Living and Loving “Decent”: Religion and Relationship Quality among Urban Parents

Abstract

Religious participation is linked to overall satisfaction among both married and unmarried couples in urban America. Less is known about what may account for the association between religious participation and relationship quality. We explore this issue using data from the first two waves of the Fragile Families and Child Wellbeing Study. Relationship-related behaviors (e.g., temperance) and relationship-specific behaviors (e.g., affection) can each account for the association between church attendance and relationship quality. Furthermore, religious participation appears to be more tightly linked to men's perceptions of relationship quality than women's.

In the United States, religion and family have a longstanding association (Bartkowski, 2001; Christiano, 2000; Edgell, 2005; Thornton, 1985). Families seek out religious institutions to consecrate key life transitions—notably young adulthood, marriage, childbirth, and death—and depend on these institutions to help guide, support, and sacralize marriage and parenting. This dependency has been strongest for married couples with children, largely because churches offer religious and normative support to marriage, and to childbearing and child-rearing within marriage (Edgell, 2005; Stolzenberg, Blair-Loy, and Waite, 1995; Wilcox, 2006).

This model of family life has become less common over the last 40 years as the U.S. has witnessed dramatic changes in family structure, notably increases in divorce, cohabitation, and nonmarital childbearing. In 2002, approximately one-third of the children born in American cities were born outside of wedlock (National Center for Health Statistics, 2002). Of these children, 82% are born into fragile families: 51% to cohabiting unions and 31% to visiting unions (McLanahan, Garfinkel, and Mincy, 2001). Although a large percentage of urban children are born outside of marriage, many are being born into fragile families where the parents are in an ongoing romantic relationship (McLanahan 2005).

Given that religious institutions have traditionally celebrated marriage and stigmatized nonmarital sexual unions, one might expect that the generally positive association between religion and relationship quality among married couples (Myers, 2006; Wilcox and Nock, 2006) would not hold for fragile families. Surprisingly, recent research suggests that both married and unmarried couples with children benefit from the religious attendance of men; specifically, couples where the father attends church several times a month are happier with their relationships and are more likely to report that their partner is emotionally supportive (Wolfinger and Wilcox, 2008). In addition, irrespective of marital status new fathers who attend church frequently are more likely to be involved with their children (Petts 2007); greater paternal involvement may in turn benefit union quality.

To date, no research has explored the mechanisms that may help account for the association between religion and relationship quality for couples, irrespective of marital status, in urban America. Family-specific norms associated with religious institutions (e.g., children ought to be reared by married parents), relationship-specific behaviors encouraged by religious institutions (e.g., affection and the absence of domestic violence), and other behaviors that affect relationships (e.g., illicit drug use) may account for the relationship between religion and relationship quality. Using data from the first two waves of the Fragile Families and Child Wellbeing Study, a national longitudinal study of urban parents who had a child between 1998 and 2000, this paper aims to answer two questions: (a) Are the associations between religious participation and the quality of intimate relationships in urban America mediated by family norms, relationship-specific behavior, and relationship-related behavior?; and (b) Do the effects of religious participation and these other intervening variables vary by gender?

The answers to these questions are important for two reasons. First, religious institutions are important civic actors in urban America, yet we know little about how they are doing in reaching out to the growing number of fragile families in U.S. cities, families which have traditionally been shamed or shunned by organized religion (Ellingson, 2004). Second, because parental relationship quality is of signal importance to the well-being of children (Amato and Booth, 1997; Carlson and McLanahan, 2006), the effects of religious participation, family norms, and relationship-related behavior on the quality of parental relationships may have important implications for children growing up in urban families.

Norms, Networks, and Nomos: Understanding the Link between Religion and Relationship Quality

In recent years, scholars have examined how religion generates social capital—that is, the resources associated with the structure and quality of relations between persons (Coleman, 1990)—in the civic and political spheres of American life (Bartkowski and Regis, 2003; Putnam, 2000; Verba, Schlozman, and Brady, 1995). But scholarship on religion has not considered how religion may generate what might be called domestic social capital. More specifically, religion may directly and indirectly foster the formation of domestic social capital: in this case, relationship quality (Coleman, 1990; Pearce and Axinn, 1998). Religious institutions promote behaviors (e.g., public expressions of forgiveness) and schemas (e.g., a willingness to put aside resentments) that have a direct influence on the quality of relationships. They also affect relationships indirectly by fostering behaviors like temperance and sexual fidelity that influence relationship quality by, among other things, promoting trust between partners. By embedding couples in social networks where they can receive social and emotional support, where relationship-specific and -related norms are supported, where departures from those norms are sanctioned, and where pro-family and pro-social beliefs are produced and legitimated through a religious nomos, religious institutions may foster trust, reciprocity between partners, and a virtuous way of life that is conducive to quality relationships (Coleman, 1990: 318−320; Putnam, 2000: 134−136). We turn now to a brief consideration of the roles that religious norms, networks, and nomos may play in influencing relationship quality, and to a consideration of how they may be particularly important for men.

Religious institutions promote specific norms relevant to relationships in urban America. Specifically, religious institutions tend to foster pro-marriage attitudes, gender traditionalism, and perhaps trust between the sexes (at least amongst fellow believers) (Wilcox, 2004). Insofar as promarriage attitudes generate trust between partners, they should encourage both partners to invest more in that relationship (Edin and Kefalas, 2005; Wilcox and Nock, 2006). Relationship-specific norms should also foster high-quality relationships. On the other hand, gender traditionalism may have varying effects for men and women, perhaps encouraging more emotional engagement for women and less for men (Edin and Kefalas, 2005; Wilcox and Nock, 2006).

In urban America, churches and other religious institutions endorse a range of relationship-related norms that coincide with what Elijah Anderson (1999) has called a code of decency. The ethos of decency in urban America encompasses norms like temperance, hard work, honesty, respect, and lawful behavior; it stands in opposition to the code of the street that encompasses substance abuse, antisocial behavior, promiscuity, and crime (Anderson, 1999). Women and especially men who internalize the ethos of decency, and are embedded in social networks that support this ethos, are more likely to avoid street behaviors that can harm their relationship, like imprisonment, drug addiction, and infidelity (Edin and Kefalas, 2005). Our ongoing ethnographic research suggests that religious institutions are key promulgators of these relationship-related norms (see also Anderson, 1999; Furstenberg, 2001). Moreover, a growing body of research on urban religion indicates that religious attendance is associated with “decent” behavior such as lawful behavior and employment, as well as higher levels of education (Freeman, 1985; Johnson et al. 2000).

The social networks found in religious congregations also play an important role in fostering behaviors that influence the quality of urban relationships. Religious networks provide couples with models of good marriages, particularly in communities where such marriages are rare; they also provide couples with emotional support and practical guidance for their relationship (Furstenberg, 2001; Wilcox and Wolfinger, 2008). More generally, these networks provide social and emotional support that can help buffer adults against the stresses—poverty, discrimination, underemployment—that can otherwise harm urban relationships or lead to behaviors like drug addiction that in turn harm relationships (Conger et al., 1990; Ellison, 1994; Edin and Kefalas, 2005). Religious networks also reinforce pro-family norms by according status to adults who conform to those norms and stigmatizing adults who reject them (Coleman, 1990).

Finally, religious institutions offer a nomos—a sacred canopy of beliefs and rituals—that can influence relationships and the behaviors that affect relationships in urban America (Berger, 1967). By providing rituals and beliefs that endow marriage with transcendent significance, religious institutions can sanctify relationships in ways that encourage partners to stay committed and to view their relationship in a more favorable light (Mahoney et al., 2003). Religious rituals such as prayer may also encourage urban parents to be more reflexive about their relationships, and seek to pursue positive behaviors (e.g., affection for one's partner) and avoid negative behaviors (e.g., ignoring one's partner's needs). In addition, religious beliefs and rituals may afford parents with a sense of purpose that helps them negotiate the challenges of urban life and buffers against the negative psychological consequences of stress (Ellison, 1994). This, in turn, may make urban parents more likely to resist street behaviors—drug use, criminal activity, and promiscuity—that can harm their relationships (Anderson, 1999; Edin and Kefalas, 2005).

Given that the association between religion and marital quality depends upon religious homogamy (Ellison et al., 1999; Lehrer and Chiswick, 1993), we expect that the association between religion and relationship quality among urban couples will be strongest for couples who share regular attendance and religious affiliations. Religious homogamy—especially in the form of joint church attendance—can reinforce and sustain a sense of marital trust, a shared commitment to relationship-related norms, and the support and control afforded by religious social networks. In theoretical terms, when religious social capital is strong for a couple, so also should their domestic capital be strong. Couples who share religious capital should reap the benefits of shared social networks, networks that encourage couples to abide by relationship-specific and relationship-related norms that are conducive to high quality relationships (Wilcox, 2006). Moreover, Stark (1996: 164) notes that religion, understood as a “group property” rather than individual attribute, is most likely to effect social behavior when it is held by a group rather than an individual. Indeed, the literature on religion and marriage indicates that religious homogamy is associated with higher reports of marital quality and stability, as well as lower reports of domestic violence, whereas the effects of individual religiosity on marital quality may be ambiguous or negative when the couples do not share the same level or type of religious faith (Ellison et al., 1999; Wilcox and Nock, 2006; but see Wolfinger and Wilcox, 2008).

The association between religion and urban relationships varies by gender, such that men's religiosity seems to be more predictive of positive relationship outcomes than women's religiosity (Wolfinger and Wilcox, 2008). There are two reasons for this. First, in society at large, religion is one of the few institutions that orients men toward their families; institutions like work, leisure, sports, and popular culture often do not encourage men to focus on their families (Wilcox and Bartkowski, 2005). Indeed, religious institutions appear to cultivate a distinctive form of masculinity, achieved through successful performance of spousal and paternal roles, rather than through professional, athletic, or leisure roles (Bartkowski 2004; Wilcox and Bartkowski 2005). On the other hand, women are encouraged by a range of institutions to focus on their relationships, and femininity is typically associated with a strong relationship orientation (Strazdins and Broom, 2004; Thompson and Walker, 1989). Second, men in urban America are less marriageable than women in part because they are more likely than women to engage in street behaviors that threaten the integrity of their relationships (Edin and Kefalas, 2005; Wilson, 1996). Insofar as religious institutions encourage “decent” norms and behaviors among their members, they could be particularly valuable to urban men who might otherwise behave in a destructive fashion.

So we predict that religious attendance, especially joint church attendance, will be positively associated with urban fathers’ and mothers’ assessments of their relationship quality. We anticipate that this association will be mediated by relationship-specific norms, relationship-specific behaviors, and relationship-related behaviors. Finally, we predict that men's religious attendance is more consequential for the quality of urban relationships than women's attendance; the emphasis churches place on family orientation and decent behavior may be particularly valuable for men. Thus both men's solo church attendance and joint church attendance by both parents should benefit interpersonal relationships among new parents in urban America.

METHOD

Data

We analyze data from the Fragile Families and Child Wellbeing Study, a national survey designed to study new unwed parents, their relationships, and their children in large U.S. cities (Reichman et al., 2001). The survey follows approximately 3,700 children born to unmarried parents and 1,200 children born to married parents in 20 cities with populations over 200,000. The baseline interviews were conducted between February 1998 and September 2000. New mothers were interviewed in the hospital within 48 hours of giving birth, and new fathers were interviewed at the hospital or elsewhere around the same time. A follow-up interview queried parents and children approximately 12 months after birth; additional interviews were conducted after 36 and 60 months.

Our analysis includes male and female couples who participated in both baseline and follow-up interviews, and report a continuing romantic relationship at both times; we do not analyze subsequent waves. Based on these criteria we exclude 1,068 fathers (22% of the total) who did not participate in the baseline interviews, 533 mothers (11% of the total) who did not participate in the 12-month follow-ups (11% of the total), 1,531 fathers (31% of the total) who did not participate in the 12-month follow-ups, 1,435 couples (29% of the total) who are not in a relationship at both waves. Eighteen women and 23 men lack data on relationship quality; these data are available for 2,723 couples. Of our independent variables, one is missing data on 6% of cases; the others, 4% or less. Given the paucity of missing data we delete affected cases listwise (see Allison, 2001 for the advantages of this technique). This yields a sample size of 2,034 male-female couples and therefore 4,068 data points. Analyses are unweighted. The weights only make respondents unmarried at the baseline interviews nationally representative, so they are problematic when studying both married and unmarried couples. They also exclude more than one quarter of cases.

In supplemental analyses we consider the relationship between religious involvement and attrition, either by survey nonresponse or union dissolution. These analyses show no substantial effects related to religious involvement. Similarly, church attendance does not affect union dissolution for either married or unmarried Fragile Families couples. These results suggest that sample selection based on religious involvement probably has no effect on our findings. In addition, our earlier work with Fragile Families data found no evidence of endogeneity in the relationship between religious participation and union quality (Wolfinger and Wilcox, 2008). Finally, we explored the possibility of sample selection based on generativity: perhaps Fragile men having their first child—or even their first child with their Fragile Families partners—may be more committed to their families and hence perceived as better partners (cf. Petts, 2007). These controls for prior fertility did not affect our results and are accordingly omitted from the findings presented here.

We measure our independent variables at the baseline and the dependent variables at the one-year follow-up as a safeguard against spurious associations. Since each set of variables is measured at only one point in time there is no need to employ the type of multilevel analysis often associated with panel data (e.g., Hox, 2002).

Dependent Variable

Our dependent variable is overall relationship quality, measured at the one-year follow-up with a question reading, “In general, would you say that your relationship with [your partner] is excellent, very good, good, fair, or poor?” Answers are coded on a five point scale, ranging from 5-“poor” to 1-“excellent”; we reverse the coding so positive scores on independent variables indicate happier relationships in regression models. We analyze responses from both male and female members of each couple. Summary statistics for all variables appear in Table 1.

Table 1.

Summary Statistics.

Variables	M	SD
Father's union satisfaction	2.12	1.11
Mother's union satisfaction	2.25	1.20
Father alone attends church frequently	.09	.29
Mother alone attends church frequently	.17	.38
Both partners attend church frequently	.21	.41
Married	.34	.47
Mother's religious tradition
None	10%	---
Black Protestant	33	---
Catholic	31	---
Conservative Protestant	17	---
Other	9	---
Father's religious tradition
None	12%	---
Black Protestant	33	---
Catholic	28	---
Conservative Protestant	17	---
Other	10	---
Mother and father same faith	.66	.48
Mother's age	25.53	6.12
Father's age	27.83	7.10
Mother's education
Less than high school	28%	---
High school graduate	35	---
Some college	24	---
College graduate	13	---
Father's education
Less than high school	28%	---
High school graduate	33	---
Some college	25	---
College graduate	14	---
Mother's race
White	28%	---
African-American	42	---
Latino	25	---
Other	4	---
Father's race
White	26%	---
African-American	44	---
Latino	26	---
Other	4	---
Father: Better if parents married	.83	.38
Father: Father earn main living	.39	.49
Father: Women cannot be trusted	.09	.29
Mother: Better if parents married	.73	.45
Mother: Father earn main living	.29	.46
Mother: Men cannot be trusted	.13	.34
Father worked last week	.84	.37
Mother worked last year	.74	.44
No sexual fidelity conflict	.70	.46
Father drinks-3 or more	.40	.49
Father drinking/drugs interfere	.10	.30
Father uses street drugs	.07	.25
Mother street drugs-pregnant	.04	.18
Mother drinking/drugs interfere	.02	.15
Father is supportive	2.72	.35
Mother is supportive	2.73	0.33
Father is not hurtful	.64	.48
Mother is not hurtful	.73	.44
No domestic violence	.88	.33

Note: Percentages may not sum to 100% due to rounding error.

N = 2,034 couples

Independent Variables

We employ one measure, frequent attendance at religious services, to measure the extent to which urban mothers and fathers are integrated into a religious institution. Our assumption is that church attendance is an indicator of a respondent's integration into the social networks and norms associated with religious integration. We also think that religious attendance is an indicator of individual religiosity (e.g., religious belief and prayer) and integration into a religious nomos. Unfortunately, the survey does not have measures of religiosity besides attendance and religious denomination in its first two waves.

The Fragile Families survey measures religion ordinally. We separately coded all mothers and fathers who report attending church “several times a month” or “once a week or more” as frequent attendees. We combine these categories because most respondents probably view “several times a month” and “once a week” as synonymous. Separate dummies measure couples where the man attends regularly, the woman attends regularly, and both attend regularly; neither attends is the reference category. Unfortunately it is not possible to know whether partners attend church together.

Church attendance, rather than denomination, is the component of religious identity that most strongly predicts relationship quality (Myers, 2006; Wilcox, 2004). Nevertheless, all analyses control for religious tradition to account for denomination-specific differences in religious participation. We use race and denominational affiliation to code respondents into five categories; an additional dummy variable measures whether mother and father have the same denominational affiliation. African-American respondents who report Protestant affiliations are coded as Black Protestants (Lincoln and Mamiya, 1990; Steensland et al., 2000). White, Hispanic, Asian American, and Native American respondents who indicate a Baptist, Pentecostal, or other sectarian denominational affiliation are coded as conservative Protestant (Steensland et al., 2000). Respondents who indicate a Catholic affiliation are coded as Catholic. Respondents who report Episcopalian, Lutheran, Congregational, Jewish, Mormon, Muslim, and Presbyterian affiliation are coded as other, as are those who report being members of other faiths. These religious traditions are uncommon, especially among unwed couples, in urban America; there are too few respondents in each category to be the basis of substantive conclusions. Finally, respondents who report no religious affiliation are coded as None (the reference category). An additional item measures whether mother and father share the same denominational affiliation.

Other independent variables, intended to account for the association between church attendance and relationship satisfaction, are organized into three groups. The first group focuses on relationship norms and attitudes. These include measures of familism (children fare better when their parents are married), gender traditionalism (men should earn a living and wives should take care of children), and gender distrust (the opposite sex cannot be trusted to remain monogamous). These items are measured separately for men and women, and are coded as dichotomies (agree vs. disagree).

The second group of independent variables taps adherence to the code of decency—that is, behaviors like temperance, legal employment, and fidelity. First, these include measures of employment status for both men and women: men were asked if they worked during the last week, women in the last year. The difference in question wording is understandable given that female respondents are interviewed within a few days of childbirth. Second are measures of alcohol and drug consumption for both respondents. Fathers were asked if there was a time in the past three months they had three or more drinks in a day, if drinking or drug use affected either employment or personal relationships in the previous year, and whether they used street drugs several times a month or more in the past three months. Mothers were asked whether they used any street drugs during their pregnancy and whether drinking or drug use affected either employment or personal relationships in the previous year. These items are all coded as yes/no binary variables. The final variable in this group, tapping the presence of conflict over sexual fidelity, is dyadic. Also coded as a dichotomy, it measures whether both partners feel their relationship is devoid of conflict over fidelity. Note that it does not specifically measure actual infidelity.

The third group of independent variables measures relationship-specific behaviors helpful or harmful to intimate relationships. Two of these items are measured separately for men and women. Respondents were asked three questions about their partners: how often is s/he “fair and willing to compromise when you have a disagreement?”; how often does s/he express “affection or love for you?”; and, how often does s/he “encourag[e] or hel[p] you to do things that are important to you?” Scores on these items were combined into a scale measuring the presence of highly supportive partners with strong relationship. A second item inquires for each respondent how often his or partner “insults or criticizes you or your ideas.” This is a dichotomous variable, coded 1 if respondents report no insulting or criticizing. Note that measures of supportive and hurtful behavior are not highly correlated (the strongest correlation between perceived supportive and hurtful behaviors is .26); more generally, we observe no evidence of collinearity. A dyadic, dichotomous variable ascertains the presence of any domestic violence (specifically, hitting or slapping) in the relationship.

Finally, we control for a variety of social and demographic factors to account for potential spuriousness in the relationship between religious participation and perceived relationship quality (Christiano, 2000; Myers, 2006): mother age and father age (in years); mother's education and father's education (from less than high school to college-educated; less than high school is the reference category); race/ethnicity (White, African-American, Hispanic, other; White is the reference category); whether each respondent was living in an intact family at age 15; and marital status (married vs. unmarried).

Analysis

We begin by presenting means for our three groups of independent variables. Means are shown separately according to the church-going behavior of each couple: male frequent attendees, female frequent attendees, both partners attend, neither partner attends. This enables us to test the hypothesis that individuals and couples who regularly attend church have beliefs and behaviors conducive to happy relationships. Significance tests between groups are performed using the Bonferonni adjustment.

Next we explore the effects of religious involvement and other factors on overall relationship quality for men and women in intimate relationships, both married and unmarried. These data are clustered within mother-father dyads, so perceived relationship quality for each partner within a couple may be affected by the same unmeasured independent variables. A solution is seemingly unrelated regression models (e.g., Greene, 2002). These provide efficient estimates because error terms are allowed to be correlated across equations for men and women. They also allow us to test differences in the effects of men's and women's religious participation across equations within each model.

We estimate a total of five regression models, each containing separate equations for male and female respondents. The first model includes only measures of church attendance and control variables: age, education, race, religious denomination, marital status, and family structure during adolescence. In the next three models we separately add our three blocks of independent variables to test various explanations for the relationship between religious participation and union quality. The fifth and final model includes all three blocks of independent variables.

RESULTS

Table 2 shows how respondent's beliefs and behaviors regarding intimate relationships vary by religious participation. Within each row, means with common superscripts are significantly different from their counterparts in different columns. The results presented in Table 2 indicate that couples in which both partners attend church frequently have more traditional attitudes about intimate relationships, a stronger adherence to values that should facilitate happy relationships (i.e., the code of decency), and a higher incidence of supportive behaviors within relationships, compared to couples where neither partner attends on a frequent basis. These are mainly dyadic effects: both partners must attend church frequently in order to benefit. (About one fifth of couples in the sample report uniformly frequent religious participation.) Indeed, on many outcomes, such as domestic violence, couples where only one partner attends are no different from couples where neither partner attends church on a frequent basis. Also, there are no statistically significant differences in attitudes, values, and behaviors between couples where only the man attends church frequently and those where only the woman attends frequently.

Table 2.

Beliefs, Attitudes, and Practices Regarding Relationship Behaviors, by Religious Participation.

	Religious Participation
	(A) Male frequent attendees n = 192	(B) Female frequent attendees n = 354	(C) Both frequent attendees n = 431	(D) Neither frequent attendees n = 1,057
Attitudes about Marriage
Father: Better if parents married	0.86	0.81 c	0.91 b,d	0.80 c
Father: Father earn main living	0.48 d	0.36 c	0.47 b,d	0.36 a,c
Father: Women can't be trusted	0.14	0.10	0.07	0.09
Mother: Better if parents married	0.73 c	0.75 c,d	0.85 a,b,d	0.66 b,c
Mother: Father earn main living	0.34	0.29 c	0.38 b,d	0.25 c
Mother: Men can't be trusted	0.14	0.17	0.10	0.14
Adherence to Values
Father worked last week	0.80 c	0.81 c	0.90 a,b,d	0.83 c
Mother worked last year	0.74	0.72	0.73	0.75
No sexual fidelity conflict	0.65 c	0.63 c	0.82 a,b,d	0.68 c
Father drinks-3 or more	0.30 d	0.39 c,d	0.27 b,d	0.47 a,b,c
Father drinking/drugs interfere	0.10	0.10	0.06 d	0.09 c
Father uses street drugs	0.04	0.07 c	0.02 b,d	0.12 c
Mother street drugs-pregnant	0.07 c	0.03	0.01 a,d	0.04 c
Mother drinking/drugs interfere	0.05	0.02	0.02	0.02
Conducive Behaviors
Father is supportive	2.71 c	2.71 c	2.80 a,b,d	2.71 c
Mother is supportive	2.75	2.70	2.77 d	2.71 c
Father is not hurtful	0.74	0.74	0.76	0.71
Mother is not hurtful	0.66	0.63	0.67	0.63
No domestic violence	0.84 c	0.87 c	0.94 a,b,d	0.86 c

Note: Means are significantly (p < .05) different from those in columns denoted by subscripted letters.

N = 2,034 couples

The results presented in Table 2 support our hypothesis that religious participation fosters traditional attitudes about relationships, adherence to a code of decency, and supportive behaviors within relationships; these attitudes and behaviors should in turn lead to happier relationships. We now turn to multivariate analysis to test this idea.

Table 3 displays the effects of church attendance and other factors on men's and women's assessments of perceived relationship satisfaction. Both men and women report significantly better relationships when both attend church several times a month or more. At one-seventh to one-eighth of a standard deviation on the dependent variable for both men and women, the effects are not large; rather, church attendance is one of a number of factors that increases relationship happiness for men and women. Still, we note that religious participation has larger effects on union quality than employment or education (result not shown).

Table 3.

Seemingly Unrelated Regressions of Union Satisfaction on Religious Attendance.

	Men		Women		Men b = Women b
	b	SE b	b	SE b
Religious Service Attendance
Father attends frequently	.22*	.08	−.02	.09	*
Mother attends frequently	.00	.07	.05	.07	n.s.
Both attend frequently	.17**	.06	.17*	.07	n.s.
Neither attends	---	---	---	---	---
Spousal Religious Denomination
Black Protestant	.15	.10	.07	.10	---
Catholic	.01	.09	−.11	.10	---
Conservative Protestant	−.03	.10	−.07	.10	---
Other	.10	.10	.23*	.10	---
None	---	---	---	---	---
Spouses same denomination	.01	.05	.08	.05	---
Married	.40***	.06	.38***	.07	---
Constant	−2.20***	.14	2.32***	.14	---
Error correlation	.38***

Note: Model controls for age, education, race, and family structure at age 15.

N = 2,034 couples

^*p < .05

^**p < .01

^***p < .001 (2 tail tests)

Women only benefit from religious participation if both they and their partners attend frequently. Their contentment with their relationship does not increase if only they attend, or only their partners attend. In contrast, men report happier relationships even if they are the only ones attending church frequently. This suggests that religion leads men to perceive their relationships more positively. Religious authorities may endorse committed intimate unions as optimal for child-rearing, or simply as a preferred venue for human existence; alternatively, churchgoers may believe that God is present in their relationships. Men may be particularly affected by the social and cultural consequences of churchgoing because they are less likely to get pro-relationship cues from other institutions. Either way, for urban couples with children, solo churchgoing seems to make men but not women look favorably on their relationship. In accordance with previous research, then, our results suggest that men more than women are affected by the institutional contexts of their interpersonal relationships (Nock, 1998; Wolfinger and Wilcox, 2008).

Consistent with prior research (Waite and Gallagher, 2000), marriage has a large and statistically significant effect on both men's and women's satisfaction with their relationships. On the other hand, with one exception religious denomination does not affect perceived union quality; nor does religious homogamy. Women whose partners fall into the ‘other’ denomination category report significantly happier relationships. We are loath to attach meaning to results involving residual categories. It is not surprising that denominational affiliation does not affect perceived union quality, given that church attendance is the component of religion that most affects relationships (Myers, 2006). Finally, the error correlation term is large and statistically significant; couples share many unobserved predictors of relationship satisfaction.

Religious participation may influence happiness in romantic relationships by affecting how people view such relationships in general. Table 4 explores this idea by determining whether attitudes and norms regarding intimate relationships affect the association between religious participation and union quality. We employ three measures of attitudes, tapping familism (children fare better when their parents are married), traditionalism (men should earn a living and wives should take care of children) and gender distrust (the opposite sex cannot be trusted to remain monogamous). All three reflect generic values--not necessarily related to specific relationships between Fragile Families respondents--that may be fostered by religious institutions. The results show that many of these cultural variables have predictable effects on perceived relationship quality. For instance, men who feel the opposite sex cannot be trusted report worse relationships, and both men and women who value marriage report better relationships. However, the attitude variables do not account for the relationship between church attendance and perceived relationship quality. The religious regression coefficients in Table 4 do not change substantially from their counterparts in Table 3, lacking the three attitudinal measures. Moreover, the coefficients measuring church attendance are statistically significant in both models.

Table 4.

Seemingly Unrelated Regressions of Union Satisfaction on Religious Attendance and Relationship Attitudes.

	Men		Women		Men b = Women b
	b	SE b	b	SE b
Religious Service Attendance
Father attends frequently	.22**	.08	−.02	.09	*
Mother attends frequently	.01	.07	−.04	.07	n.s.
Both attend frequently	.17*	.06	.17*	.07	n.s.
Neither attends	---	---	---	---	---
Spousal Religious Denomination
Black Protestant	.15	.10	.06	.10	---
Catholic	.02	.09	−.11	.10	---
Conservative Protestant	−.01	.10	−.07	−.10	---
Other	−.10	.10	.23*	−.10	---
None	---	---	---	---	---
Spouses same denomination	.02	.05	.08	.06	---
Attitudes about Relationships
Father: Better if parents married	.20***	.06	.14*	.07	---
Father: Father earn main living	.01	.05	.01	.05	---
Father: Women cannot be trusted	.23***	.08	−.08	.09	---
Mother: Better if parents married	−.07	.05	−.11	.06	---
Mother: Father earn main living	−.07	.05	−.00	.06	---
Mother: Men cannot be trusted	.16*	.07	.19*	.08	---
Married	.37***	.06	.37***	.07	---
Constant	−2.26***	.15	−2.33***	.15	---
Error correlation	.38***

Note: Model controls for age, education, race, and family structure at age 15. N = 2,034 couples

^*p < .05

^**p < .01

^***p < .001 (2 tail tests)

Almost all religious institutions in contemporary urban America support personal responsibility, fidelity, and temperance. In all likelihood, these norms are both causes and consequences of prosperity and personal stability. Can they account for the relationship between religious attendance and union quality? Table 5 explores this question by adding measures of employment, alcohol and drug use, and sexual infidelity to Table 3. Together these factors explain the relationship between churchgoing and union quality for women. The absence of sexual jealousy or a partner with a substance abuse problem improves women's happiness with their relationships; as a result, the regression coefficient measuring the effect of religious participation shrinks and loses statistical significance (We have little insight into why women who used street drugs during pregnancy are happier in their relationships. Perhaps the ability to make such a socially unacceptable admission in a survey indicates an unusually high degree of candor.) Men also benefit when they do not have a problem with drugs or alcohol, or conflict regarding sexual infidelity. These results also suggest that couples who both attend church worry less that their partners will be unfaithful, which in turn leads to happier relationships.

Table 5.

Seemingly Unrelated Regressions of Union Satisfaction on Religious Attendance and Adherence to Values.

	Men		Women		Men b = Women b
Religious Service Attendance	b	SE b	b	SE b
Father attends frequently	.21*	.08	−.02	.09	*
Mother attends frequently	−.00	.07	−.05	.07	n.s.
Both attend frequently	.13*	.06	.13	.07	n.s.
Neither attends	---	---	---	---	---
Spousal Religious Denomination
Black Protestant	.14	.09	.07	.09	---
Catholic	.02	.09	−.11	.10	---
Conservative Protestant	.−01	.10	−.05	.10	---
Other	−.12	.10	.24*	.10	---
None	---	---	---	---	---
Spouses same denomination	.00	.05	.07	.06	---
Adherence to Values
Father worked last week	−.09	.07	.01	.07	---
Mother worked last year	.08	.05	−.07	.06	---
No sexual fidelity conflict	.34***	.05	.35***	.06	---
Father drinks-3 or more	−.04	.05	−.04	.05	---
Father drinking/drugs interfere	−.25**	.08	−.21*	.09	---
Father uses street drugs	−.10	.10	−.22*	.10	---
Mother street drugs-pregnant	.06	.13	.29*	.14	---
Mother drinking/drugs interfere	−.18	.16	−.26	.17	---
Married	.31***	.06	.30***	.07	---
Constant	−2.32***	.15	−2.45***	.16	---
Error correlation	.37***

Note: Model controls for age, education, race, and family structure at age 15. N = 2,034 couples

^*p < .05

^**p < .01

^***p < .001 (2 tail tests)

The story is different for couples in which only the man attends church frequently. The variables measuring employment, sexual jealousy, and substance use cannot explain the relationship between men's attendance and their perceptions of a happy relationship. This result supports our earlier contention: religious participation affects how men perceive their relationships, above and beyond whatever effects it has on their behavior.

We estimated an additional model, not shown, containing only measures of religious participation, control variables, and men's and women's employment. Although working may signify adherence to norms of self-reliance, it might also reflect local economic conditions. Our analysis suggests that employment status alone does not account for the relationship between religious participation and perceived relationship quality. After including the work variables, the coefficients measuring the effects of church attendance are virtually the same as they are in Table 3. Measures of “decency” aside from work—temperance and sexual fidelity—help explain the effect of church attendance on union quality.

Table 6 ascertains whether frequent church attendance increases relationship quality by promoting interpersonal dynamics conducive to good relationships. These include a dyadic measure of domestic violence, as well as separate items for each partner indicating supportive and hurtful behaviors. Together these measures largely account for the association between joint religious participation and perceptions of relationship quality. However, fathers who attend regularly by themselves still perceive their relationships as better. The other participation coefficients are generally smaller than in models with other blocks of independent variables (Tables 3-5) and not statistically significant. For both men and women, most of the coefficients measuring domestic violence, supportive, and hurtful behaviors are statistically significant and in the predicted direction. These results suggest that religious participation benefits romantic relationships through the positive behaviors it encourages—and, conversely, the negative behaviors it discourages.

Table 6.

Seemingly Unrelated Regressions of Union Satisfaction on Religious Attendance and Behaviors Conducive to Relationships.

	Men		Women		Men b = Women b
	b	SE b	b	SE b
Religious Service Attendance
Father attends frequently	.18*	.08	−.06	.09	**
Mother attends frequently	−.01	.06	−.09	.07	n.s
Both attend frequently	.10	.06	.05	.07	n.s
Neither attends	---	---	---	---	---
Spousal Religious Denomination
Black Protestant	.14	.09	.10	.09	---
Catholic	.02	.09	−.05	.09	---
Conservative Protestant	−.01	.09	.03	.10	---
Other	−.12	.10	.21*	.10	---
None	---	---	---	---	---
Spouses same denomination	−.01	.05	.07	.05	---
Conducive Behaviors
Father is supportive	.52***	.07	.26***	.07	---
Mother is supportive	.29***	.07	.86***	.08	---
Father is not hurtful	.20***	.05	.10	.05	---
Mother is not hurtful	.11*	.05	.30***	.06	---
No domestic violence	.13	.07	.14	.08	---
Married	.37***	.06	.34***	.06	---
Constant	−4.74***	.28	−5.89***	.29	---
Error correlation	.34***

Note: Model controls for age, education, race, and family structure at age 15.

N = 2,034 couples

^*p < .05 ;

^**p < .01 ;

^***p < .001 (2 tail tests)

Table 7 includes all blocks of independent variables that had been separately introduced in Tables 4-6. Predictably these variables attenuate most of the relationship between church attendance and union quality. None of the religious attendance coefficients corresponding to women's perceived relationship quality are significant. As in Table 6, the coefficient reflecting just men's attendance is the only statistically significant predictor of men's perceived relationship quality.

Table 7.

Seemingly Unrelated Regressions of Union Satisfaction on Religious Attendance, Attitudes about Relationships, Adherence to Values, and Behaviors Conducive to Relationships.

	Men		Women		Men b = Women b
	b	SE b	b	SE b
Religious Service Attendance
Father attends frequently	.19*	.08	−.06	.09	**
Mother attends frequently	−.00	.06	−.08	.07	n.s
Both attend frequently	.09	.06	.05	.07	n.s
Neither attends	---	---	---		---
Spousal Religious Denomination
Black Protestant	.13	.09	.10	.09	---
Catholic	.03	.09	−.06	.09	---
Conservative Protestant	.02	.09	.03	.10	---
Other	−.13	.10	.23*	.10	---
None	---	---	---	---	---
Spouses same denomination	−.00	.05	.07	.05	---
Attitudes about Relationships
Father: Better if parents married	.17*	.06	.11	.06	---
Father: Father earn main living	.01	.05	−.00	.05	---
Father: Women cannot be trusted	.16*	.08	−.05	.08	---
Mother: Better if parents married	−.08	.05	.10	.06	---
Mother: Father earn main living	−.05	.05	.03	.06	---
Mother: Men cannot be trusted	−.06	.07	−.04	.07	---
Adherence to Values
Father worked last week	−.07	.06	.00	.07	---
Mother worked last year	.03	.05	.10	.06	---
No sexual fidelity conflict	.22***	.05	.16**	.06	---
Father drinks-3 or more	−.02	.05	.08	.05	---
Father drinking/drugs interfere	.18*	.08	−.12	.08	---
Father uses street drugs	.02	.09	−.14	.10	---
Mother street drugs-pregnant	.07	.13	.33*	.13	---
Mother drinking/drugs interfere	−.17	.15	−.24	.16	---
Conducive Behaviors
Father is supportive	.48***	.07	.24**	.07	---
Mother is supportive	.21**	.07	.82***	.08	---
Father is not hurtful	.18***	.05	.08	.05	---
Mother is not hurtful	.09	.05	.29***	.06	---
No domestic violence	.09	.07	.10	.08	---
Married	.32***	.06	.32***	.06	---
Constant	−4.47***	.29	−5.73***	.31	---
Error correlation	.34***

Note: Model controls for age, education, race, and family structure at age 15.

N = 2,034 couples

^*p < .05

^**p < .01

^***p < .001 (2 tail tests)

Taken together, these results suggest three ways in which religious participation benefits intimate relationships. First, couples benefit when both partners attend church. There appear to be two mechanisms that account for this result. Religious attendance is correlated with a lifestyle that accords with the code of decency, including temperance and a lack of conflict over sexual fidelity. Adhering to these practices improves relationship quality. However, we cannot know whether these practices cause or are caused by religious attendance. Second, religious attendance appears to increase the incidence of interpersonal behaviors conducive to good relationships, such as affection. Perhaps these behaviors influence joint religious attendance (Booth et al. 1995).

When just men attend, the story is different. Irrespective of other independent variables, men's religious participation improves their own perceptions of union quality. Recall also from Table 2 that men who attend alone are not more likely to engage in supportive relationship behaviors than men in a relationship where neither partner attends. Thus, our results suggest another way that religion may promote better relationships among urban parents: religious attendance may increase the likelihood that men view their relationships in a more rose-colored light.

CONCLUSION

This study finds that religious attendance—men's attendance alone and dyadic attendance—is associated with higher reports of relationship quality for both married and unmarried couples with children in urban America. We suspect that religious attendance influences both partners’ relationship behaviors and their perceptions of the relationship itself, independent of those behaviors. In other words, shared religious attendance appears to encourage supportive, nonviolent behaviors that improve relationship quality, and men's attendance alone seems to promote a more positive view of the relationship. Thus, from a theoretical perspective, this study suggests that shared religious social capital is associated with higher levels of domestic capital; in this case, relationship quality.

Somewhat to our surprise, the relationship between religion and relationship quality is not explained by marriage or gender norms. This may be because urban churches often choose to avoid speaking directly about marriage, sex, nonmarital childbearing and other relationship-specific issues (Ellingson, 2004). They do so because there is often a tension between their theology, which tends to support more traditional family and sexual norms, and the behaviors of their congregants, who often hail from nontraditional families. Thus, if churches are not emphasizing marriage norms, it would help explain why these norms do not account for the association between religion and the quality of urban relationships. It might also explain why religious attendance seems to affect married and unmarried couples in fairly similar ways, at least when it comes to relationship quality (Wolfinger and Wilcox, 2008). More generally, our results are consistent with the finding that declines in religious authority over family life have made religiously-grounded family norms less important for contemporary relationships (Myers, 2006).

But the norms, networks, and nomos associated with urban churches do appear to foster a code of decency among urban parents—especially men—that makes for happier relationships. Specifically, we find that the association between religious attendance and relationship quality seems to be partly accounted for by the man's alcohol and drug consumption, and the couple's conflict over infidelity. These variables have strong implications for relationship quality. We also find that the association between religion and relationship quality is partially accounted for by relationship-specific factors—lower levels of domestic violence, lower levels of hurtful behavior, and higher levels of supportive behavior. This suggests that religious partners are more likely to bring certain skills to their relationships, perhaps because the norms, network, and nomos they encounter in their churches encourages them to treat others with respect and affection. More generally, consistent with the larger literature on social capital (Coleman, 1990; Putnam 2000), our results suggest that shared religious social capital fosters positive behaviors that in turn lead to higher levels of reciprocity and trust not only in the civic and political arenas but also in the domestic arena. Consistent with the literature on religion and masculinity (Bartkowski 2004; Wilcox and Bartkowski 2005), our results provide evidence that religion fosters a “decent” form of masculinity that stands in stark contrast with the “street” form of masculinity that often dominates low-income urban communities (Anderson, 1999); this decent masculinity seems to be linked to beneficial outcomes for urban couples. It is also possible that religion—especially insofar as individual religiosity fosters emotional coping among believers (Ellison, 1994)—helps urban couples deal with stresses in their lives in a constructive fashion, thereby allowing them to steer clear of negative relationship behaviors like drug use and domestic violence.

Finally, urban men who attend church frequently are more inclined to take a favorable view of their relationship, even after controlling for their partners’ relationship behaviors. This suggests that religious attendance fosters a positive outlook on the part of urban men that makes them view their relationship in a favorable light, above and beyond anything that their partner is doing in the relationship. It may be that their religious beliefs, or the rituals they engage in, enable them to see their relationship as sanctified; that is, they may see God working in their lives in ways that lead to a more positive assessment of the relationship (Mahoney et al., 2003).

Several limitations of this research should be acknowledged. First, given that we study new urban parents our results do not necessarily generalize to childless couples, or to couples living in suburban and rural America. Second, we cannot be sure that we have adequately specified the direction of causality. It may be, for instance, that men who are striving to live “decent” lives in urban America turn to religious institutions to support their lifestyle (Anderson, 1999). Likewise, urban couples who enjoy quality relationships may be more likely to join and attend a church together than couples who do not enjoy happy relationships (Booth et al., 1995). Future research will have to determine with more precision how the code of decency, relationship-specific behaviors, and global relationship quality are related to one another. Finally, future research should explore how parents’ religiosity is linked to offspring well-being and, how, if at all, any relationships between parental religiosity and child well-being are accounted for by the quality of the couple relationship between the parents.

In sum, this study suggests that churches play a modest but noteworthy role in fostering higher-quality relationships among married and unmarried couples with children living in urban America. They seem to do so by fostering “decent” norms and behaviors, especially among men who attend church with their partners, that lead directly and indirectly to quality relationships. But relationship-specific norms do not appear to play a key role in accounting for the link between faith and family life among these urban couples. That is, urban churches do not promote happy relationships by focusing on marriage; rather, they appear to promote happy relationships by encouraging their members—especially their men—to live decent lives and to treat their partners with decency.