worldviews (Arnett, 2000). Reference group theory suggests that identifying with a particular religion will lead individuals to avoid sexual behaviors because of the teachings of their religion (Zaleski & Schiaffino, 2000). Importance placed on religion, then, may be in part assessing the extent to which individuals use their religion as a reference group. Studies using both one-item or composite measures of the importance of religion in daily life have found that more religious individuals engage in fewer sexual activities and have more conservative attitudes about sex (Jessor & Jessor, 1975; Mahoney, 1980; Samuels, 1997; Sheeran et al., 1993; Whitbeck, Yoder, Hoyt, & Conger, 1999; Zaleski & Schiaffino, 2000).

Less research has addressed perceptions of religion's negative sanctions toward sexual behavior or the adherence to these sanctions, lf emerging adults use their religious institutions as reference groups, then perceptions of their religions' negative sanctions should influence their sexual behavior and attitudes. Similarly, the extent to which individuals adhere to these sanctions should also influence their sexual behaviors. Sheeran et al. (1993) asked participants the extent to which religious beliefs influence their decisions about sex. They found that those who reported a stronger reliance on religious beliefs when making sexual decisions tended to engage in sexual intercourse less frequently. Holder et al. (2000) examined nine aspects of religiosity, including attendance and importance, and their association with sexual behavior. They found that feeling spiritually interconnected with friends and believing that religion is important were associated with lower likelihood of voluntary sexual activity. Although they used a multidimensional measurement of religiosity, their measure of sexuality assessed only one aspect--whether participants were sexually active--and thus did not address other sexual behaviors or attitudes.

In the current study, we examined associations between religiosity and sexual attitudes and behaviors. We used five measures of religiosity (identity, behavior, attitudes, perceptions, and practice) to identify how different measures of religiosity may have unique associations with sexual behaviors and attitudes. Based on both reference group theory and prior empirical findings, we predicted the following:

1. More religious individuals would be more likely to be sexually abstinent, and if sexually active to have a later onset of intercourse, fewer partners, and less condom use than less religious individuals. We also expected that Catholics and Protestants would be less sexually experienced than nonbelievers (those who identified as agnostic or atheist or who reported that they had no religion).

2. More religious individuals would have more conservative sexual attitudes than less religious individuals. In addition, Catholics and Protestants would be more conservative than nonbelievers.

3. More religious individuals would perceive less vulnerability to HIV than less religious individuals. In addition, Catholics and Protestants would perceive less vulnerability than would nonbelievers.

4. More religious individuals would have less positive attitudes toward condoms and lower self-efficacy about condoms than would less religious individuals. In addition, Catholics and Protestants would have less positive attitudes and lower self-efficacy about condoms than would nonbelievers.

METHOD

Participants

We recruited 220 participants at a large public Eastern university. Several participants were excluded from the final sample because of established exclusion criteria (e.g., participants could not be graduate students and had to be age 25 or younger). As a result, the total sample consisted of 205 students.

Students were contacted through flyers distributed in general education and required summer courses, as well as through flyer distribution in high traffic areas on campus. The flyer described a study of college students' personal relationships. They read and signed consent forms before completing the questionnaires. Students received $10 for participating.

Sixty-one percent of the sample was female. Participants ranged in age from 18 to 25 (M = 20.5, SD = 1.7). The sample was predominantly European American (78%). The remaining individuals identified as African American (9%), Asian American (9%), Latino American (2%), and mixed ethnicity (3%). The majority of the sample identified as heterosexual (95%), with a small minority of students identifying as gay, lesbian, or homosexual (2%), bisexual (2%), and other (1%).

The participants reported a diverse range of socioeconomic backgrounds. Two percent of participants' mothers and 3% of fathers had not graduated from high school, 36% of mothers and 30% of fathers held a high school diploma as their highest degree, 43% of mothers and 33% of fathers had earned a 2-year or a 4-year college degree, and 19% of mothers and 34% of fathers had earned a graduate degree. Students represented all 10 colleges at the university, and all possible years of study (31% first year; 7% second year; 18% third year; 19% fourth year; and 24% fifth year or more).

The majority of participants identified as Catholic (41%) or Protestant (33%). There was great variability among the Protestants, with Baptists (5%), Presbyterians (5%), and Methodists (4%) the most frequently reported. The sample was 6% Jewish, 2% Muslim, 2% Hindu, 1% Buddhist, and 1% other. For ease of description, the term nonbeliever is used to denote the 14% of individuals who identified as agnostic or atheist or who reported that they had no religion. Given the small number of participants in certain religious groups, only Catholics, Protestants, and nonbelievers were considered in the religious affiliation comparisons.

It is clear that the sample was varied in terms of a number of factors, including religion, socioeconomic status, and sexual identity. The experiences of religiosity and sexuality may differ depending on these personal characteristics. For instance, the meaning of religiosity for sexual behavior may differ for Catholics compared to Buddhists, or for heterosexuals compared to bisexuals. Our sample size does not allow us to make comparisons across all religions or across specific sexual identities. We chose to include all eligible participants in analyses that did not involve religious group comparisons, so as not to exclude subgroup populations from the analyses and to maximize power to detect significant results.

Measures

Religious service attendance. We assessed frequency of attendance at religious services by asking participants to report the number of times they attended religious services during the past 12 months.

Religion in daily life. We used a 7-item measure of religiosity (Rohrbaugh & Jessor, 1975) to assess the importance of religion in daily life. Each item contained 4 or 5 response choices. For instance, the question "Which of the following best describes your practice of prayer or religious meditation?" had the following response options: "Prayer is a regular part of my daily life," "I usually pray in times of stress or need but rarely any other time," "I pray only during formal ceremonies," and "I never pray." Responses were summed across the 7 items with higher scores indicating that religion had more importance in the respondent's daily life. This measure demonstrated good reliability in previous research with college students ([alpha] = .90; Rohrbaugh & Jessor, 1975) and in the current sample ([alpha] = .90).

Religious adherence. This instrument measures individuals' adherence to their religion's principles in four different sex-related domains: using birth control, premarital sex, extramarital sex, and abortion (Wyatt et al., 2000). Respondents answered each question on a 5-point scale that ranged from not at all to very closely. In a previous study of females the measure demonstrated adequate reliability ([alpha] = .71; Wyatt et al., 2000). Reliability for the current sample was similar ([alpha] = .68).

Religion's negative sanctioning of behaviors. Because different religious groups endorse different views about sexual behaviors, we created a 4-item measure for the current study to assess the degree to which participants perceive their religion to negatively sanction the four behaviors described above (using birth control, premarital sex, extramarital sex, and abortion). Participants were asked the following: "Please rate how much your religion thinks it is wrong to do each of the following things." They responded using a 5-point scale that ranged from not at all against my religion to very much against my religion. Reliability was satisfactory ([alpha] = .84).

Sexual behaviors. We asked participants if they had ever had penetrative sex, defined as "sex in which the penis penetrates the vagina or anus." Those participants who answered "yes" were considered sexually active, whereas those who responded "no" were considered sexually abstinent. Sexually active respondents were also asked to report their age at first intercourse and number of lifetime sexual partners. In addition, lifetime frequency of condom use was assessed by the question "How frequently have you had penetrative sex without using a condom?" Response options were "never," "once," "a couple of times," "some of the time," and "most of the time."

Sexual attitudes. We used a shortened 21-item version of the Sexual Attitudes Scale (Hudson, Murphy, & Nurius, 1983) to measure respondents' general conservative attitudes toward sex (e.g., "I think that increased sexual freedom undermines the American family," "Movies today are too sexually explicit"). Respondents were asked to rate their agreement on a 5-point scale ranging from strongly disagree to strongly agree. Reliability in the current sample ([alpha] = .88) was similar to that reported in previous research ([alpha] = .94; Hudson et al., 1983).

Perceived vulnerability. The Fear of AIDS subscale of the Multidimensional AIDS Anxiety Questionnaire (Snell & Finney, 1996) is a 6-item measure that we used to assess the extent to which participants experienced fear or concern pertaining to AIDS (e.g., "I feel scared when I think about catching AIDS from a sexual partner"). We asked respondents to rate their agreement on a 5-point scale. Reliability for the current sample on this subscale ([alpha] = .91) fell within the range reported in previous research ([alpha] = .85-.94; Snell & Finney, 1996).

Attitudes about condoms. We used the Outcome Expectancies of Condom Use Scale (Jemmott & Jemmott, 1992) to measure respondents' positive attitudes about condoms. The first subscale, prevention expectancies, is a 3-item measure that assesses the extent to which participants think that condoms can provide physical protection from pregnancy and sexually transmitted diseases (e.g., "Condoms can prevent pregnancy"). The second subscale, hedonistic expectancies, is a 5-item measure of participants' positive attitudes toward condoms (e.g., "Sex is natural when condoms are used"). Responses are on a 5-point scale. Reliability in the current sample ([alpha] = .81 for prevention, .74 for hedonistic) was comparable to that reported in previous research ([alpha] = .50-.73; Jemmott & Jemmott, 1992).

Condom use self-efficacy. The Condom Use Self-Efficacy Scale was used to measure participants' self-efficacy for condom use (Basen-Engquist et al., 1996). It consists of three subscales (3 items each) that assess self-efficacy for communicating about condom use (e.g., "... how sure are you that you could tell your partner that you want to start using condoms?"), self-efficacy for buying and using condoms (e.g., "If you wanted to get a condom, how sure are you that you could go to the store and buy one?"), and barriers to condom use (e.g., "I would feel uncomfortable carrying condoms with me"). Higher scores on the first two subscales are indicative of high self-efficacy regarding condom use, whereas higher scores on the third subscale are indicative of low self efficacy. Similar to reliability demonstrated in previous research ([alpha] = .66-.73; Basen-Engquist et al., 1996), reliability in the current sample was adequate ([alpha] = .64-.74).

RESULTS

Preliminary and Descriptive Analyses

We performed correlations to examine associations among the 4 measures of religiosity (see Table 1). Correlations ranged from .07 to .43. Given this magnitude of association, we determined that these measures were unique enough to be examined as separate measures.

There were no significant gender differences in religiosity on any of the 4 measures, ts(182-202) < 1.3, ps > .05. We found associations between individuals' ages and their religiosity. Older individuals reported attending religious services less frequently, r(200) = -.29, p < .001, that religion was less influential in their daily lives, r(198) = -.19, p < .01, and that their religion had fewer negative sanctions against sexual behaviors, r(184) = -.18, p < .05.

Religiosity and Sexual Behaviors

Our first hypothesis was that more religious individuals would be more likely to be sexually active than less religious individuals, and that Catholics and Protestants would be less sexually experienced than nonbelievers. We used t tests to compare sexually active to sexually abstinent individuals on the four measures of religiosity (see Table 2). Sexually abstinent youth reported on average attending religious services almost weekly, compared to less than once a month for sexually active youth. Sexually abstinent individuals reported that religion had more influence in their daily lives and that they adhered to their religion's teachings more than did sexually active individuals. These significant differences represented small to moderate effect sizes (Cohen, 1988). The groups did not differ on their perceptions of their religion's negative sanctions against sexual behavior. There were no significant differences in proportions of sexually active individuals who were Catholic, Protestant, or nonbelievers, [chi square] = 1.11, p > .05.

We used correlations to examine associations among religiosity and other measures of sexual behavior for sexually active individuals (n = 154). We found no associations for age at first sexual intercourse or regularity of condom use. Youth who attended services more frequently reported having fewer lifetime sexual partners compared to less religious youth, r(151) = -.21, p < .01. In addition, we used ANOVAs to examine whether sexual behaviors differed by religious affiliation; when the ANOVA was significant, we computed effect size ([eta], see Rosenthal, Rosnow, & Rubin, 1999). No group differences were found for age at first intercourse or regularity of condom use, Fs(2, 130) < 1.0, ps > .05. We found a significant difference for number of lifetime partners, F(2, 129) = 3.68, p < .05, [eta] = .23. Tukey post-hoc tests revealed that sexually active Catholics (M = 3.8, SD = 4.2) reported fewer lifetime sexual partners than sexually active Protestants (M = 6.6, SD = 6.7), p < .05. Nonbelievers did not significantly differ from Catholics or Protestants. Therefore, our hypothesis was supported for sexual abstinence and number of lifetime partners, but not for age of onset or condom use.

Religiosity and Sexual Attitudes

To examine associations between religiosity and sexual attitudes, we calculated partial correlations followed up by regressions. Our descriptions focus on the regressions because we were interested in understanding each measure of religiosity in the context of the other measures of religiosity. In all these analyses, gender and sexual activity status (abstinent vs. active) were entered as controls because both religiosity and sexual attitudes have been shown to differ by gender and sexual behavior. If these variables were not controlled, associations between the two constructs could be spurious. First we computed partial correlations, controlling for gender and sexual activity status (see Table 3). Next, we ran a series of regression analyses to examine measures of religiosity in the context of the other religiosity measures. In each regression, gender and sexual activity status were entered in the first step and the four religiosity measures were entered in the second step.

Our second hypothesis was that more religious individuals would have more conservative sexual attitudes than would less religious individuals (see Table 4). The final model accounted for 30% of the variance. In the first step, sexual behavior significantly related to sexual attitudes. In the second step, two of the religiosity measures significantly related to sexual attitudes. Individuals for whom religion was more a part of their daily lives and those who adhered to their religion in regard to sexual behaviors tended to have more conservative sexual attitudes.

We used ANOVAs to compare Catholics, Protestants, and nonbelievers on sexual conservatism and found significant differences, F(2, 173) = 4.78, p < .01, [eta] = .23. Tukey post-hoc tests revealed that Protestants (M = 53.4, SD = 11.8) were more conservative than were nonbelievers (M = 45.5, SD = 10.2), p < .05. Catholics did not significantly differ from Protestants or nonbelievers. Thus, our second hypothesis was mostly supported.

Religiosity and Perceived Vulnerability to HIV

Our third hypothesis was that more religious individuals would perceive less vulnerability to HIV than would less religious individuals. A regression explained 9% of the variance in fear of HIV, with only the second step significant (see Table 4). Specifically, individuals who attended services more frequently had less fear about HIV. In contrast, individuals who saw religion as playing a more important role in their daily lives tended to fear HIV more. An ANOVA revealed no significant differences in perceived vulnerability to HIV by religious group affiliation, F(2, 173) = 0.87, p > .05. Thus, our third hypothesis was only partially supported.

Religiosity and Condom-Related Beliefs

Our fourth hypothesis was that more religious individuals would have less positive attitudes toward condoms and lower self-efficacy about condoms than would less religious individuals. We performed two regressions to examine associations between religiosity and attitudes about condoms (see Table 4). The regressions explained 8% of the variance in hedonistic outcome expectancies and 10% of preventive expectancies. Only gender was significantly associated with hedonistic attitudes about condom use. Gender also predicted prevention attitudes about condom use, as did the extent to which an individual adhered to his or her religion's rules about sexual behavior. Individuals who followed their religions more closely were less likely to believe that condoms could prevent negative outcomes such as pregnancy or sexually transmitted diseases.

We performed three regressions to examine how religiosity was associated with condom use self-efficacy, explaining 4% of the variance in communication, 22% of the variance in using and buying, and 17% of the variance in perceived barriers. In all three models, sexual behavior was a significant predictor (though for communication self-efficacy the overall regression model was not significant). Sexually active individuals tended to have higher condom use self-efficacy than did others. For using and buying condoms, gender was also significant in the first step. The second step did not significantly add to the explanation of variance in self-efficacy for using and buying condoms. However, in the final model, religion's negative sanctions toward sexual behavior was also significant. Specifically, individuals who viewed their religion as having more sanctions against sexual behaviors had less self-efficacy for using and buying condoms. For barriers to condom use, the extent to which an individual adhered to his or her religion in regard to sexual behavior was a significant predictor. Individuals who adhered to their religion more tended to perceive more barriers to condom use.

Finally, we calculated five ANOVAs to examine differences in condom beliefs by religious affiliation. The only significant finding was for self-efficacy for communicating about condoms, F(2, 173) = 3.05, p < .05, [eta] = .18. Tukey post-hoc tests revealed only marginal differences. Nonbelievers (M = 14.3, SD = 1.2) reported marginally higher confidence in their ability to talk to a partner about condoms than did Protestants (M = 13.3, SD = 2.1) or Catholics (M = 13.3, SD = 2.0), ps = .06. Thus, our fourth hypothesis was partially supported.

DISCUSSION

Our four measures of religiosity were generally associated, but at relatively modest levels. This finding suggests that these constructs are overlapping but include distinct features. That is, religious behavior and practice are associated with religious attitudes and perceptions, but attitudes do not fully determine behaviors (in this case, the behavior was attendance at religious services).

Older youth reported fewer religious behaviors, fewer perceived negative sanctions from their religions, and less importance of religiosity in their daily lives than did younger individuals. These age differences in behavior are supported by prior research on decreases in religious service attendance during the transition to adulthood (Johnston et al., 1995). In contrast to our findings, research also suggests that attitudinally, individuals become more religious during this transition (De Haan & Schulenberg, 1997; Kotesky et al., 1990; Stolzenberg et al., 1995). It may be that the oldest individuals in the current sample are still exploring their religions and therefore have yet to become more committed, as emerging adulthood is a period of exploration in the areas of religion and religiosity (Arnett, 2000). There are many environmental transitions that may contribute to these changes, including independent decisions about attending religious services, increased exposure to other religions, and the desire to explore behaviors inconsistent with the teaching of one's religion, including sexual behaviors.

Some differences emerged based on religious identity. Specifically, Catholics had fewer lifetime sexual partners than did Protestants. Protestants reported more conservative sexual attitudes than did nonbelievers. Nonbelievers reported marginally higher confidence in their ability to talk to a partner about condoms than did Protestants or Catholics. These differences in sexual behaviors and attitudes by religious group identification, however, were inconsistent. These inconsistent findings are similar to those examining sexual behaviors (e.g., Bearman & Bruckner, 1999; Thornton & Camburn, 1989). Our finding of group differences in conservative attitudes also is supported by the literature (Sheeran et al., 1993; Thornton & Camburn, 1989). The other attitudinal measures that we assessed, such as attitudes toward HIV and toward condoms, have not generally been measured in previous studies, so we cannot compare these findings to past work. It may be that because one's religious affiliation--although an indicator of religious identity--does not assess commitment, behavior, or attitudes, it is not as important as these factors when examining associations with sexuality. Because most religions disapprove of premarital sex, reference group theory suggests that variations based on religious group membership may be less significant than variations based on religious behaviors or attitudes (Thornton & Camburn, 1989).

In support of this distinction and as frequently reported in the literature (e.g., Jessor & Jessor, 1975; Samuels, 1997) religious behavior differed between sexually active and abstinent youth and was associated with number of lifetime sexual partners. These associations are important because emerging adults' decisions about religiosity and sexual behaviors were made while living away from home and, presumably, under less immediate influence from their parents. Although we did not replicate findings of an association between religious attendance and conservative sexual attitudes in general (Sheeran et al., 1993; Thornton & Camburn, 1989), frequency of attendance at religious services was associated with perceived vulnerability to HIV and perceived barriers to condom use. Individuals who attended services more frequently were less fearful of acquiring HIV and perceived more barriers to using condoms, even after controlling for their sexual behavior. However, when other aspects of religiosity were taken into account, religious behavior had little explanatory power for sexual and condom attitudes. Thus, religious behavior may be more important in explaining sexual behavior than in explaining sexual attitudes. Most of the prior studies that have examined associations between religious service attendance and sexual attitudes have not considered other measures of religiosity simultaneously, which may explain the difference in findings.

Sexually abstinent youth reported that religion was more important in their daily lives than did sexually active youth. In addition, religious attitudes were associated with conservative sexual attitudes and perceived vulnerability to HIV. Youth who perceived religion as more important in their daily lives tended to be more conservative sexually and to fear HIV more. Previous researchers have consistently found associations between religious attitudes and sexual behaviors and conservative attitudes (e.g., Jessor & Jessor, 1975; Samuels, 1997; Sheeran et al., 1993). In the current study, religious attitudes were also a powerful predictor of sexual behaviors and attitudes, even after taking other measures of religiosity into account. Importance of religion in daily life may be assessing the extent to which emerging adults use their religion as a reference group (Zaleski & Schiaffino, 2000). This measure did not, however, relate to any of the measures of condom-related beliefs.

Sexually abstinent youth reported more adherence to their religions' negative sanctions than did sexually active youth. In addition, adherence to one's religions' negative sanctions was associated with conservative sexual attitudes, preventive expectancies of condom use, and perceived barriers to condom use. Thus, in line with previous work, these sex-specific practices of religion explained some of the variance in many of the sexual attitude measures, as well as in sexual behavior (Sheeran et al., 1993). This finding may be important to consider when examining associations between religion and outcomes other than sexuality. The explanatory power of religion may be stronger when examining practices specific to the outcome of interest, compared to examining more generalized religiosity (e.g., religious service attendance).

Sexual behavior was not associated with youth's perceptions of their religions' negative sanctions toward sexual behavior. It appears that individuals' practice of their religions' negative sanctions is more important in determining their behavior than is their perception of what those sanctions are. As with religious group membership, this finding supports reference group theory in that the generalized disapproval of premarital sex in most religions means that perceived negative sanctions may be less important than sexual behaviors or attitudes (Thornton & Camburn, 1989). Religion's negative sanctions against sexual behavior (including use of birth control) did explain youth's preventive expectations and self-efficacy to use and buy condoms. Those who viewed their religion as having more negative sanctions had less faith in condoms' preventive abilities and felt less confident in their abilities to purchase or use condoms. Individuals may feel uncomfortable about engaging in behaviors (premarital sexual activity and birth control use) that their religious institutions forbid. However, the role of negative sanctions was less important when other religiosity factors were taken into account.

Surprisingly, the association between religious attitudes and fear of HIV was in the opposite direction of the association for religious behaviors. In other words, youth who attended services more frequently feared HIV less, but youth who felt religion's presence in their daily lives more strongly feared HIV more. It is possible that some religions promote fear about HIV, or that certain personality types turn to religion more than others. However, given how little is known about attitudes toward HIV and religiosity, more research is needed before drawing firm conclusions. These findings do suggest, though, that religious behaviors and religious attitudes perform in different ways when it comes to HIV attitudes.

We know of no prior studies that have examined associations between religiosity and condom attitudes. Research does suggest that among those who are sexually active, more religious individuals are less consistent contraception users (Studer & Thornton, 1987). Although measures of religiosity explained some aspects of condom attitudes, none of these measures were associated with hedonistic expectancies of condom use. In other words, beliefs about whether condoms interfere with pleasure were independent from religious beliefs. Similarly, none of the religiosity measures were associated with feeling self-efficacious about communicating with one's partner about using condoms. These findings suggest that religious reference groups may be less important for some aspects of condom use--particularly those involving pleasure or interpersonal negotiation--than they are for beliefs about function (prevention) and barriers. It will be important to further examine how aspects of religious attitudes are uniquely associated with specific components of condom attitudes.

The percentage of variance explained in sexual attitudes ranged from 4% to 30%. Religiosity had the most explanatory power for conservative sexual attitudes, one of the measures most frequently used in studies of religiosity (Miller & Olson, 1988; Thornton & Camburn, 1989). This finding suggests that general sexual attitudes may be better explained by religiosity than more specific attitudes about HIV or communicating about condoms. Many aspects of religiosity are associated with a very general type of sexual attitude, but when more specific sexual attitudes are examined, different dimensions of religiosity function in different ways. It is possible that general conservatism is most influenced by reference groups, in this case religion. Religiosity never explained the majority of the variance in any sexual attitude. Attitudes about sex, HIV, and condoms are clearly multifaceted constructs, and thus likely to be influenced not only by religious reference groups but also by groups such as family and friends, as well as romantic partners and individual experiences.

It is important to note some of the limitations of the current study. Use of a college sample prevents generalization to youth of similar ages who are not enrolled in college. In addition, our sample did not include enough youth from Jewish or non-Judeo-Christian religions to include comparisons beyond the three groups considered here. It is also important to note that the grouping of Protestant includes a wide range of individuals of varying degrees of religious conservatism, but the current sample did not allow for comparisons within this group. As with religious group affiliation, our sample did not include enough sexual minority youth to make comparisons between heterosexuals and those of other sexual identities. Finally, because the current study was cross-sectional, we cannot determine causal mechanisms. Just as it is possible that religiosity causes sexual behaviors and attitudes, it is also possible that sexual behaviors and attitudes (and changes in them) cause degree of religiosity.

Despite these limitations, important findings emerged. We found that the five aspects of religiosity (identity, behavior, attitudes, perceptions, and practice) were associated with sexual behaviors and attitudes in unique ways. Religious behavior may be the strongest predictor of sexual behavior, whereas more attitudinal measures of religiosity may be better predictors for sexual attitudes. The findings suggest the importance of identifying the specific religious and sexual constructs of interest in a particular study before designing measurement. Findings also suggest that, despite evidence of changes in sexuality and religiosity during emerging adulthood, these domains continue to be associated with each other during this developmental period. In future studies, it will be important to examine these multiple dimensions of religiosity and sexuality in other cultural and religious groups and subgroups, as well as to examine how religiosity and sexuality predict each other over time.

Table 1. Associations Among Religiosity Measures

                                  Religious
                                  attendance   Religion in
Measure                           frequency    daily life

Religious attendance frequency        --         .42 **
Religion in daily life                            --
Religion's negative sanctions
Religious adherence

                                  Religion's
                                   negative    Religious
Measure                           sanctions    adherence

Religious attendance frequency      .40 **       .15 *
Religion in daily life              .43 **       .40 **
Religion's negative sanctions        --          .07
Religious adherence                               --

Note. Due to missing data, sample size ranged from n = 183 to 201.
* p < .05. ** p <.01.

Table 2. Differences in Religiosity Between Sexually Active and
Sexually Abstinent Youth

                                   Sexually      Sexually
                                    active      abstinent
Measures of religiosity             M (SD)        M (SD)

Religious attendance frequency
  (times per year)                 8.9 (13.1)   23.1 (22.5)
Religion in daily life            13.2 (6.2)    15.3 (6.6)
Religion's negative sanctions     15.1 (4.7)    15.6 (4.7)
Religious adherence               10.7 (4.0)    14.2 (4.1)

Measures of religiosity              t       [eta]

Religious attendance frequency
  (times per year)                4.23 ***    .29
Religion in daily life            2.01 *      .14
Religion's negative sanctions     0.61        .04
Religious adherence               5.05 ***    .35

Note. Due to missing data, sample size ranged from n = 47-51 for
sexually abstinent and n = 137-154 for sexually active individuals.
* p < .05. *** p < .001.

Table 3. Associations Between Religiosity Measures and Sexual and
Condom Attitudes (Gender and Sexual Activity Status Partialled Out)

Measure                            Religious    Religion
                                   attendance   in daily
                                   frequency      life

Conservative sexual attitudes        .20 **      .36 ***
Perceived vulnerability to HIV      -.14 *       .12
Hedonistic outcome expectancies     -.04         .05
  of condom use
Preventive outcome expectancies     -.08        -.14 *
  of condom use
Communication about condoms         -.09        -.04
Using & buying condoms              -.11        -.02
Barriers to condom use               .15 *       .02

Measure                            Religion's   Religious
                                    negative    adherence
                                   sanctions

Conservative sexual attitudes        .24 ***      .38 ***
Perceived vulnerability to HIV       .03         -.04
Hedonistic outcome expectancies      .06          .08
  of condom use
Preventive outcome expectancies     -.17 *       -.20 **
  of condom use
Communication about condoms         -.07         -.02
Using & buying condoms              -.17 *       -.08
Barriers to condom use               .12          .14

Note. Due to missing data, sample size ranged from n = 171 to 200.
* p<.05.**n<.01.***p<.001.

Table 4. Regression Model Predicting Measures of Sexual Attitudes
From Religiosity

Variable                             B     SE B    [beta]    [R.sup.2]
                                                              [DELTA]

                    Conservative sexual attitudes
                    (n = 181, [R.sup.2] = .30 ***)

Step 1                                                         .08 ***
  Sexual activity                  -7.99   2.05   -0.28 ***
  Gender                           -0.72   1.84   -0.03
Step 2                                                         .21 ***
  Sexual activity                  -3.01   2.08   -0.11
  Gender                           -1.21   1.64   -0.05
  Religious service attendance      0.06   0.06    0.08
  Religion in daily life            0.37   0.16    0.18 *
  Religion's negative sanctions     0.34   0.20    0.12
  Religious adherence               0.89   0.22    0.31 ***

                    Perceived vulnerability to HIV
                     (n = 182, [R.sup.2] = .09 *)

Step 1                                                         .02
  Sexual activity                   -.59   1.02    -.04
  Gender                            1.66    .91     .14
Step 2                                                         .07 *
  Sexual activity                  -1.99   1.14    -.15
  Gender                            1.43    .90     .12
  Religious service attendance      -.09    .03    -.26 **
  Religion in daily life             .25    .09     .24 **
  Religion's negative sanctions      .08    .11     .06
  Religious adherence               -.19    .12    -.14

             Hedonistic outcome expectancies of condom use
                     (n = 173, [R.sup.2] = .08 *)

Step 1                                                         .06 **
  Sexual activity                    .90    .67     .10
  Gender                            1.66    .57     .22 **
Step 2                                                         .02
  Sexual activity                    .75    .75     .08
  Gender                            1.56    .57     .21 **
  Religious service attendance      -.02    .02    -.11
  Religion in daily life             .04    .06     .07
  Religion's negative sanctions      .05    .07     .06
  Religious adherence                .03    .08     .03

             Preventive outcome expectancies of condom use
                     (n = 180, [R.sup.2] = .10 **)

Step 1                                                         .04 *
  Sexual activity                    .41    .48     .06
  Gender                           -1.08    .43    -.19 *
Step 2                                                         .06 *
  Sexual activity                   -.12    .53    -.02
  Gender                           -1.02    .42    -.18 *
  Religious service attendance      -.00    .01    -.02
  Religion in daily life            -.00    .04    -.01
  Religion's negative sanctions     -.08    .05    -.12
  Religious adherence               -.13    .06    -.20 *

        Condom use self-efficacy: Communication about condoms
                     (N = 182, [R.sup.2] = .04)

Step 1                                                         .03
  Sexual activity                    .72    .34     .16 *
  Gender                             .24    .30     .06
Step 2                                                         .01
  Sexual activity                    .57    .39     .12
  Gender                             .22    .31     .05
  Religious service attendance      -.01    .04    -.07
  Religion in daily life             .02    .03     .05
  Religion's negative sanctions     -.03    .04    -.07
  Religious adherence               -.02    .04    -.04

           Condom use self-efficacy: Using and buying condoms
                     (n = 181, [R.sup.2] = .22 ***)

Step 1                                                         .18 ***
  Sexual activity                   2.37    .42     .39 ***
  Gender                           -1.09    .37    -.20 **
Step 2                                                         .04
  Sexual activity                   1.98    .47     .32 ***
  Gender                           -1.12    .37    -.20 **
  Religious service attendance      -.01    .01    -.10
  Religion in daily life             .05    .04     .12
  Religion's negative sanctions     -.09    .05    -.16 *
  Religious adherence               -.07    .05    -.12

            Condom use self-efficacy: Barriers to condom use
                     (n = 181, [R.sup.2] = .17 ***)

Step 1                                                         .11 ***
  Sexual activity                  -1.72    .36    -.34 ***
  Gender                             .27    .32     .06
Step 2                                                         .05 *
  Sexual activity                  -1.17    .40    -.23 **
  Gender                             .30    .32     .07
  Religious service attendance       .02    .01     .16
  Religion in daily life            -.03    .03    -.09
  Religion's negative sanctions      .05    .04     .10
  Religious adherence                .09    .04     .17 *

* p < .05. ** p < .01. *** p < .001.

REFERENCES

Arnett, J. (1992). Reckless behavior in adolescence: A developmental perspective. Developmental Review, 12, 339-373.

Arnett, J. J. (1996). Sensation seeking, aggressiveness, and adolescent reckless behavior. Personality and Individual Differences, 20, 693-702.

Arnett, J. J. (2000). Emerging adulthood: A theory of development from the late teens through the twenties. American Psychologist, 55, 469-480.

Basen-Engquist, K., Masse, L., Coyle, K, Parcel, G. S., Banspach, S., Kirby, D., et al. (1996). Validity of scales measuring the psychosocial determinants of HIV/ STD-related risk behavior in adolescents. Unpublished manuscript.

Bearman, P. S., & Bruckner, H. (1999). Power in numbers: Peer effects on adolescent girls' sexual debut and pregnancy. Washington, DC: The National Campaign to Prevent Teen Pregnancy.

Beck, S. H., Cole, B. S., & Hammond, J. A. (1991). Religious heritage and premarital sex: Evidence from a national sample of young adults. Journal for the Scientific Study of Religion, 30, 173-180.

Bock, E. W., Beeghley, L., & Mixon, A. J. (1983). Religion, socioeconomic status, and sexual morality: An application of reference group theory. The Sociological Quarterly, 24, 545-559.

Brewster, K. L., Cooksey, E. C., Guilkey, D. K., & Rindfuss, R. R. (1998). The changing impact of religion on the sexual contraceptive behavior of adolescent women in the United States. Journal of Marriage and the Family, 60, 494-504.

Centers for Disease Control and Prevention. (2000). Health, United States. Hyattsville, MD: National Center for Health Statistics.

Chara, P. J., & Kuennen, L. M. (1994). Diverging gender attitudes regarding casual sex: A cross-sectional study. Psychological Reports, 74, 57-58.

Cohen, J. (1988). Statistical power analysis for the behavior sciences (2nd ed.). Hillsdale, NJ: Erlbaum.

Crockett, L. J., Bingham, C. R., Chopak, J. S., & Vicary, J. R. (1996). Timing of first sexual intercourse: The role of social control, social learning, and problem behavior. Journal of Youth and Adolescence, 25, 89-111.

De Haan, L. G., & Schulenberg, J. (1997). The covariation of religion and politics during the transition to young adulthood: Challenging global identity assumptions. Journal of Adolescence, 20, 537-552.

Hoge, D. R., Johnson, B., & Luidens, D. B. (1993). Determinants of church involvement of young adults who grew up in Presbyterian churches. Journal for the Scientific Study of Religion, 32, 242-255.

Holder, D. W., Durant, R. H., Harris, T. L., Daniel, J. H., Obeidallah, D., & Goodman, E. (2000). The association between adolescent spirituality and voluntary sexual activity. Journal of Adolescent Health, 26, 295-302.

Hudson, W. W., Murphy, G. J., & Nurius, P. S. (1983). A short-form scale to measure liberal vs. conservative orientations toward human sexual expression. The Journal of Sex Research, 19, 258-272.

Jemmott, L. S., & Jemmott, J. B. (1992). Increasing condom-use intentions among sexually active black adolescent women. Nursing Research, 41, 273-279.

Jessor, S. L., & Jessor, R. (1975). Transition from virginity to nonvirginity among youth: A social-psychological study over time. Developmental Psychology, 11, 473-484.

Johnston, L. D., Bachman, J. G., & O'Malley, P. M. (1995). Monitoring the future: Questionnaire responses from the nation's high school seniors, 1993. Ann Arbor, MI: Institute for Social Research.

Kotesky, R. L., Walker, J. S., & Johnson, A. W. (1990). Measurement of identity from adolescence to adulthood: Cultural, community, religious, and family factors. Journal of Psychology and Theology, 18, 54-65.

Mahoney, E. R. (1980). Religiosity and sexual behavior among heterosexual college students. Journal of Marriage and the Family, 49, 97-113.

Miller, B., & Moore, K. (1990). Adolescent sexual behavior, pregnancy, and parenting: Research through the late 1980s. Journal of Marriage and the Family, 52, 1025-1044.

Miller, B. C., & Olson, T. D. (1988). Sexual attitudes and behavior of high school students in relation to background and contextual factors. The Journal of Sex Research, 24, 194-200.

Mirande, A. (1968). Reference group theory and adolescent sexual behavior. Journal of Marriage and the Family, 30, 572-577.

Morrill, A. C., Ickovics, J. R., Golubchikov, V. V., Beren, S. E., & Rodin, J. (1996). Safer sex: Social and psychological predictors of behavioral maintenance and change among heterosexual women. Journal of Consulting and Clinical Psychology, 64, 819-828.

Rohrbaugh, J., & Jessor, R. (1975). Religiosity in youth: A personal control against deviant behavior. Journal of Personality, 43, 136-155.

Rosenthal, R., Rosnow, R. L., & Rubin, D. B. (1999). Contrasts and effect sizes in behavioral research: A correlational approach. New York: Cambridge University Press.

Rostosky, S. S., Wilcox, B. L., Wright, M. L. C., & Randall, B. A. (in press). The impact of religiosity on adolescent sexual behavior. Journal of Adolescent Research.

Samuels, H. P. (1997). The relationships among selected demographics and conventional and unconventional sexual behaviors among black and white heterosexual men. The Journal of Sex Research, 34, 85-92.

Sheeran, P., Abrams, D., Abraham, C., & Spears, R. (1993). Religiosity and adolescents' premarital sexual attitudes and behaviour: An empirical study of conceptual issues. European Journal of Social Psychology, 23, 39-52.

Snell, W. E., Jr., & Finney, P. (1996). The Multidimensional AIDS Anxiety Questionnaire. Unpublished manuscript.

Stolzenberg, R. M., Blair-Loy, M., Waite, L. J. (1995). Religious participation in early adulthood: Age and family life cycle effects on church membership. American Sociological Review, 60, 84-103.

Studer, M., & Thornton, A. (1987). Adolescent religiosity and contraceptive usage. Journal of Marriage and the Family, 49, 117-128.

Thornton, A., & Camburn, D. (1989). Religious participation and adolescent sexual behavior and attitudes. Journal of Marriage and the Family, 51, 641-653.

Toon, T. W., & Semin, G. R. (1999). Sexual efficacy of adolescents, permissiveness, and the transition toward nonvirginity: From intention to behavior. Journal of Applied Social Psychology, 29, 903-917.

Whitbeck, L. B., Yoder, K. A., Hoyt, D. R., & Conger, R. D. (1999). Early adolescent sexual activity: A developmental study. Journal of Marriage and the Family, 61, 934-946.

Wyatt, G. E., Carmona, J. V., Loeb, T. B., Guthrie, D., Chin, D., & Gordon, G. (2000). Factors affecting HIV contraceptive decision-making among women. Sex Roles, 42, 495-521.

Yankelovich, D. (1974). The new mortality: A profile of American youth in the 1970s. New York: McGraw-Hill.

Zaleski, E. H., & Schiaffino, K. M. (2000). Religiosity and sexual risk-taking behavior during the transition to college. Journal of Adolescence, 23, 223-227.

Manuscript accepted September 10, 2003

This work was supported by National Institute of Child Health and Human Development Grant #R-01 HD 41720 to Eva S. Lelkowitz. We are grateful to Shelley Hosterman, Eric Loken, Susan McHale, Lyndsey Sturm, Elizabeth Tempio, and Georgette Villareal for their help with study design, data collection, data scoring and entering, data cleaning, and statistical analyses.

Address correspondence to Eva S. Lefkowitz, 110 South Henderson Building, Pennsylvania State University, University Park, PA 16802: e-mail: [e-mail].

Eva S. Lefkowitz, Meghan M. Gillen, and Cindy L. Shearer 

Pennsylvania State University

Tanya L. Boone

California State University, Bakersfield