Religiosity, Sex Frequency, and Sexual Satisfaction in Britain: Evidence from the Third National Survey of Sexual Attitudes and Lifestyles (Natsal)

ABSTRACT Previous studies on the relationship between religiosity and sexual behavior have yielded mixed results, partly due to variations by gender and marital status. Furthermore, less is known about this relationship in relatively secularized societies, as in the case of Britain. In this study, we used data from the third British National Survey of Sexual Attitudes and Lifestyles (Natsal-3) to explore the link between religiosity (11% of men and 16% of women stated that religion and religious beliefs were very important to them) and sex frequency and satisfaction among men and women in different types of relationships. Women and men who saw religion as more important in their lives reported having less sex on average, though this was mainly driven by the significantly lower sex frequency among non-cohabiting religious individuals compared to their less religious peers. At the same time, religiosity was linked with overall higher levels of sex life satisfaction. This relationship appeared to be largely mediated by attitudes on the appropriate context for sexual intercourse. These findings highlight the importance of sociocultural norms in shaping sexual behavior and sexual satisfaction.


Introduction
In recent decades, the religious composition of Western European countries has fundamentally changed. Across different countries, many have left religious belief, and particularly younger cohorts now exhibit high levels of secularization in countries such as the United Kingdom (Crockett & Voas, 2006;Stonawski et al., 2015). At the same time, Western nations have seen increases in the ages of marriage and family formation (Abeynayake et al., 2012;Allendorf et al., 2021). Besides secularization, important drivers for this postponement include longer education, later ages at first job, rising housing prices, higher costs associated with having children and new norms for family formation (Brauner-Otto, 2021;De la Rica & Iza, 2005). The postponement of union formation is also associated with a decline in marriage and a rise in the non-married share of the population (Di Giulio et al., 2019;Sobotka & Berghammer, 2021). As formal unions and sexual activity are closely interlinked (Christopher & Sprecher, 2000;Lei & South, 2021), the postponement of marriage could have important consequences for sexual activity. For example, the reported decrease over the past few decades in sex frequency among men and women in the UK and other developed countries has been mainly attributed to the decline in marriage and the increasing proportion of people living without a partner (Lei & South, 2021;Mercer et al., 2013). At the same time, the median age at first intercourse has been declining among younger cohorts in Britain (Lewis et al., 2017), which entails higher exposure to sex outside a formal union. Despite these important potential consequences of religious decline and related societal changes such as new family arrangements, the relationships between religiosity, sex frequency and sexual satisfaction are not well understood.
To date, research on religion and sexual behavior has been mainly done in the context of risk behaviors among adolescents and unmarried young adults in the United States (Burdette et al., 2015). These studies emphasized the role of religion in promoting sexual abstinence and discouraging premarital sex; for example, higher religiosity has been found to be associated with delayed initiation of sexual intercourse (Bearman & Brückner, 2001;Meier, 2003), reduced likelihood of engaging in casual sex (Burdette et al., 2009;Kuperberg & Padgett, 2016), and having fewer sexual partners (Barkan, 2006).
Other studies on the relationship between religion and sexual frequency and satisfaction produced inconsistent findings. This can be partly attributed to differential religious teachings about marital and non-marital sex; while Abrahamic religions discourage sexual activity outside of marriage, sex among (heterosexual) married couples is not only accepted, rather, it is considered a sacred and vital aspect of marriage life (Hackathorn et al., 2016;Hernandez-Kane & Mahoney, 2018;McQuillan, 2004). Thus, as religious individuals are more likely to view sexual intimacy within marriage as having divine properties, this is likely to enhance both frequency and quality of sex among married religious people. In line with this approach, Hernandez- Kane and Mahoney (2018) found that greater sanctification of marital sexuality around the first year of marriage predicted higher sexual frequency as well as greater sexual and marital satisfaction in the following year. Also, according to Dew et al. (2020), religious couples who perceive their marriage as sacred are more likely to hold attitudes and invest in practices that enhance relationship quality, including greater commitment, relationship maintenance behaviors (e.g., expressions of kindness, love, and affection), and time spent together, which in turn contribute to marital sexual satisfaction. Similarly, Waite and Joyner (2001) found that regular church attendance was positively associated with sexual satisfaction for both men and women, though the estimated effect was found to be lower when controlling for attitudes on sexual exclusivity. Thus, higher sexual satisfaction among religious individuals may at least partly be attributed to higher investment in long-term partnerships and relationship intimacy.
In contrast, sexual relationships outside marriage are less acceptable or even considered sinful in many religious traditions; in Christianity, Roman Catholic teachings emphasize that sexual intercourse should only take place within marriage, and that marriage is based on a lifelong commitment between men and women (Richards, 2009). Non-marital sex is also discouraged in Protestant teachings (including the Church of England), which place high value on chastity and marital faithfulness (Creighton, 2009). Similar restrictions on nonmarital sex are also included in Islamic religious texts, and are particularly restrictive of female premarital sex (Dialmy, 2010). This may lead to reduced sexual activity as well as lower sexual satisfaction among unmarried religious individuals.
Previous studies have shown that those who attend church more frequently are more likely to be married, to have lower probability of divorce, and to have more children on average (Berghammer, 2012;Brini, 2020;Halman & van Ingen, 2015). In general, religious individuals hold more negative attitudes toward non-committal lifestyles and sexual behavior (Baker et al., 2020;Pearce et al., 2019). Low levels of religiosity and secularism have also been found to be associated with more positive views of premarital sex, teenage sex, and extramarital sex (Baker et al., 2015;McPhetres & Zuckerman, 2018).
According to Hackathorn et al. (2016), internalization of religious beliefs is associated with sexual guilt, which in turn leads to reduced sexual satisfaction for unmarried people, though this relationship is not found among those who are married. On the other hand, Cranney (2020) found that unmarried religious individuals reported higher levels of satisfaction from sex life than non-religious ones, which could be partly attributed to differential expectations of nonmarital sex within each group. According to Cranney, unmarried religious people are more likely to be celibate or rarely sexual by choice, which is in line with the ideal sexual script for their current marital status. Therefore, despite having less frequent sex than their non-religious peers, they are more likely to express higher sex life satisfaction. Thus, the mixed findings on the relationship between religiosity and sexual satisfaction among unmarried people may be the result of selection, as the inclusion or exclusion of sexually inactive respondents may influence findings.
Some scholars have suggested that religious teachings on sexuality, including norms against non-marital sex, are likely to have a stronger effect on women than on men, due to the greater emphasis on the sexual status of female individuals in religious texts, or higher social pressure to comply with those scripts (Burdette et al., 2015;Rostosky et al., 2004). This assumption was supported by McFarland et al. (2011), who found an inverse relationship between degree of religious integration and sexual activity within the past year among unmarried adult women, but not among men. However, no relationship was found in this study between frequency of religious service attendance and sexual activity for either married or unmarried individuals.
As religion might influence sexual attitudes and behaviors through values and beliefs about the sanctity of marriage and marital sex, it has been suggested that intrinsic aspects of religiosity would matter more to sexual functioning than public expressions of religiosity, such as religious service attendance (Ashdown et al., 2011;Hackathorn et al., 2016). Indeed, studies that analyzed different types of religious measures suggested that personal measures of religiosity, such as private prayer or other in-home religious activities are better predictors of sexual behavior and satisfaction compared to public or institutionalized indicators of religious adherence (Cranney, 2020;Dew et al., 2020). These findings are also consistent with the study by Pargament (2002) on religion and well-being. According to Pargament, religious involvement is more likely to have beneficial consequences on well-being when it is based on intrinsic motivation and internalized beliefs than when it is being externally imposed on individuals. Thus, the implications of religion on sex frequency and satisfaction may vary across different aspects of religiosity, and by the given relationship status.

The Intersection of Religiosity, Sex Frequency, and Sexual Satisfaction
While sex frequency has consistently been found to be correlated with satisfaction from sex life, recent studies have suggested that this relationship is not necessarily straightforward. For example, Schoenfeld et al. (2017) found that affectionate, supportive, and caring behaviors between spouses -e.g., nonsexual expressions of affection -were linked to both higher sex frequency and sexual satisfaction. Thus, they concluded that in order to understand variations in sexual satisfaction, the broader relationship climate should also be taken into account. Other studies have also emphasized the importance of relationship dynamics to sexual satisfaction. For instance, Waite and Joyner (2001) have shown that men and women who expected their relationship to last longer had higher emotional satisfaction and physical pleasure from sex than those who perceived their relationship as a short-term one. Similarly, a qualitative study on sexual experiences in New Zealand reported that both men and women expressed ambivalence toward casual sex, which was often described as "unfulfilling," and stated an overall preference for sex within a long-term relationship (Farvid & Braun, 2017). These findings are consistent with a study by Muise et al. (2016), showing that sex frequency was significantly associated with happiness only among those who were in a romantic relationship.
As previous studies have demonstrated, religiosity is associated with having less permissive attitudes toward extramarital sex, including the importance of sexual fidelity, and the view that sex should only occur when it is motivated by love or a wish to have children (Hardy & Willoughby, 2017;Iveniuk et al., 2016). Furthermore, religiously committed individuals show higher preference for marriage over other forms of relationships (Henderson et al., 2018;Lehrer, 2004), and are less likely to engage in sex outside a long-term committed relationship (Burdette et al., 2009;Kuperberg & Padgett, 2016). These relatively traditional approaches to romantic relationships and sexuality could therefore explain some of the differences found in sexual satisfaction by religiosity.
While previous research pointed to the differential implications of religion on sexual behavior and satisfaction among married and unmarried individuals, less is known about this association in other contexts of committed relationships (e.g., non-marital cohabitation or living apart together). For example, approval for pre-marital sex among more religious people may be higher when it occurs within a committed relationship that is likely to lead to marriage (Uecker, 2008). Furthermore, the vast majority of studies on religion and sexuality were conducted in the United States, where the proportion of religiously observant people is substantially higher compared to that in most other Western countries (Evans, 2018). Less attention has been given to this relationship in more secularized societies. In Britain, the proportion of people affiliated with a religion declined from around two-thirds during the 1980s to just under half of the adult population in 2018 (Curtice et al., 2019). About 40% of British people are identified as Christians, which includes Anglicans (Church of England), Roman Catholics, and other Christians, while close to a tenth of the population are affiliated with non-Christian denominations, including Muslims, Hindus, Sikhs, and other religions.
In terms of ethnic diversity, about 85% of the population in England and Wales identify as White, and the main ethnic minority groups include people of South Asian origin (close to 8%), and those who identify as Black African or Black Caribbean (about 3%; Office for National Statistics, 2021). According to previous studies, ethnic minority populations in Europe tend to be more religious than the majority group (Kaufmann et al., 2012;Stonawski et al., 2016), and hence, may have a less accepting view of non-marital sex than White British populations.
In the present study, we examined the link between religiosity and sex frequency, and the relationship between religiosity and reported level of satisfaction from sex life among British men and women in varying relationship statuses. In addition, we explored the role of sexual attitudes and behaviors in explaining differences in sexual satisfaction by religiosity. Given the rapid changes to the religious composition of Britain and other post industrialized societies, alongside the ongoing postponement of union formation and other developments in partnering dynamics, it is increasingly important to address the relationships between religion, sex patterns, and predictors of sexual satisfaction.

Research Hypotheses
As previous studies indicate that more religious individuals have lower sexual activity outside of marriage or other types of committed relationships, our first hypothesis was that individuals with a higher level of religiosity would have less frequent sex outside formal unions (either non-marital cohabitation or marriage) compared to their less religious peers.
Avoidance of sex outside marriage, as well as lower approval of casual sex, infidelity, or sex without love, are in line with religious norms and values about the sanctity of marriage and marital sex. These norms may yield differential expectations from sex life among religious and non-religious individuals and, among the former, conforming to these norms by limiting sexual activity outside a committed relationship and investing more in long-term relationships is likely to contribute to a positive view on one's sex life. These norms are in accordance with religious teachings from major religions, which emphasize that one should avoid excesses and be content. For example, for Christians and Jews, three of the Ten Commandments relate to being content and satisfied with what one has (You shall not commit adultery; You shall not covet your neighbor's wife; You shall not covet your neighbor's good). Therefore, our second hypothesis postulated that more religious individuals would have higher levels of satisfaction from sex life within marriage.

Data and Methods
To test these hypotheses, we used data from the National Survey of Sexual Attitudes and Lifestyles (Natsal-3), which was conducted in 2010-2012 (Johnson & University College London, Centre for Sexual Health and HIV Research, 2021). This is a large-scale stratified probability sample survey of 15,162 men and women, which is representative of the resident population in Britain aged 16 to 74. Participants completed the survey through a combination of computer assisted face-toface interviews alongside computer assisted self-interviews for the more sensitive questions that involved sexual experiences and sexual function . The response rate for Natsal-3 was 57.7%, which is similar to that of other major social surveys conducted in Britain around the same time .
For the purposes of the current study, we capped the bottom age limit of our sample at 18, to include only adults and to minimize the problem of changes in religious participation during adolescence (Petts, 2009). In addition, we excluded those aged 60 and over, since the prevalence of sexual activity tends to decline at older ages, and older adults are more susceptible to health conditions that affect sexual functioning (Camacho & Reyes-Ortiz, 2005;Lindau et al., 2007). Furthermore, to reduce heterogeneity bias, we did not include those who identified as attracted only or mostly to people from the same gender. 1 The dependent variables in our study were sex frequency and level of satisfaction from sex life. Sex frequency was measured by the number of occasions of heterosexual sex in the last 1 This group formed about 2% of respondents aged 18-59.
four weeks, including vaginal intercourse, oral sex, and anal sex. This question was addressed to all respondents, where those who had never had sex received the value of zero. The question on sexual satisfaction was also addressed to all respondents, regardless of sexual activity, and was phrased as "Thinking about your sex life in the last year, how much do you agree or disagree with the following statements: 'I feel satisfied with my sex life.'" The answer categories to this question were on a five-point Likert scale, ranging from "Disagree strongly" to "Agree strongly." In addition, we conducted a sensitivity analysis with the outcome measure of sexual function among those who were sexually active. Sexual function was measured using the Natsal-SF, a 17-item validated measure especially developed for Natsal-3 (Mitchell et al., 2012). This measure incorporates physiological, psychological, and relational aspects of sexual function for those who have had sex in the past year, including interest in sex, enjoyment and arousal during intercourse, sexual difficulties (e.g., erectile dysfunction for men, dry vagina for women), feeling emotionally close to partner during sex, sharing the same sexual likes and dislikes, and general satisfaction from sex life. For the purpose of this study, those in the lowest quintile of the resulting scores for the Natsal-SF, were classified as having lower sexual function.
The key independent variable in our analysis was religiosity. Religious adherence may include different aspects, such as religious beliefs, religious practices, and importance of religion in one's life. Since intrinsic measures of religiosity have been found to be more important determinants of sex behavior and satisfaction than public expressions of religiosity (Ashdown et al., 2011;Cranney, 2020;Dew et al., 2020;Hackathorn et al., 2016), we used a measure of subjective religiosity. In the survey, respondents were asked "How important are religion and religious beliefs to you, now?" with the following answer options: very important, fairly important, not very important, and not important at all. In addition, for comparison purposes, we ran a separate analysis using frequency of religious service attendance. This measure included three levels: monthly attendance, yearly attendance, and those who never or almost never attended religious services. We also accounted for religious denomination, which included the following categories: no religion, Anglicans (Church of England), Roman Catholics, other Christians, and non-Christian denominations.
Apart from religion, we also controlled for sociodemographic characteristics, including age, relationship status, presence of children in the household, level of education, and ethnicity. Relationship status included the categories of unpartnered (which included those who had never had a partner, or did not currently had a partner, as well as those who have casual sex partners), living apart together (LAT) -to represent those in a steady non-cohabiting relationship, non-marital cohabitation, and married. Children in the household included three binary variables to represent the presence of children (including biological and non-biological) in various age groups: children aged 0-1, children aged 2-11, and children aged 12 or older. The reason for that was to control for potentially varying effects of having a newborn, young children, and older children in the household on sexual relationships. Educational attainment included the following categories: no qualifications, lower secondary (corresponding to General Certificate of Secondary Education (GCSE) or Ordinary Level (O-level) qualifications), typically completed at the age of 16), upper secondary (corresponding to Advanced Level (A-levels) qualifications, typically completed at the age of 18), and degree level education (bachelor's degree or above). The measure for ethnicity was based on the main ethnic groups in Britain, and included "White," "South Asian," "Black," and "other." Furthermore, we controlled for subjective health status; the question addressed to respondents was formulated as "How is your health in general? Would you say it is . . . very good, good, fair, bad, or very bad?." Since only a small number of respondents rated their health as very bad (less than 1%), the latter two categories were merged. In addition to health status, we also controlled for whether the respondent was taking any antidepressant drugs, which are known to be closely associated with sexual dysfunction (Clayton & Montejo, 2006). Another relevant covariate was whether the respondent was currently trying to conceive, which may affect both sexual frequency and satisfaction.
In order to examine the role of sexual experiences and attitudes on the level of satisfaction from sex life, we included a measure of the total number of lifetime sexual partners (including same and opposite sex partners), as well as level of approval of casual sex, and sex without love. The first was phrased as "What is your opinion about a person having one night stands?" with the following answer categories: "Always wrong," "Mostly wrong," "Sometimes wrong," "Rarely wrong," "Not wrong at all," and "Depends/don't know." The second was phrased as "It's OK to have sex with someone without being in love with them," with a five-point scale of agreement from "Disagree strongly" to "Agree strongly."

Analytical Approach
First, we estimated the distribution of sociodemographic and sexual characteristics for men and women in our sample. For this purpose, we employed weights provided in the Natsal survey to account for sample design and non-response . We also tested for gender differences for each of these variables. To examine religious differences in the frequency of sexual intercourse, we ran a separate regression analysis for men and women, estimating the number of sex occasions in the past four weeks by level of religiosity. The regression models also controlled for religious affiliation, socio-demographic factors, and health status. As sexual activity among religious individuals was expected to vary by type of relationship status, we also analyzed each category separately (unpartnered, living apart-together, cohabiting, and married). Since the number of sex occasions in the past four weeks was a count variable and was overdispersed (i.e. the variance was larger than the mean), an ordinary least squares regression was less appropriate and might lead to biased results. We therefore opted for a negative binomial regression model, which accommodates for count outcomes without being susceptible to overdispersion (Long & Freese, 2006). Next, we explored the relationships between religiosity and level of satisfaction from sex life by running a multivariable ordered logistic regression model. The basic model included the same covariates as in the negative binomial regression, with additional controls for sex frequency in the past four weeks and its squared term. This was in order to account for non-linearity in the relationship between sex frequency and sexual satisfaction. The number of lifetime sexual partners, and sexual attitudes (including approval of casual sex and sex without love), were then added to the full model, to examine their implications on the estimates of religiosity in relation to sexual satisfaction. Furthermore, as the number of lifetime sexual partners was found to be relatively lower both among younger and older respondents, we included a squared term for age. The regression analysis was repeated for each relationship category, to explore potential differences in the link between religiosity and sexual satisfaction by relationship status. We also conducted a robustness analysis, using an alternative outcome measure of sexual function, which was based on selfappraisal of one's sexual relationship and sexual functioning (see above description), within a sub-sample of sexually active individuals. This analysis was conducted using a logistic regression model, where those in the lowest quintile in the sexual function scale received the value of 1, and all others received the value of 0. Additional robustness analyses were conducted using a measure of religious attendance instead of subjective religiosity for predicting sex frequency, sexual satisfaction and sexual function.
Finally, estimating multiple models for each relationship status can increase the probability of obtaining a significant result by chance (Streiner & Norman, 2011). Therefore, we addressed this issue by applying a Bonferroni correction in the following way; since there were five different versions of each model (all men/women and four additional models for each relationship status), the threshold for a significant result was set to P< .01 (α = 0.05/5 = 0.01). Table 1 presents the distribution of the sociodemographic characteristics for men and women in our study. It is shown that most respondents stated that religion was not very important or not important at all to them, with a minority of 11% of men and 16% of women saying that religion and religious beliefs were very important to them, and 22% and 27% respectively regarded religion as fairly important (based on weighted proportions). Similarly, over two-thirds of respondents reported that they never or almost never attended religious services, and 14-17% of men and women (respectively) attended religious services at least once a month. Furthermore, around half of respondents reported not being affiliated with any religion. Around 40% of respondents stated a Christian affiliation (Anglican, Roman Catholic, or other), and less than a tenth of respondents were affiliated with a non-Christian religion. Across all religious measures, women showed significantly higher levels of religiosity compared to that of men. These figures largely correspond with findings from other major social surveys in Britain, such as the British Social Attitudes survey (Curtice et al., 2019). In terms of ethnicity, the vast majority of respondents were White (86-88% of men and women, respectively). Around 6% were of South Asian origin, 4% were Black, and 3% were classified as other. These proportions correspond with data on ethnicity from the 2011 census, except for a slight under-representation of those from South Asian origin .

Results
Nearly half of respondents were married, and a further 17% lived in non-marital cohabitation. Over a fifth of respondents were not in a steady partnership (unpartnered), and just over a tenth of respondents were in a noncohabiting steady relationship (living apart together). A higher proportion of women had children aged 2-11 or older children in the household compared to men, though a higher proportion of men reported they were currently trying to conceive (9% of men compared to 5% of women). Over 80% of respondents reported being in a good or a very good health, with no significant differences by gender. On the other hand, a significantly higher proportion of women reported taking anti-depressant drugs than men (9% compared to 4%, respectively).
The descriptive statistics for sexual behaviors and attitudes for men and women are presented in Table 2. On average, men reported higher frequency of sex occurrences in the past four weeks compared to women (4.4 compared to 4.0, respectively). However, women reported higher levels of sexual satisfaction compared to men; 25% of women expressed strong agreement with the statement "I feel satisfied with my sex life" compared to 24% of men, while 14% of women and 17% of men stated they disagreed or strongly disagreed with that statement.
Only 2% of men and women in our sample reported that they had never had any sexual partners. Among those who had sexual partners, the number of lifetime partners was considerably higher for men compared to women, where nearly 40% of men reported having ten or more sexual partners in their lifetime compared to a quarter of women. Differences between men and women were also shown in relation to sexual attitudes, as men expressed higher approval of casual sex and sex without love than women; 65% of men stated that they either agreed or strongly agreed that it was okay to have sex with someone without being in love with them, compared to 50% of women. Furthermore, women were more likely to disapprove of casual sex than men, with about a third of women saying that one-night stands are "always wrong" compared to a quarter of men.
Next, we present the results of the multivariable regression analysis for sex frequency in the past four weeks among men and women. The findings from the negative binomial regression model show that men and women who stated that religion was very important and men who stated that religion was fairly important to them, had significantly less frequent sex compared to those who stated that religion was not important at all (see Tables 3 and 4, Model 1). However, when observing the results for each relationship category separately, it is shown that the negative relationship between religiosity and sex frequency was only significant for those who are unpartnered or in a non-cohabiting steady relationship (living apart together). By contrast, no significant relationship was found between religiosity and sex frequency among cohabiting and married people (see Tables 3 and 4, Models 2-5). These findings are in line with the first hypothesis of the study, according to which individuals with higher religious devotion would have less sex outside marriage.
Other notable findings include the significantly lower sex frequency among women with ethnic minority background compared to those identifying as White. As in the case of religiosity, this relationship also appeared to hold only among those who were unpartnered or living apart together. The regression analysis also revealed a curvilinear relationship between education and sex frequency, as men and women with lower or upper secondary education showed higher frequency compared to those with degree level education (Tables 3 and 4, Model 1).
When the measure of religious importance was replaced with religious service attendance, the findings showed a similar pattern of reduced sex frequency among those attending religious services at least once a month, compared to those who never, or almost never attended religious services. However, among men, significant differences in sex frequency by religious service attendance were only found for those in a non-cohabiting steady relationship, and for women, the significant relationship was found among those who were unpartnered. Results from this analysis are available in Supplemental Table S1(a,b).
In what follows, we explore the relationship between religiosity and level of satisfaction from sex life for men and women by relationship status. Tables 5 and 6 present the results of the ordered logistic regression for satisfaction from sex life among men and women. The findings point to a generally higher satisfaction from sex life among those who considered religion as fairly or very important compared to those stating religion was not important at all. In line with the second hypothesis, more religious married women reported higher sexual satisfaction than less religious women, though no significant differences in sexual satisfaction by religiosity were found among married men. However, a positive relationship between religiosity and sexual satisfaction was found among unpartnered and cohabiting men. It should be noted that when using religious service attendance instead of subjective religiosity, no significant differences in sexual satisfaction were found across the different relationship categories (not shown).
Tables 7 and 8 present the results of the ordered logistic regression for sex life satisfaction after controlling for number of lifetime sexual partners, and attitudes to casual sex and sex without love. It is shown that after including these attitudes, the positive relationship between religiosity and sexual satisfaction largely disappears for both men and women. One exception to that was among cohabiting men, where those who stated that religion was fairly important to them, still reported higher satisfaction from sex life compared to the least religious group. For women, having no sexual partners, as well as having ten or more lifetime sexual partners, was associated with lower satisfaction from sex life. Among men, on the other hand, no relationship was found between the number of lifetime sexual partners and sexual satisfaction. However, disapproval of sex without love and of casual sex was linked with higher satisfaction from sex life among both men and women.
Another finding of interest is the curvilinear relationship found between sex frequency and sexual satisfaction; while sexual satisfaction initially increases with sex frequency, it declines again at a higher number of sex occasions. Therefore, having "too much" sex may lead to lower levels of satisfaction from sex life.
Next, we show findings from the robustness analysis on the relationship between religiosity and sexual function, which combines physiological, psychological, and relational aspects of sex life for those who have been sexually active in the past year. Tables 9 Table 2. Distribution of sexual attitudes and behaviors.  Ref.
Ref.  and 10 present the findings from the logistic regression analysis, estimating the likelihood of being in the lower quintile of the sexual function scale, controlling for sociodemographic variables and sex frequency in the past four weeks. According to these findings, religiosity was not related to sexual function among men. However, among married women, those who considered religion as fairly important showed a lower likelihood of being in the lowest quintile of sexual function compared to women who stated that religion is not important at all. These findings provide further support for the positive association between religiosity and sexual satisfaction among married women. In addition, as in the analysis for sexual satisfaction, after controlling for lifetime sexual partners and sexual attitudes, the differences in sexual functioning by religiosity among married women were no longer significant (not shown). Furthermore, in contrast to subjective religiosity, religious service attendance was not found to be associated with sexual function (not shown).

Discussion
Our findings suggest that both sex frequency and sexual satisfaction vary by religiosity, though this relationship differs across types of unions. In accordance with our first hypothesis, among single non-cohabiting individuals, the more religious had less frequent sex compared to their less religious peers. This finding was consistent when religiosity was measured either by subjective importance of religion or religious service attendance. Nonetheless, those who attributed greater importance to religion and religious beliefs reported higher satisfaction from sex life. In line with our second hypothesis, more religious married women reported higher sexual satisfaction than their less religious peers, though this relationship was not found among married men. Interestingly, unpartnered religious men also reported higher satisfaction from sex life, though this relationship disappeared after we included controls for attitudes to casual sex and sex without love, or when the sample was limited to sexually active respondents. In addition, we found a non-linear relationship between number of lifetime sexual partners and sexual outcomes for women, where having no or many partners was linked to lower sexual satisfaction. Higher approval of casual sex or sex without love was also found to be negatively associated with sexual satisfaction for both men and women. As our study shows, the relationship between sex frequency and sexual satisfaction is neither simple nor straightforward; across all relationship types, too little or too much sex was associated with lower sexual satisfaction, suggesting that an optimum exists in terms of frequency related to higher satisfaction levels. This is in line with Kornrich et al. (2013, p. 18), who argued that "couples are not purely interested in the amount of sex they have -they undoubtedly also care about the quality of sex." Previous studies have shown that increased investments in exclusive long-term partnerships and greater time to develop satisfactory trusting relationships can matter for sexual satisfaction, while sex outside a committed relationship is often related to lower sexual satisfaction (Farvid & Braun, 2017;Waite & Joyner, 2001). As religious individuals are less likely to engage in casual sex (Burdette et al., 2009;Kuperberg & Padgett, 2016), and are more likely to limit sexual activity to a relationship based on love (Hardy & Willoughby, 2017;Iveniuk et al., 2016), this can lead to lower expectations of sexual activity outside a formal union, as well as increased satisfaction from sex life in general.
However, it is possible that religious sentiments about the sanctity of marital sex, as well as disapproval of sex outside marriage, matter more for women's than for men's sexual satisfaction. This is also evident by the relatively higher levels of sexual satisfaction among more religious cohabiting men when all other variables were held constant, while no similar relationship was found among cohabiting women.
As expected, the findings on ethnic minority groups showed similar patterns to that of more religious people, as women who identified as South Asian or Black reported lower sex frequency compared to women who identified as White. Furthermore, this relationship appeared among women who were unpartnered or in a steady non-cohabiting relationship, but not among those who were cohabiting or married. According to Krull et al. (2021), since ethnic minority groups can be at a relatively disadvantaged position, having sex outside a stable union and the prospect of unintended pregnancy could be perceived as particularly risky and stigmatizing.
Our findings also showed a significant association between educational attainment and sexual frequency and satisfaction; overall, highly educated individuals reported having less frequent sex, as well as reduced satisfaction from sex life compared to those with lower qualifications. This may be the result of several factors, including higher work load among the highly educated, greater work related stress levels, or increased investment in labor market capital and careers over relationshipbased capital (Abdoly & Pourmousavi, 2013). However, the complex pathways underlying the relationships between education and sexual outcomes require further investigation, which is beyond the scope of this study.
Our research suggests that changes in sexual behavior need to be understood in a context of changes in religious norms and beliefs and other societal level trends. The postponement of union formation is related to less frequent sex, while also increasing the exposure to casual sex among those with weaker religious orientation. Therefore, the decline in religiosity and the rise in the single population are likely to exacerbate these trends, which may potentially result in lower sexual satisfaction.
Our study had several strengths. We used representative data and focused on a topic that so far has received insufficient attention in sex research -the role of religion, and how religiosity relates to sexual frequency and sexual satisfaction. There is a scarcity of studies which have looked at religion by relationship type and our study did this. We believe our findings and analyses can provide valuable and novel insights that can be of use for scholars interested in the intersections of sex and religion in contemporary societies.
The present study had some limitations in terms of information in our dataset. For example, we lacked information on religiosity and religious beliefs during childhood, which meant that we could not investigate how religion changes over the life course, and how this relates to sexual behaviors. Further, the dataset did not include detailed information on physiological and mental illnesses and disease histories, health risk factors, personality, and labor market histories -all of which may relate to both sexual behavior and religious trajectories. In addition, the Natsal-3 survey relied on self-reported data, which may be subject to desirability bias. However, this bias is minimized by the use of computer-assisted self-interview technology for the questions concerning sexual experiences and sexual function . Another potential limitation is the underrepresentation of Asian men and women in the sample, given the observed differences in sexual behavior between British Asian and the population of British White, who form the majority in the UK. Nevertheless, the Natsal-3 sample is still largely representative of the resident population in Britain.
Recent decades have seen widespread secularization, with declines in religiosity and decreasing levels of religious affiliation in Western countries. At the same time rapid changes in family forms have taken place, with later transitions into stable relationships, higher proportions not forming families, more cohabitation, increased levels of family dissolution, and greater proportions remaining single in younger adulthood than earlier. These changes in the religious and demographic makeup of the United Kingdom and other Western countries can have implications on many life domains, including sexual activity patterns and sexual satisfaction levels.
Given continued societal level changes in terms of demography, living arrangements, religiosity, and education in a context of population aging, one needs a broad research approach in order to better foresee future developments and consider ways that can improve sexual satisfaction. It is therefore necessary to collect detailed longitudinal data on sexual attitudes and behaviors which includes information on religiosity -and study these. Health and individual characteristics, but also normative and faith-related factors can have important effects. Future studies should pay more attention to religion when assessing sexual behavior and satisfaction, including when studying population level trends and differences among population subgroups.