Christopher G. Ellison and Xiaohe Xu
Social scientists are once again actively interested in the connections between religion and family, two institutions that have enjoyed a long and complementary relationship (Wilcox, 2004; Edgell, 2006). This interest can be observed in the increased number of works on this topic that have appeared in recent years in major peer-reviewed outlets in sociology, family studies, psychology, religious studies, and allied fields and the significant volumes now published on prominent university and trade presses. In this chapter, we will review the existing theoretical and empirical literature linking multiple dimensions of religion with various facets of family life.
Several caveats are in order. First, because much of the work on this topic has been conducted about and within the United States, our review will necessarily concentrate on this context, and we will return to this as a limitation in the concluding section. Second, most studies on religion and family in the United States and other Western societies center on the Judeo-Christian tradition, and consequently, our work will emphasize this theme as well. Third, although there are important literatures on religious variations toward family-related issues (e.g., women’s roles, equal rights for gay, lesbian, bisexual, and transgender persons), due to space limitations, we will focus primarily on specific family-related behaviors.
The remainder of this chapter is organized as follows. First, we will briefly review key patterns and trends in religion and spirituality in the United States over the past 25 years. Then, we will review recent theory and research linking religion with the following areas of family life, organized primarily by life course stage: (i) premarital interactions, (ii) cohabitation and timing of marriage, (iii) relationship dynamics and quality, (iv) childbearing, (v) child-rearing, and (vi) intergenerational relations. Finally, we will call attention to substantive and methodological issues and gaps in religion–family research. We will underscore the importance of research in non-Western contexts, where attention to religion has often been consigned to studies of fertility and related issues.
Because much of the research on religion and family has been conducted in US context, it is appropriate to begin with a very brief glimpse of the contemporary American religious scene. Although much of the developed West has been heavily influenced by secularization – manifested by diminished public role of religious institutions and values and in many settings by reduced individual religious commitment and belief – for the most part, the United States has been regarded as a stubborn outlier (Norris and Inglehart, 2004). The United States is also characterized by substantial religious diversity; it has been estimated that more than 2000 religious groups and faith traditions are active in the United States. Despite the absence of any single official source of data on religion in the United States, it is possible to combine data from a number of nationwide surveys of individuals and congregations to reach several broad conclusions (Chaves, 2011; Ellison and McFarland, 2013).
First, when asked, most American adults report an affinity (but not necessarily affiliation) with a Christian religious group or tradition. Specifically, approximately 30% of Americans associate themselves with conservative (evangelical, fundamentalist, or charismatic) Protestant groups (e.g., Southern Baptist, Assemblies of God, many smaller denominations and independent churches, as well as many nondenominational groups, including many of the rapidly growing megachurches). Significant proportions of African Americans, as well as increasing numbers of Latinos and Asian Americans, belong to such conservative faith communities. These groups and many of their members tend to affirm the inerrancy and authority of the Bible as well as beliefs about Original Sin, the imperative of salvation through grace, and others (Hempel and Bartkowski, 2008). Around 12% of Americans are affiliated with moderate and liberal Protestant groups (e.g., Episcopal, Methodist, most Presbyterian churches, etc.). In general, the more conservative groups have gained membership and social influence over the past 30–50 years, while the more moderate and liberal groups have declined. Approximately 25% of US adults self-identify as Roman Catholics. The percentage of US adults reporting no religion on surveys has climbed to roughly 20%. The remainder (13%) of the American public consists of members of non-Christian faiths (e.g., Judaism, Islam), sectarian groups (e.g., Mormons, or Latter-Day Saints), and a myriad others. One of the fastest growing segments of the US population self-identifies as spiritual but not religious. This group – now approximately 25% of the adult population and higher among younger cohorts – includes some individuals who retain tenuous denominational identities, as well as many who have rejected organized religion entirely.
Second, religious affiliation, practice, and belief are socially patterned. For example, scholars have long referred to the social sources of American religious affiliation, especially region and socioeconomic status (Ellison and McFarland, 2013). Conservative Protestants have been historically concentrated in the Southeast and lower Midwest, while Catholics have long maintained cultural strongholds in the Northeast, upper Midwest, and Southwest. Liberal Protestants have been concentrated in the Northeast, while Jews have typically resided in coastal urban contexts. Mormons have long dominated Utah, and now maintain majorities or pluralities in many other parts of the Mountain West. With respect to education and SES, conservative Protestants have historically lagged behind most other groups in education, income, and wealth. This has also been true of Catholics, although they have experienced considerable upward mobility across generations. Liberal Protestants and Jews have traditionally enjoyed comparatively high levels of education and social and economic standing. Although many of these social sources remain influential, specific patterns are shifting under the weight of geographical mobility, assimilation, intermarriage, and other factors. Moreover, during the past few decades, a number of factors have shifted the focus of many Americans’ religious attentions and loyalties away from denominations and toward local congregations (Chaves, 2011).
It has proven to be difficult to estimate levels of religious attendance with precision, due to the biases inherent in self reports of behavior in general, and especially this particular behavior (Chaves, 2011). Observers cautiously estimate that perhaps 25% of US adults attend some type of religious service in a given week. Large majorities of Americans pray weekly or more, while the frequency of meditation is much lower. Two-thirds or more of US adults say that religion is very important to them. In terms of core doctrine, levels of support for the concepts of Heaven and Hell (especially the latter) have declined over the past few decades, while beliefs in some type of God and some type of afterlife have remained strong and stable (Schwadel, 2011).
By most indicators of personal religiosity (e.g., attendance at services, prayer, religious salience), women are more religious than men. Comparatively high levels of religiousness are found among racial and ethnic minorities, married people (especially those with children), and older adults. Contrary to many popular impressions, the associations between religious involvement and socioeconomic status are uneven. For example, income and education are positively associated with the frequency of attendance and the occupation of church leadership positions. On the other hand, less educated and less affluent persons tend to engage in more frequent devotional activities (e.g., prayer, scriptural study) and to hold more conservative or orthodox religious beliefs (Ellison and McFarland, 2013).
In recent years, social scientists have assessed trends in American religion by decomposing age, period, and cohort influences. Overall, most forms of religious practice and belief increase modestly with age. Controlling for these age-related patterns, recent studies have revealed significant cohort-based declines in the likelihood of regular attendance, weekly prayer, and belief in biblical inerrancy, which have accelerated for those born after the mid-1940s (Schwadel, 2011). It is also important to note that Americans’ confidence in organized religion has been on the decline for some time, and is plummeting even more rapidly than their confidence in most other social institutions (Chaves, 2011). Thus, although the United States clearly remains one of the most conventionally religious nations in the developed West, there are also signs of change and possible secularization on the American religious scene.
At the broadest level, can religion be said to influence family life? If so, how might this occur? First, denominations and faith traditions may embrace distinctive religious teachings regarding the family or aspects thereof. For example, conservative Protestants tend to believe that the Bible is the inerrant Word of God and that it is the authoritative source of guidance in all human affairs; that is, levels of support for these core doctrines are significantly higher in conservative Protestant denominations than in other religious groups. In practice, however, not all parts of the Bible receive equal emphasis among self-described inerrantists. Given their beliefs about the doctrine of original sin and the primacy of the (heterosexual) family unit, conservative Protestant readings of the Bible often stress passages about human sexuality (e.g., fornication, adultery, homosexuality), marital and gender roles, and other family-related themes (Hempel and Bartkowski, 2008). Mormons (Latter-Day Saints) have strong beliefs about marriage, sexuality, and procreation as well, all of which are intertwined with a very distinctive view of the family as an eternal unit. These and other faith communities may constitute subcultures within which specific understandings of family life are socialized.
In addition, religious participation and salience may also have implications for family attitudes and behaviors. Although certain denominations and traditions hold particularly distinctive ideologies, most religious groups attempt, to varying degrees and with varying levels of effectiveness, to guide the morality and behaviors of their members. This may be especially true in certain areas, such as sexuality and other family matters. The frequency of attendance at services may reflect one’s level of religious commitment, because attendance requires the expenditure of scarce resources such as time and energy (and may involve financial contributions as well). This form of religious involvement may also reflect the degree of one’s immersion in religious teachings about the family, which may be communicated formally (via sermons and religious education classes) and informally (via social network ties).
Compliance with religious teachings about family-related matters may result from the threat of shame due to the actual or prospective violation of internalized norms, among persons for whom religion is important, a significant component of their personal identity. In addition, for those enmeshed in networks of coreligionists, deviance may be deterred by social sanctions, or rewards for desirable conduct, or diminished opportunities within religious networks for certain types of deviant behavior, such as illicit sexual conduct (Hoffmann and Bahr, 2005). Of course, it is also possible that apparent religious effects may partly reflect the role of selectivity, as individuals who do not wish to adhere to the guidelines of a particular faith (or any religion) may switch denominations or abandon organized religion altogether.
Dating is an extremely important element of the lives of adolescents. In contrast to the patterns observed among previous cohorts of youth, who tended to transition into marriage and childbearing at much earlier ages, today’s adolescents and young adults engage in a succession of premarital romantic relationships. Although the nature of these relationships can vary widely, they are often characterized by emotional attachment and physical (and sometimes sexual) attraction, and they meet a number of needs and desires, ranging from companionship, personal discovery, and anticipatory socialization. Researchers have linked dating experiences with subsequent family and life course transitions as well as adolescent mental health and well-being. Although investigators have explored a number of influences on dating behavior, with few exceptions, the role of religion has received short shrift.
Recent research suggests that this has been a significant oversight. Work by Bartkowski, Xu, and Fondren (2011) revealed a number of noteworthy links between teen and family religion and dating behavior. For example, they have shown that members of conservative Protestant and sectarian groups tend to date more often than others and do so primarily with coreligionist partners. Jewish teens date less often, on average, but tend to engage in comparatively high levels of interfaith dating. Teen beliefs in the religious exclusivity, that is, the superiority of their own faith, are associated with less frequent dating and lower levels of interfaith dating. Although teens’ organizational religious involvement and family religious engagement are weakly and inconsistently associated with the frequency of overall dating and interfaith dating, more religiously devoted teens are less inclined to date persons from other faiths.
Over the years, a much larger body of work has explored the links between religious factors and adolescent sex, with a particular focus on the timing of coital debut and to a lesser extent the frequency of sexual activity and the number of sexual partners (e.g., Hardy and Raffaelli, 2003; Meier, 2003). Most researchers have concluded that religious teens hold more conservative attitudes about premarital sex, due partly to parental religious socialization (Pearce and Thornton, 2007), tend to delay their sexual debut, and to have sex less often and with fewer partners than other adolescents. However, investigators have differed over the precise nature of these patterns, as well as denominational variations in adolescent sexual behavior, due to differences in data sources and analytic strategies.
Perhaps, the most exhaustive project in this area has been executed by Regnerus (2007), who combined the data from multiple nationwide surveys with the data from a large number of in-depth interviews with a representative sample of teens. Although his results were complex and defy easy summary, his key results showed that, despite their distinctive sexual scripts, for example, strong beliefs and rhetoric regarding teen chastity, conservative Protestant sexual behavior does not differ greatly from that of other religious groups. Mormon (Latter-Day Saint) youths are more consistent in their conservative beliefs and practices. Jewish youth combine liberal attitudes with relatively conservative sexual behaviors in most areas. In addition, consistent with much previous research, religiosity, defined as the frequency of religious attendance and the degree of religious salience, is a much stronger predictor of most sexual attitudes and behaviors than denominational identification.
Regnerus also showed that the virginity pledge movement, started by conservative Protestants and exemplified by such groups as True Love Waits, has involved only a small minority of adolescents (Bearman and Bruckner, 2001). While not entirely curtailing premarital sex even among pledgers, the movement has nevertheless had several notable effects. On average, pledgers delay first sex longer, have fewer sexual partners prior to marriage, and choose partners who themselves have had fewer sexual partners.
Finally, and somewhat paradoxically, Regnerus showed that new forms and rationales of sexual restraint are emerging among youths who come from those religious groups with comparatively relaxed sexual mores, especially Jews and liberal Protestants. Adolescents from these religious backgrounds (and especially males) appear to be comparatively likely to engage in oral sex, masturbation, and pornography (especially from the Internet) as substitutes for vaginal intercourse. According to Regnerus, the motivations for these patterns may be largely strategic, and may reflect the social class composition of these religious denominations. This new middle-class sexual morality, as he labeled it, is aimed at avoiding the risks associated with early sexual activity in order to enhance the likelihood of educational completion and the formation of rewarding nuclear families.
One important topic that has received recent attention in the literature is the phenomenon of hooking up among young adults, particularly on college campuses. Hookups are loosely defined as casual, unplanned physical liaisons, often sexual in nature, with no expectation of any future relationship. A number of scholars have observed that such encounters have largely supplanted more formal dating among college students and young adults. Although some commentators view these practices as relatively benign, critics have expressed concern over the potential negative consequences for social, emotional, and physical well-being, especially for women (Freitas, 2009). Researchers have identified religious variations in hooking up using both quantitative and qualitative data. For example, in one sample of 1000 college women, Burdette et al. (2009) found that (i) Catholic women were more likely to have hooked up than others; (ii) conservative Protestant women were less likely to have hooked up, largely due to their higher levels of religious attendance; and (iii) women attending Catholic institutions were especially prone to have hooked up, over and above the estimated net effects of individual affiliation. Although there were limitations to their data, the results dovetailed with the findings of other research on hooking up (Freitas, 2009), and with broader observations that the depth of religious and moral socialization of young Catholics may be declining (Smith and Denton, 2005). It has been suggested that conservative Protestant colleges, by contrast, are relatively successful in deterring the hookup culture.
A significant body of theoretical and empirical work has demonstrated that more religious individuals, and especially those with conservative Protestant affiliations and beliefs, were more likely to oppose cohabitation and to embrace promarriage attitudes (e.g., Pearce and Thornton, 2007; Ellison, Acevedo, and Ramos-Wada, 2013). Although numerous studies document associations between multiple dimensions of religiousness and family-related attitudes, does religious involvement actually shape patterns of cohabitation and marriage? The available evidence suggests an affirmative answer to this question. Young adults with low rates of religious salience were much more likely to cohabit than others, and less likely to marry. Parental religiosity also influenced these behaviors, largely (but not entirely) by shaping their religiosity. Religious affiliation effects on cohabitation and marriage were weak and inconsistent (Thornton, Axinn, and Hill, 1992). Fine-grained analyses of religious effects on marriage and cohabitation have generally yielded several conclusions. Conservative Protestants, especially those with high levels of personal religiosity, and Mormons were particularly unlikely to cohabit. Individuals from these religious backgrounds also tended to marry earlier than other persons. Moderate Protestants and Catholics tended to marry at roughly similar ages, and thus occupied a middle ground in terms of age at (first) marriage. Liberal Protestants, Jews (except for Hasidim), and nonreligious persons married significantly later than other persons (Xu, Hudspeth, and Bartkowski, 2005; Lehrer, 2011).
Another important issue involves the choice of marital partners. Specifically, a long tradition of theory and research addresses causes, patterns, and trends in religious homogamy and intermarriage, with religion typically measured via denominational affiliation. Researchers have been interested in the probability of intermarriage and the distance of intermarriage if it occurs (Sherkat, 2004; Lehrer, 2011). Several key findings have emerged from this literature. First, in the United States, although religious homogamy remains an important phenomenon, rates of interfaith marriage have increased significantly among younger cohorts (Sherkat, 2004). This overall pattern may reflect declining theological differences and degrees of social distance among many religious groups, as well as reduced ability of many religious groups to sanction marital decisions. Educational upgrading, especially among women, geographical mobility, the growth of mass media and culture, and perhaps other factors are also thought to influence the general trend toward greater intermarriage. Second, although interfaith marriage is clearly on the rise, the trend toward greater exogamy in the United States has occurred more slowly among Catholics, conservative Protestants, and Mormons, as compared with more liberal denominations; however, the relative ordering of distances between Catholics and conservative Protestants stands out as a particularly formidable barrier to intermarriage (Sherkat, 2004). Third, the study of religious homogamy and interfaith marriage has long been complicated by the issue of religious switching; many partners who married outside the faith tended to convert to the religion of their spouse, resulting in higher apparent levels of religious homogamy. However, such tendencies toward conversion for marital reasons are waning among younger cohorts (Waite and Lewin, 2010).
In the United States and most Western European societies, Catholics have generally remained more endogamous than Protestants, and the more conservative Protestant groups (e.g., the conservative Reformed Protestants in the Netherlands) have maintained comparatively low rates of interfaith marriage (Kalmijn, 1998). However, the prevailing trends point to increasing rates of exogamy among most religious groups throughout Europe. Although many studies on this topic have omitted Jews due to their small numbers in most datasets, it has been widely assumed that rates of interfaith marriage have been quite low (Kalmijn, 1998). However, at least among non-Orthodox Jews in the United States, rates of exogamy have increased significantly, raising concerns about the intergenerational transmission of Judaism (Lazerwitz, 1995).
A number of distinguished scholars have explored religious differentials in fertility and fertility-related behaviors. According to one well-cited formulation, religion should be expected to make a difference when three important conditions are met: (i) religious groups and traditions hold distinctive theological beliefs regarding ideal family size, contraception, and other relevant factors; (ii) these religious groups have sufficient cultural and organizational strength to enforce these norms and to sanction violators; and (iii) religion is an important component of individuals’ identities, such that they would be deterred from violating these norms (McQuillan, 2004). To this list, one might also add the importance of parental religion in the intergenerational transmission of attitudes and preferences regarding ideal family size and other fertility-related matters (Pearce, 2002). For several decades, most research conducted in the United States focused on Catholic–Protestant fertility differences; Catholics tended to have significantly higher fertility rates, due to lower levels of contraceptive use as well as the higher percentage of European American immigrants in the Catholic population. However, these differences declined with rising rates of contraceptive use by Catholic women, and increases in assimilation and upward mobility by most European American ethnic groups (Westoff and Jones, 1979). Indeed, with the exception of specific groups such as Hispanic immigrants to the United States, evidence of Catholic distinctiveness in fertility rates has virtually disappeared in the United States and much of Europe (Frejka and Westoff, 2008).
By the early 1990s, investigators in the United States redirected their gaze toward the comparatively high fertility rates of three other groups (Mosher, Williams, and Johnson, 1992). First, although their Total Fertility Rates (TFRs) have declined over time and across cohorts, Mormons (Latter-Day Saints) still outpace nearly all other religio-ethnic groups in the United States. Second, although most Protestant denominations across the theological spectrum had reached broad agreement in support of family planning by the 1970s, some segments of conservative Protestantism began to reconsider this view 10–15 years later. Thus, as a group, conservative (i.e., fundamentalist, evangelical, and charismatic) Protestants now tend to have TFRs that surpass most other segments of the US population (Mosher, Williams, and Johnson, 1992). These patterns stand in marked contrast to the relatively low (below replacement value) TFRs of members of liberal Protestant groups and religiously unaffiliated persons. Third, although most segments of the Jewish population (Reform, secular, and even Conservative) have TFRs well below replacement level (i.e., below 2.0), Orthodox, and especially Hasidic, Jews have extremely high TFRs (over 3.0 for Orthodox and well over 6.0 for Hasidim). This high level of reproduction, accompanied by exceptionally low rates of intermarriage, has led some observers to conclude that they will dominate the US Jewish population in scarcely more than a generation (Stark, 2012).
In addition to these subgroup variations in fertility, several scholars have emphasized the role of religious commitment in general (e.g., frequency of attendance at services, religious salience) in predicting fertility intentions and behavior. According to one exhaustive set of analyses, both attendance and salience bear independent associations with the likelihood of having two or more children among women ages 18–44 in the United States (Frejka and Westoff, 2008). Another study in the United States revealed broadly consistent findings and also showed that a substantial portion of the link between religious salience (especially high salience) and fertility intentions was linked with a broader complex of traditional value orientations with respect to gender roles, marriage, and sexuality (Hayford and Morgan, 2008).
Religious attendance and salience are linked with fertility patterns in various European nations, albeit less strongly and less consistently than in the United States. After noting considerable regional variation in fertility patterns across Europe, Frejka and Westoff (2008) concluded that “if Europeans were as religious as Americans one might theoretically expect a small fertility increase for Europe as a whole, but considerably more for Western Europe.” Indeed, other scholars have gone much further, even suggesting that the low fertility rates among the least religious segments of the European population may imply that secularization (now widely heralded in many parts of Europe) may be a demographically self-limiting process, and that the differential fertility rates now observed between more and less (or non-) religious subgroups could lead to a de facto religious revival in Europe (Kaufmann, 2010).
Finally, one noteworthy recent contribution has explored religious differences in the timing of first births (Pearce, 2010). Briefly, and perhaps surprisingly, conservative Protestant and Catholic affiliations were positively associated with, while religious attendance was inversely related to, the likelihood of having a premaritally conceived first birth. The denominational patterns may reflect a combination of the focus on abstinence-based sex education and lack of contraception in some quarters and the greater cultural pressure to legitimize childbearing through marriage. Although there were no meaningful denominational differences, religious attendance was associated with faster rate of maritally conceived first births.
For several decades, scholars have investigated adult and parental child-rearing values, or the traits that are most preferred in children. Much of this attention has centered on two such values: obedience to authority and intellectual autonomy (or thinking for oneself). Studies have shown that the preference for autonomy in children increased steadily throughout much of the twentieth century, while the emphasis on obedience has declined (Alwin and Felson, 2010).
Early studies focused on Catholic–Protestant differences in child-rearing orientations, showing that Catholics expressed greater enthusiasm for obedience and less support for intellectual autonomy. However, by the 1980s, such differences had all but disappeared (Alwin and Felson, 2010), and the gaze of researchers has shifted to conservative Protestants, whose preferences for obedience in children have now been well established in the literature (Ellison and Sherkat, 1993a; Starks and Robinson, 2005). One trend analysis using NORC General Social Survey Data from 1986 to 2002 revealed no change in the valuation of obedience among Catholics, but an increase in support for obedience among frequently attending conservative Protestants over the 16-year study period (Starks and Robinson, 2005). One notable finding in some of these studies is that while early research on religion and child-rearing values presumed that obedience and autonomy were antithetical preferences, contemporary research suggests that conservative Protestants tend to value both of these traits highly, perhaps because they want children to obey parental and religious authority figures but also want them to exercise independent judgment in sifting through secular information and cultural messages that may be incompatible with their religious worldviews (Ellison and Sherkat, 1993a).
Although most work on religion and child-rearing values has been conducted in the United States, a handful of studies have explored these issues in other contexts. For example, research among Turkish Muslims has linked higher levels of religiousness (orthodox belief, frequency of prayer, religious salience) with greater preference for the traits of obedience and good manners in children while also showing that more religious adults are less supportive of intellectual autonomy and creativity (Acevedo, Ellison, and Yilmaz, 2013). A comparative study of child-rearing values in 11 European nations reported that adults in societies with higher proportions of Protestants were more prone to value imagination and creativity in children than those persons in predominantly Catholic countries. However, religious cultural tradition was unrelated to the valuation of obedience and religious faith in children (Fjellvag, 2011).
Beginning in the 1990s, an expanding body of research in the United States has focused on the distinctive parenting philosophies of conservative Protestants. One study of advice manuals produced for parents revealed sharp differences between materials generated within these religious circles and those written by secular child-rearing experts. According to this work, conservative Protestants are distinctive with regard to long-term parenting objectives (authority-mindedness in preparation for spiritual salvation, as opposed to emotional health and social skills), structure of parent–child relations (hierarchal rather than egalitarian), definition of parental roles (preference for clearly defined and gendered parental roles), and firm child discipline, including corporal punishment (Alwin and Felson, 2010).
Building on this research, subsequent scholarly inquiry has focused on religious variations in approaches to child discipline, including corporal punishment of children. Those who initiated research in this area found that members of conservative Protestant groups were more prone to support, and to use, mild-to-moderate forms of corporal punishment (i.e., spanking and slapping) to discipline children than persons from other (or no) religious backgrounds (Ellison and Sherkat, 1993b; Ellison, Bartkowski, and Segal, 1996). These patterns have been traced to several specific theological tenets that are disproportionately embraced by conservative Protestants, including the belief that the Bible is without error and is the authoritative guide for human behavior, the belief in original sin, the belief that God is judgmental as well as loving, and the belief in a literal Hell. To achieve spiritual salvation, conservative Protestant theology holds that individuals must repent of their sinful ways, submit to divine authority, and accept salvation through God’s grace (Hempel and Bartkowski, 2008). These core conservative Protestant beliefs help to explain the high premium placed on obedience, as parents believe they have a duty to shape the will of children, making them more submissive and setting the stage for their salvation (Ellison and Sherkat, 1993b). Moreover, corporal punishment (the rod) is prescribed as the appropriate punishment for willful disobedience and rebellion against parental authority (Ellison, Bartkowski, and Segal, 1996). Research has shown that conservative Protestant mothers tend to employ corporal punishment as part of a coherent parenting strategy, rather than a spontaneous expression of anger or frustration, and that these mothers are unconvinced by the arguments of secular experts about the harm that this practice might do (Gershoff, Miller, and Holden, 1999).
In recent years, investigators have also explored religious differentials in other aspects of parental – and especially paternal – involvement in the lives of children. Here, as in many other domains of family life, some scholars have been particularly interested in the potential distinctiveness of conservative Protestantism. Some observers in this area have conceptualized paternal involvement as a three-dimensional construct, comprising paternal supervision, affective fathering, and father–child interaction. Contrary to some popular assumptions, this work found that conservative Protestant fathers tended to engage in greater supervision and affective fathering than other fathers; the contrast with their nonreligious counterparts was especially marked (Bartkowski and Xu, 2000). Other studies have explored the association between the overall religiousness of parents (fathers and mothers) and their involvement in the lives of their children. For example, one study using data from a nationwide probability sample of parents linked the frequency of attendance with paternal supervision, affective parenting, father–child interaction, and several specific paternal behaviors with school-age children, including one-on-one activities, having dinner together, and youth-related activities such as school or community activities, and scouting (Bartkowski and Xu, 2000; Wilcox, 2002). An important alternative hypothesis – that these associations stemmed from a broader set of values, mores, and practices, termed middle-class conventionalism – received no empirical support (Wilcox, 2002).
Although most of this research has been conducted using probability samples drawn from the general population, a parallel set of studies have explored similar issues among fragile families, that is, samples of urban, lower SES, disproportionately minority, and mostly unmarried parents of young children. Several findings are noteworthy. First, paternal religious attendance was positively associated with a summary measure of paternal engagement with children ages 1–5, and this pattern was especially strong for fathers who increased their attendance immediately following the birth of a new child (Petts, 2007). Second, in contrast to findings in the general population, conservative Protestant affiliation was inversely associated with an omnibus measure of paternal engagement (Wildeman, 2008). After examining associations between parental (and especially maternal) religious attendance and paternal and child well-being in fragile families, one researcher concluded that religious practices and communities offer valuable resources for single mothers, in particular. Such resources could conceivably encompass formal programs and ministries, informal support networks, and norms of successful, spiritually grounded child-rearing (Petts, 2012).
In addition to parental values and child-rearing practices, researchers have increasingly focused on religious differences in child outcomes. One area of interest has involved corporal punishment, which is widely presumed to foster externalizing behavior problems (aggression, antisocial conduct) and internalizing problems (emotional distress), as well as other developmental issues for children. An emerging body of work has suggested that mild-to-moderate forms of corporal punishment (i.e., spanking and slapping) may bear fewer (or no) negative associations on children from conservative Protestant families, as compared with those from other backgrounds (Gunnoe, Hetherington, and Reiss, 2006; Ellison, Musick, and Holden, 2011). Indeed, in one longitudinal study, children of conservative Protestant mothers who were spanked at the initial data collection point (when the children were ages 2–4) but were no longer spanked 5 years later exhibited particularly low levels of antisocial behavior and emotional problems compared with other children. Several factors could mitigate any deleterious effects of corporal punishment for this specific subgroup. These may include (i) distinctive modes of administering corporal punishment, about which much advice has been offered by conservative Protestant leaders and writers, and (ii) a cultural climate within this religious community – clergy, peers, relatives, and others – that may normalize and support corporal punishment as an appropriate disciplinary response to certain types of childish misbehavior (Ellison, Musick, and Holden, 2011). In addition, any harmful effects of this type of discipline may be mitigated by the lower levels of yelling and verbal abuse (Bartkowski and Wilcox, 2000) and the higher levels of parental warmth and nurturance (Wilcox, 1998), which are exhibited by conservative Protestant parents. Clearly, this is a topic that begs for additional investigation in the future.
In addition to the research summarized earlier, many studies have explored relationships between facets of parental, child, and/or adolescent religiousness and a range of other child and adolescent outcomes, often reporting salutary patterns in the samples of the general population as well as major subgroups (Regnerus, 2003). For example, studies have demonstrated that children (school-age children and adolescents) from more religiously active backgrounds have often exhibited fewer internalizing and externalizing behaviors. In several studies, they achieved better academic outcomes and interpersonal relationships and tended to follow instructions and avoid disciplinary problems more than those from less religious backgrounds (Muller and Ellison, 2001; Bartkowski, Xu, and Levin, 2008). Many other studies – too numerous to review here – have linked aspects of religiousness (on the part of parents and children or adolescents) with a host of outcomes, ranging from reduced delinquency to increased levels of volunteering and civic involvement (Regnerus, 2003). There is evidence that these links between parental religiousness and child outcomes may be due to higher levels of parental supervision and expectations, as well as greater intergenerational closure within religious families and networks (Smith, 2003).
One important area of research involves the role of religious similarity or dissimilarity (in terms of denominational affiliation, attendance patterns, or theological beliefs) among family members in shaping child outcomes. For example, couples with highly dissonant affiliation patterns have reported increased marital conflict and reduced religious participation compared to those couples with similar religious affiliations; according to one study, the adolescent children of these unions were more likely than others to engage in marijuana use and underage drinking (Pearce and Haynie, 2004). Other research has revealed that the more religious mothers and adolescents were, the less likely adolescents were to engage in acts of delinquency; however, significant religious discord between mothers and adolescents – especially when mothers valued religion much more than their children did – was associated with elevated levels of delinquent behavior (Stokes and Regnerus, 2009). Nor are these patterns restricted to the quality of relationships between parents and adolescents. Similar findings have surfaced in research on the associations between the religious congruence between mothers and their adult children and positive perceptions of their relationships by both parties (Pearce and Axinn, 1998). Finally, one recent study of mothers’ perceptions in late life revealed that their perceptions of religious value similarity to their adult children predicted assessments of greater affection and lower levels of conflict. These patterns were especially pronounced in African American families (Sechrist et al., 2011).
Studies of general parental religiousness highlight its possible role in shaping relationship quality between parents and older adult children. Some of these works have focused on paternal religiousness and relationship quality among fathers and older adult children. In the work by King (2010), father’s religious salience predicted positive assessments of father–child relationship quality (measured in terms of understanding, trust, fairness, and affection) as reported by older adult offspring. It is striking that these patterns persisted despite statistical adjustments for parental involvement during the teen years, offspring reports of antisocial behavior during adolescence, and multiple indicators of marital quality. Maternal religiousness (attendance, salience) has also been linked with positive relationships with adult children (Pearce and Axinn, 1998). Further, more religious grandparents tend to be more active in the lives of their grandchildren than their less religious counterparts, partly because they are more enmeshed in family relationships of many types (King and Elder, 1999). Such findings clearly underscore the need to understand the role of religion in shaping family relationships across generations.
A long tradition of theory and research has linked religion and spirituality with marital quality. However, over the past 10–15 years, there have been several innovations in this literature, such as the emergence of new measures in the domain of religion and spirituality, the development of fresh ways of conceptualizing and modeling the links between religion and marital quality, and attention to an expanded array of marital outcomes, including satisfaction and happiness, dependence, conflict and conflict resolution, domestic violence, and risk of marital dissolution.
A wealth of studies has centered on the association between religious factors – primarily affiliation, attendance, and salience or commitment – and positive indicators of marital quality, such as marital happiness and satisfaction. With only a few notable exceptions, the vast majority of work in this area has reported salutary associations between religious involvement and positive indicators of marital quality. Although many studies in this area have used generic measures of religiosity such as religious attendance and salience, as reported by one partner, more recent works have focused on dyadic information, with particular attention to the couple’s joint religious activities, such as devotional activities (e.g., prayer, scriptural study) in the home (Lichter and Carmalt, 2009; Ellison, Burdette, and Wilcox, 2010).
Why might we expect a connection between the institutions of religion and marriage? One conceptual scheme points to three elements: norms, networks, and nomos (Wilcox and Wolfinger, 2008). Briefly, most religious groups and traditions promote norms that can facilitate happy and stable relationships. These may be general norms, involving kindness, generosity, forgiveness, and Golden Rule ethics. However, religious norms regarding marriage may be more specific. Religious groups often embrace certain expectations about appropriate marital roles (i.e., what it means to be a good husband, a good wife), as well as particular types of behaviors to be avoided. Both types of norms are typically grounded in scripture and theology, and they may be communicated through written materials and sermons. In addition to these two types of norms, religious communities consist of networks that can also foster positive marital outcomes. This may occur through informal mechanisms, such as relationships among individuals and couples within the congregation, or through formal mechanisms such as premarital counseling, couples’ ministries and marital enrichment programs, and religious education classes. Finally, the term nomos refers to a religious interpretive framework, through which the marital bond takes on added spiritual significance. Such a religious perspective, a version of which is discussed more fully later, may foster a shared understanding between partners that the union is more covenantal than contractual in nature.
Family researchers influenced by developments in psychology have increasingly conceptualized the links between religion and marital quality in terms of a relational spirituality approach, which often distinguishes between distal and proximal religious constructs. Examples of the former include religious practice and commitment, while examples of the latter include interconnected relational virtues such as forgiveness and sanctification, among others (Mahoney, 2010). The latter has become a particularly important construct in this literature. Briefly, sanctification refers to the extent to which individuals ascribe sacred characteristics to the relationship or to one’s partner, or view the relationship having sacred meaning or significance. Researchers have linked sanctification with a range of desirable outcomes, including greater marital satisfaction, happiness, and bonding, as well as lower levels of negative interaction and conflict, and improved communication and conflict resolution skills (Mahoney et al., 1999; Mahoney, 2010). In addition, controls for sanctification often eliminate the associations between religious practice or commitment and these marital outcomes, sometimes even reversing the signs of the coefficients. Sanctification has also been shown to moderate the deleterious effects of general relationship stress, financial strain, and marital inequity on relationship quality (Ellison et al., 2011).
Several additional factors have been advanced to account for the links between religiousness and intimate relationship quality. Among the most prominent explanations are forgiveness (Fincham, Hall, and Beach, 2006), joint communication, especially around religious or spiritual matters (Mahoney et al., 1999), self-sacrifice, and the expression of love and affection in the form of compliments, emotional supportiveness, and gratitude (Wilcox and Wolfinger, 2008; Wolfinger and Wilcox, 2008). To date, however, the precise nature of these associations remains unclear (Stafford, David, and McPherson, 2013), and additional research is needed to establish their role in marital and relationship quality.
Another important topic involves the association between religious homogamy, or religious similarity of partners, and marital quality. Although investigators focused on denominational homogamy, the importance of denomination as a marker of same-faith (or mixed-faith) unions appears to have declined over the past several decades (Myers, 2006). Thus, in more recent studies, investigators have tended to focus on other facets of religious similarity among partners. Some researchers have emphasized the importance of shared religious commitment, often measured in terms of similar or joint attendance at worship services (Wilcox and Wolfinger, 2008). In a recent study that considered multiple aspects of couples’ religion, shared religious and spiritual values – along with joint family religious activities – strongly predicted relationship quality among working-age couples (Ellison, Burdette, and Wilcox, 2010).
In addition to marital happiness and satisfaction, social scientists have also examined several additional facets of marital quality, and we turn to those briefly. One such dimension is marital dependency, or the extent to which either spouse believes that the quality of his or her life (in terms of money, companionship, sex, or various other factors) would decline should the marriage end. This gauges the extent to which marital partners depend upon the union for various types of resources. One study of a large sample of US adults has explored the associations between spouses’ religiosity and dependency using dyadic data. This work found that conservative Protestants (especially non-Baptists), members of quasiethnic religions (e.g., Lutherans, Mormons), regular churchgoers, and persons in same-faith unions expressed greater marital dependency than others, although those patterns varied somewhat by gender (Wilson and Musick, 1996).
Studies have also addressed religious variations in marital conflict. For example, various forms of religious similarity and dissimilarity have been shown to be linked with the levels and types of arguments in a large sample of US married couples (Curtis and Ellison, 2002). The estimated effects of denominational heterogamy on marital conflict were modest and inconsistent; however, spousal differences in attendance patterns, and especially differences in core theological beliefs (e.g., the belief that the Bible is the inerrant Word of God), were linked with more frequent arguments. In addition to the frequency of such conflicts, research has also provided information on religious differences in the risk of specific types of conflicts. For example, sexual infidelity is a key source of marital strife, and studies have shown that regular churchgoers, persons who believe that the Bible is error-free, and conservative Protestants with high religious commitment and inerrantist beliefs are especially unlikely to engage in extramarital sex (Burdette et al., 2007). Theological dissimilarity among partners has been associated with more frequent conflicts in two specific areas: finances and household labor. One possible explanation for housework conflicts between very conservative and very liberal partners involves the tendency of the former to embrace separate spheres ideology, according to which household labor is divided fairly rigidly along gender lines, with women primarily responsible for daily domestic chores (e.g., cooking, cleaning) and men responsible for sporadic tasks (e.g., repairs). Among couples in which both partners share conservative beliefs about the inerrancy and authority of the Bible, women perform more housework than their counterparts from other backgrounds, and in those households, domestic tasks are also more gender segregated (Ellison and Bartkowski, 2002).
Turning to the issue of more severe forms of conflict, a modest body of research has also examined the connections between religion and intimate partner violence. Briefly, nationwide survey data indicate that frequency of religious attendance is strongly inversely associated with the perpetration of acts of violence by men and women. This pattern among men was stronger for African Americans than men from other ethnic backgrounds, due largely to elevated rates of perpetration among nonattending African American men. In addition, dissimilarities of theological conservatism were associated with violence; specifically, the risk of violence was elevated among couples in which the man held much more conservative beliefs about the Bible than the woman (Ellison, Bartkowski, and Anderson, 1999). A parallel literature has explored partner violence within religious communities in more detail (Nason-Clark, 1997). Although the findings defy easy summary, this line of work has suggested that many congregations and (male) clergy members may be poorly informed and ill-suited to address women’s victimization by intimate partners. For example, some clergy dismiss the possibility of violence within the church community, while others can give inappropriate advice (e.g., counseling abused women to remain in violent relationships for religious or other reasons). On the other hand, informal networks (primarily fellow churchwomen) can be valuable sources of solace and assistance (Nason-Clark, 1997).
Finally, a number of studies have explored the role of religious factors in shaping the risk of divorce or marital dissolution. Clearly, religious groups have differed sharply on their views regarding marriage and divorce. For example, in recent nationwide studies, conservative Protestants have expressed particular discomfort with the phenomenon of divorce, and greater enthusiasm for restricting divorce, than other Americans (Stokes and Ellison, 2010; Ellison, Wolfinger, and Ramos-Wada, 2013). Indeed, religious conservatives were the key force behind the phenomenon of covenant marriage, a specific, legally recognized form of marriage in some parts of the United States, in which couples agree to premarital counseling and a more restricted set of conditions for divorce (Nock, Sanchez, and Wright, 2008). Although the goals of this movement included reducing divorce rates and strengthening individual marriages (and strengthening marriage as a social institution), recent research has cast doubt on the effectiveness of covenant marriage in at least one state (DeMaris, Sanchez, and Krivickas, 2012). Despite the deep concerns in religious circles about the current state of the institution of marriage, it is interesting to note that conservative Protestant congregations have been no more likely than others to offer marital counseling or other marital support programs (Wilcox, Chaves, and Franz, 2004).
Early research centered on whether mixed-faith unions, that is, those in which partners belonged to disparate denominations or faith traditions, faced heightened risk of divorce. This was a reasonable line of inquiry to the extent that denominational differences were also indicative of broader differences in lifestyles, cultural preferences, and social networks. However, most recent studies have reported only modest links between denominational homogamy and marital stability, with a few caveats (Lehrer, 2011). In particular, certain patterns of mixed-faith marriage, especially those in which one partner belonged to an exclusivist (or conservative) denomination and the other belonged to an ecumenical (or more liberal) denomination, were at particularly high risk of dissolution. In addition, specific types of homogamy (e.g., among Mormons) appeared to be more insulated from dissolution than other unions. Finally, although there is no evidence of a protective effect for conservative Protestants overall, homogamous conservative Protestant couples in which both partners attended services regularly had lower rates of marital dissolution than other types of unions (Vaaler, Ellison, and Powers, 2009).
Many investigators have turned their focus to other facets of religion that may influence marital stability. Some observers argued that religious participation, especially attendance at services, was more closely linked with marital stability, due to shared networks and activities, as well as common religious socialization among partners. However, religious homogamy remains a topic of interest; several researchers have concluded that spousal differences in attendance patterns and core theological beliefs are linked with heightened risk of marital dissolution (Call and Heaton, 1997; Vaaler, Ellison, and Powers, 2009). These findings may track broader changes in American religion (e.g., restructuring along conservative–liberal or religious–secular lines), as religious beliefs such as biblical inerrancy increasingly cross denominational boundaries. One study has reported gendered patterns, indicating that unions in which the men attend much more often than the women, and those in which the women are much more conservative than the men, may face elevated risk of dissolution than others (Vaaler, Ellison, and Powers, 2009), although this finding needs additional replication and investigation.
In this chapter, we have reviewed a large and growing body of empirical research on the associations between multiple aspects of religion and a wide array of family-related attitudes and behaviors. As we noted at the outset, space limitations required that we focus on a limited set of outcomes, and the structure of scholarly literature dictated that much of our review centered on (Judeo-)Christian religion and on studies conducted primarily in the United States. We conclude the chapter with several general observations about the limitations of the literature in this area, as well as broad suggestions for future work.
First, we have dealt solely with studies that have treated religion as an independent variable, demonstrating or suggesting effects on family-related outcomes. However, it is also important to remember that the religion–family connection is bidirectional, that is, that family structures and relations can shape religious affiliations, practices, and beliefs. This is particularly evident in the research on religious socialization among adolescents and young adults. For example, adults may alter their religious participation when they marry or have children, at least partly with the needs of partners and offspring in mind. In addition, parental marital discord and divorce can alter or impair the intergenerational transmission of faith. There are many additional examples of this other religion–family connection.
Second, the religion–family literature has focused almost exclusively on heterosexual families. At one level, this may not be surprising, given the widespread derogation and exclusion of GLBT persons and same-sex couples within many religious communities. However, new religious denominations (e.g., the Metropolitan Community Church) with a focus on the GLBT population are flourishing, existing denominations (e.g., the United Methodist Church) are reevaluating their positions with respect to homosexuality and same-sex unions, and new expressions of GLBT spirituality are emerging beyond the bounds of organized religion. Limited evidence has linked religiosity with positive relationship functioning among same-sex couples as well (Oswald et al., 2008). The role of religion and spirituality among GLBT couples and families clearly warrants further research.
Third, expanding on the previous point, nontraditional families (not just GLBT but also divorced, cohabiting, and single-parent families) are becoming increasingly common in the United States and other Western societies. Most religious institutions have been heavily invested in the traditional model of the two-parent heterosexual family for both theological and institutional reasons. Others have not always been well served by existing religious institutions, in terms of formal programming or informal social climates. In a very real sense, however, the long-term viability of organized religion in the United States may well depend upon how quickly and successfully existing religious structures adapt to changing family definitions and realities (Edgell, 2006).
Although we have focused heavily on general population samples, we note an important paradox: some of the most highly religious segments of the US population, notably African Americans, also face some of the greatest obstacles to marriage and family stability. A few studies have used data on a large sample of fragile families, which includes a disproportionate share of ethnic minority parents (e.g., Wolfinger and Wilcox, 2008), and a small number of studies have compared the associations between religion and family among multiple ethnic groups (Brown, Orbuch, and Bauermeister, 2008; Ellison, Burdette, and Wilcox, 2010). But there are few studies on African Americans specifically, and there remains a dearth of information on the role of religion in the family lives of Latinos, Asian Americans, and other minority populations. Remedying this defect should be an urgent priority for future investigators.
Finally, with the exception of fertility-related attitudes and behaviors, to our reading, there is a significant shortage of systematic empirical research on the links between religion and family life outside the United States and especially in non-Western contexts. (Our search was necessarily limited to English language materials.) Thus, additional comparative research on varied religions and family life should be strongly encouraged. It is hoped that future inquiry along the lines sketched earlier will further illuminate the continuing, and changing, associations between the social institutions of religion and family.