Remarriage Timing: Does Religion Matter?

Abstract

Using pooled data from the National Survey of Family Growth (NSFG 2006–2010), we examine the effects of denominational affiliation, worship service attendance, and religious salience on remarriage timing. Survival analyses indicate that both men and women affiliated with conservative Protestant faith traditions are significantly more likely than their unaffiliated and Catholic counterparts to remarry at an accelerated pace following divorce. Results further show that, net of religious affiliation and socio-demographic characteristics, worship service attendance accelerates remarriage timing, whereas the effects of religious salience are weaker or unobserved. These results are largely consistent with prior research on denominational variations in the timing of first marriage and underscore the robust influence of religion on the institution of marriage.

1. Introduction

The intersections between religion and family have received renewed attention during the past several decades, with scholars now recognizing that religion influences virtually every facet of family life and intimate relationships, such as union formation, housework, household decision-making, child development, parenting, paternal involvement, marital quality, premarital sexual behavior, marital infidelity, and risk of divorce (e.g., Bartkowski 2001; Bartkowski and Ellison 2009; Bartkowski and Xu 2000; Bartkowski et al. 2008; Burdette et al. 2007, 2009; Ellison and Xu 2014; Mahoney 2010) (see Ellison and Xu 2014 for review). Social scientists have established that religion influences people’s entry into marriage, as well as their exit from it. Where entry into first marriage is concerned, Mormons (Latter-day Saints) and conservative Protestants are considerably more likely to marry early, while their counterparts in other faith traditions and those who claim no religious affiliation tend to marry later (Rendon et al. 2014; Xu et al. 2005; Uecker 2014; Uecker et al. 2016). Scholars point to the pro-family character of Latter-day Saints and conservative Protestants in explaining their eagerness to marry young (Xu et al. 2005; Uecker and Stokes 2008). And, as it turns out, religion also influences married persons’ propensity to exit marriage through divorce (Call and Heaton 1997; Ellison and Xu 2014; Lehrer and Chiswick 1993; Vaaler et al. 2009). Generally, religion inhibits divorce; however, marriages marked by shared religious commitments are particularly less likely to dissolve than are religiously heterogamous marriages (Ellison and Xu 2014).

This study is designed to advance research on religion, marriage, and divorce by turning its attention to religion as an antecedent of remarriage. Only one previously published study has examined religion and remarriage (Brown and Porter 2013). However, this is not to say that there has been complete inattention given to religion and remarriage among family and religion scholars. An exchange of opinions—bereft of data—over religion and remarriage occurred decades ago but shed no light on actual empirical patterns (see Halliday 1980; Cherlin 1980). Hout’s (2000) examination of Catholic remarriage drew primarily on descriptive statistics for a lay audience, while Wu examined a host of predictors of Canadian remarriage, with religion being one of many variables explored (Wu 1994). The lone study mentioned previously (Brown and Porter 2013) detected significant effects for denominational affiliation, worship service attendance, and religious salience on the propensity of remarriage among American women. It was reported that women who were conservative Protestants, regular attenders, and those for whom religion was highly salient were more likely to remarry. Regrettably, however, much of the data used in that previous study was two decades old. By contrast, we use more recent and multi-cycle data from the National Survey of Family Growth. In addition, our study includes both men and women, thus allowing us to explore gendered trajectories in the religion-remarriage relationship even as we render findings that are more reflective of current social trends. Finally, while the previous study used binary logistic regression to estimate the effects of religion on the propensity of remarriage, our study employs survival analysis to examine the timing of remarriage while accounting for censoring problems in data that could significantly bias the results.

Given the pervasiveness of divorce in contemporary American society, and the high likelihood for couples to remarry following divorce (Cherlin 2009; Sweeney 2010), this study contributes significantly to sociological scholarship on the religion-family nexus and the growing body of research on the antecedents of remarriage. In particular, this study addresses the following key research questions. Given signs of intergenerational decline in Americans’ religiosity (e.g., attendance) (Schwadel 2011), what influence, if any, does religion exert in remarriage timing among young Americans who have recently experienced a divorce? And, if religion does influence remarriage trajectories, are variations among denominational subcultures evident similar to those observed in the timing of first marriage (Rendon et al. 2014; Uecker 2014; Xu et al. 2005)? In addition, what role, if any, does religious and spiritual capital play in the timing of remarriage? It is possible, for example, that integration within religious networks through regular worship service attendance could serve as a remarriage market among persons of faith. Finally, since both religion and marriage have been characterized as gendered institutions (Bartkowski 2001; Bernard 1972; De Vaus and McAllister 1987; Miller and Stark 2002) and prior research underscores gender differences in remarriage timing (James and Shafer 2012), do gender differences moderate the effect of religion on remarriage timing among young Americans? In the pages that follow, these questions are critically examined.

1.1. Correlates of Remarriage and the Case for Religious Influence

What factors influence the timing of remarriage? Research on remarriage and stepfamilies has been conducted for about forty years now.¹ Early research focused on the prevalence, antecedents, and character of remarriage, and more recent studies have investigated the effects of remarriage on marital quality and child development (see Coleman and Ganong 1990; Van Eeden-Moorefield and Pasley 2013 for reviews). Although still considered an “incomplete institution” given the lack of clear social norms governing stepfamilies, remarriage is now quite common in the United States. In fact, about a third of all Americans will remarry following divorce, and scholars have identified the types of people who are most likely to remarry (Bumpass et al. 1990; Cherlin 2009; Cherlin and Furstenberg 1994; Coleman et al. 2000). For example, studies show that while men are generally more likely to remarry, and do so more quickly than their female counterparts, women who marry at a young age or who divorce younger are more likely to remarry than their older female counterparts. In addition, women with more children are less likely to get remarried, probably because they find it difficult to attract a partner willing to support a larger family (Cherlin 2010; Wilson and Clarke 1992). It appears that remarriage also varies by race-ethnicity, such that non-Hispanic whites are considerably more likely to remarry than their Latino or African American counterparts. It is estimated that about one in two white women will remarry within five years of marital separation as compared with one third of Mexican-American women and one fifth of black women. Of course, socioeconomic status explains a part of these racial-ethnic variations, because the affluent are more likely to remarry than are the poor (Sweeney 1997). When calculated as a proportion of divorced persons, remarriage rates have declined somewhat today compared with previous eras. In large part, this decline is due to post-divorce cohabitation, which has become increasingly common and is one of the most pronounced influences on the delay and decline in remarriage (Xu et al. 2006, 2011). Scholars suspect that “the provisional nature of informal unions may be part of their appeal for individuals who may be hesitant, at least temporarily, to recommit to formal marriage” (Cherlin and Furstenberg 1994, p. 362).

As noted above, a great deal of scholarly attention has been paid to the character of stepfamilies and the effects of remarriage on a wide range of familial, interpersonal, and developmental outcomes (Coleman et al. 2000). However, in the move from outcomes to antecedents associated with remarriage, the effects of religion on the timing and propensity of remarriage have been explored in only one previous investigation (Brown and Porter 2013). That study indicates that religiosity, and especially affiliation with a conservative Protestant faith, hastens remarriage. However, that study also exhibits methodological shortcomings by focusing only on women, using data as much as two decades old, and employing regression methods that are not ideally suited for examining temporal trajectories. The present study fills a research void in scholarship on family life and religious involvement.

1.2. The Linkages between Religion and Remarriage Timing

1.2.1. The Religious Subculture Thesis

Why might religion have effects on the timing of remarriage? There are several compelling reasons. According to the religious subculture thesis, religious denominations can be rank-ordered along a liberal-moderate-conservative continuum across a range of “pro-family” issues, which can help account for both between-denominational differences and within-denominational homogeneity/heterogeneity (Gay et al. 1996; Hoffmann et al. 2017). Employing this religious subculture thesis, recent scholarship has highlighted distinctive denominational subculture variations in key patterns of family formation, including the timing of first marriage. Early evidence revealed that those claiming a religious affiliation are more likely to marry than are the religiously unaffiliated (Heaton and Goodman 1985). And where marriage timing is concerned, conservative Protestants and Latter-day Saints (Mormons) marry, on average, at the youngest ages (Xu et al. 2005). Other, less culturally distinctive religious groups are more likely to postpone marriage at rates akin to those in the general population.

As indicated by the religious subculture thesis, attitudes toward marriage and divorce vary by religious affiliation. Conservative religious groups are significantly more likely to view marriage as a lifetime commitment only to be ended under the most extreme circumstances (Wilson and Musick 1995). Such attitudinal opposition to divorce is connected to conservative Protestants’ view of the Bible, which they esteem as divine, inerrant revelation to humankind. The Bible valorizes marriage and describes God’s “hatred” for divorce (Malachi 2:16). In addition, Conservative Protestants are more dependent on marriage for their subjective wellbeing (Wilson and Musick 1996), and demonstrate more subjective commitment to their own marriage than do those in other faith traditions and the religiously unaffiliated (Wilson and Musick 1995). For their part, mainline religious groups have shown considerably more accommodation to divorce than their conservative counterparts. Catholicism presents a more complicated picture, such that the official doctrine of the Catholic Church does not recognize the legitimacy of divorce, although American Catholics are somewhat accepting of it (Sander 1995). Moreover, some Catholic leaders wish to reform the Catholic Church’s longstanding condemnation of divorce, as evidenced in debates that occurred during a recent worldwide meeting of Catholic bishops, namely, the 2014 Synod on the Family. In short, conservative Protestants are the most pro-marriage major faith tradition in the U.S., and it is no surprise that conservative Protestants are at the forefront of a movement to reinvigorate marriage in a culture of divorce (Gay et al. 1996; Xu et al. 2005). Therefore, there is a good reason to anticipate similar religious subculture variations in the case of remarriage timing.

1.2.2. The Religious Capital Thesis

Recently echoing popular theoretical perspectives associated with social or cultural capital (Bourdieu 1986; Coleman 1990; Putnam 2000), scholars of religion have developed a theoretical perspective known as religious capital, which is defined as “the degree of mastery of and attachment to a particular religious culture” (Stark and Finke 2000, p. 120). Resembling their social or cultural capital counterparts, the religious capital thesis emphasizes the investment in institutionalized faith traditions, especially integration in religious networks that are often measured by attendance at worship services. Prior research showed that religious capital was positively associated with marital stability and inversely associated with marital dissolution (Call and Heaton 1997; Heaton and Pratt 1990; Lehrer and Chiswick 1993; Vaaler et al. 2009). Very generally, the accumulation of religious capital has a protective effect against divorce. Such protections are amplified for couples who share the same denominational affiliation (Heaton and Pratt 1990; Lehrer and Chiswick 1993; Vaaler et al. 2009) and who both attend worship services frequently (Call and Heaton 1997; Lehrer and Chiswick 1993). The protective effects of religious homogamy against divorce have been shown to persist over the duration of marriage (Heaton et al. 1985). These effects are likely connected to the ways in which religious homogamy promotes marital satisfaction (Heaton 1984) and reduces marital conflict (Curtis and Ellison 2002). More recent scholarship has revealed a somewhat more complicated portrait, demonstrating that interfaith marriages can provide protections against divorce for couples whose husbands are theologically conservative (Vaaler et al. 2009). However, religious capital is not uniformly protective in relation to divorce. The risk of divorce is elevated for couples whose husbands attend more frequently than their wives and for those in which wives embrace more theologically conservative beliefs than their husbands (Vaaler et al. 2009).

Turning to remarriage timing, religious capital measured as attendance at worship services can be important in several ways. First, those who have experienced a divorce but attend worship services frequently can regularly receive moral proclamations from the pulpit about the importance of marriage or remarriage and other pro-family messages or teachings, which can help rebuild confidence in the institution of marriage after divorce. Second, frequent attendance at worship services can provide opportunities to cultivate and increase religious capital through which one can interact with co-religionists to enhance their views of remarriage and perhaps find a remarriage partner. And finally, frequent attendance at worship services can serve as an indication of continued religious commitment, including commitment to remarriage as a particular type of recoupling.

1.2.3. The Spiritual Capital Thesis

Distinct from religious capital, spiritual capital treats religious knowledge, competencies, and preferences as positional goods that can be exchanged (Verter 2003). Spiritual capital can be manifested in the investment in non-institutionalized or individualized religious activities, such as in-home devotional efforts, private prayers, personal scripture study, or religious salience. This form of capital can also affect remarriage timing largely because it reflects how important religion is to a person and the extent to which they have internalized the religious norms, values, and teachings pertaining to remarriage and family life after divorce. Individuals who possess a great deal of spiritual capital are more likely to internalize these norms and values and are more inclined to utilize religious teachings and resources to inform their remarriage decisions. As such, there are good reasons to believe that spiritual capital can promote remarriage, thus accelerating remarriage timing after divorce. This pattern might be especially true for those who are affiliated with conservative Protestant congregations because this faith tradition places a premium on developing a “personal relationship” with deity through private forms of religious devotion. Such a pattern may be considerably less evident in more institutionalized forms of religion such as Catholicism because these traditions emphasize organizational attachment to a specific denomination.

1.2.4. The Gendered Institution Thesis

Marriage and religion have long been viewed as gendered institutions. As initially observed by Bernard, there are two different marriages, his and hers, such that marriage typically benefits husbands more than wives (Bernard 1972). Contemporary scholarship on marriage continues to document gendered boundaries, segregated marital roles, and gender-differentiated meanings in the institution of marriage. And, even remarriage itself is subject to gendered trajectories (e.g., James and Shafer 2012). Where the gendered character of religiosity is concerned, scholars have observed that women tend to be more religious than men (see Trzebiatowska and Bruce 2012) and that religious convictions vary by gender (Bartkowski and Hempel 2009; Hoffmann and Bartkowski 2008). Three of the most prevalent explanations generally offered for women’s greater religiosity are as follows. First, some scholars contend that women are more inclined towards religion because they are more risk-averse. Based on this principally psychological explanation, irreligiosity is a form of risk-taking on a cosmic scale (risk of sin in this life and eternal damnation hereafter). Second, others advance a socialization argument. Based on this explanation, women’s greater exposure to values such as nurturance, meekness, and empathy during their childhood lead them to gravitate toward similar religious values in adulthood. Finally, some sociologists of religion point to women’s structural locations in society. This social location argument charges that women’s greater responsibility for childrearing combines with their lower rates of labor force participation and prioritization of family life, all of which lead to a stronger religious orientation than their male counterparts. These arguments suggest that denominational subculture differences in remarriage and the effects of religious and spiritual capital on remarriage timing could be significantly different for men and women.

Guided by the four theoretical perspectives reviewed above, the following hypotheses are proposed.

• H1: Denominational subculture hypothesis—Conservative Protestants will exhibit earlier entry into remarriage after divorce than those who are affiliated with other faith traditions, especially Catholics, as well as those who are religiously unaffiliated.
• H2: Religious capital hypothesis—Stocks of religious capital as measured by attendance at worship services will accelerate remarriage, that is, reduce the amount of waiting time between divorce and remarriage.
• H3: Spiritual capital hypothesis—Stocks of spiritual capital as measured by religious salience will accelerate remarriage, that is, reduce the amount of waiting time between divorce and remarriage.
• H4: Gendered institution hypothesis—The strength and significance of associations between denominational affiliation, religious capital, and spiritual capital and remarriage timing will be different for men and women.

In summary, it is hypothesized that there will be significant denominational differences in remarriage timing, with conservative Protestants conforming to a pattern of accelerated remarriage. Conservative Protestants might couple pro-marriage and anti-divorce orientations in way that is anticipated to hasten remarriage as atonement for the “stigma” of divorce, a pattern called “redemptive remarriage” in earlier research (Brown and Porter 2013). By contrast, it is expected that Catholicism will be marked by a pattern of delayed remarriage, given this tradition’s institutional barriers to divorce and remarriage. Such requirements as the annulment of a first marriage prior to a church-sanctioned remarriage will dramatically lengthen post-divorce waiting time for remarriage among Catholics.

2. Research Methodology

2.1. Data

To address the research questions and test the hypotheses as specified previously, the present study draws on data from the 2006–2010 cycles of the National Survey of Family Growth (NSFG 2006–2010). The Institute for Social Research (ISR) at the University of Michigan conducted the surveys, from June 2006 through June 2010, under contract from the National Center of Health Statistics. Like their previous cycles, the NSFG 2006–2010 was designed to provide reliable national data on cohabitation, marriage, divorce, remarriage, contraception, infertility, and the health of women and infants in the United States. The pooled NSFG 2006–2010 sample is nationally representative of the civilian, non-institutionalized population, consisting of 10,403 men and 12,279 women ages 15–44. Given the focus of the present study on remarriage timing among young Americans, 1979 divorced women and 1218 divorced men were included as the analytic sample for statistical analysis. The NSFG 2006–2010 surveyed young Americans aged 44 or younger because of its focus on reproductive health. This data collection factor can lead to an age truncation problem, which may limit the estimation of the religious effects on remarriage timing because of its disproportionate inclusion of divorced but unmarried (not remarried) respondents, thus potentially limiting the generalizability of the study findings. Despite this possible age truncation problem, the NSFG 2006–2010 contains excellent life course transition questions pertaining to recoupling such as post-divorce cohabitation and remarriage, making the data suitable for the present study.

2.2. Dependent Variable: Waiting Time to Remarriage

Consistent with previous research on marriage timing, this study uses an event history approach, more precisely survival analysis, to analyze the NSFG 2006–2010 data. Under the event history analysis framework, the dependent variable was conceptualized and operationalized as the waiting time to remarriage following (first) divorce, which was computed using two different procedures. First, for respondents who were remarried, remarriage timing was calculated by subtracting date of divorce from date of remarriage in century month (year × 12 + month). Second, for respondents who were not remarried, their waiting time to remarriage was calculated by subtracting date of divorce from date of interview (also in century month). Respondents who were not remarried at the time of interview represent censoring cases. One of the major advantages of using event history methods for data analysis is to include these censored observations. It should be noted that logistic regression models used in Brown and Porter’s (2013) earlier study with older data cannot account for censoring problems.

2.3. Focal Covariates

To operationalize the religious denomination subculture thesis, respondents’ current denominational affiliation was used. Because the National Center of Health Statistics did not release the detailed denominational membership variables, the denominational affiliation available in the public use file was pre-collapsed, and was thus incongruent with previous studies that utilized detailed denominational membership (Xu et al. 2005). This variable was first recoded into five broader categories and then dummy-coded into four variables: (1) Catholic, (2) mainline Protestant (e.g., Methodists, Lutherans, Presbyterians, and Episcopal groups), (3) other faith traditions (e.g., Muslims, Jews, Latter-day Saints and Jehovah Witnesses), and (4) the unaffiliated, with conservative Protestant (e.g., Baptists and other fundamentalist Protestants) serving as the reference group.

The religious capital thesis was operationalized as integration in religious networks in terms of frequency of worship service attendance. The NSFG 2006–2010 recorded this variable on an ordinal scale with seven response categories, ranging from 1 = “never attend” to 7 = “attend more than once a week.” For ease of interpretation, worship service attendance was treated as a continuous variable in the statistical analysis.

The spiritual capital thesis was operationalized as a form of subjective identification (internalization of one’s faith), in this case, religious salience. The NSFG 2006–2010 asked how important religion was in respondents’ daily life, which was dummy-coded with 1 = salient (“very important” and “somewhat important”) and 0 = not salient (“not important”). The category of “not salient” was used as the reference. While these focal covariates (or independent variables) represent an important addition to the analyses presented here, a broader set of control variables must also be considered due to their potentially confounding influence.

2.4. Other Covariates Serving as Statistical Controls

In the survival analysis, the following covariates were controlled to avoid possible spuriousness or confounding effects. These covariates typically served as statistical controls in previous research on remarriage. Race/ethnicity was dummy-coded into Black, Hispanic, and other race/ethnicity, with white serving as the reference category. Post-divorce cohabitation combined several measures into a count variable, namely, number of post-divorce cohabitations. Current educational attainment in years of education was dummy-coded into two variables: high school and more than high school with less than high school serving as the reference group. Since respondents’ employment status at time of remarriage was unavailable, current employment status was used and dummy-coded with 1 = working and 0 = not working. To control for intergenerational transmission of union formation and recoupling, family structure at age 14 was used and dummy-coded into 1 = biological two-parent family and 0 = other family arrangement. Current family resources were measured by whether the family received public assistance, which was dummy-coded into 1 = poor and 0 = not poor. While region of residence was not provided in the public use file, metro statistical area was used and dummy-coded into 1 = urban and 0 = rural to control for remarriage market differences. Lastly, the survey year was dummy-coded into three variables: 2007, 2008, and 2010, with 2006 serving as the reference category.

2.5. Analytic Strategies

Consistent with previous research on marriage timing (Rendon et al. 2014; Xu et al. 2005), parametric survival time models or accelerated failure-time (AFT) models were employed. These models are log-time parameterized and denoted as lnt_j = x_j β + z_j, where x_j is a matrix of covariates, β is a vector of regression coefficients to be estimated, and z is the error term (Allison 1995; Cleves et al. 2004; Stata 2015). With this parameterization, the log-transformed waiting time to remarry was regressed on the religion variables and other covariates as a series of nested log-logistic models. Model 1 tests the religious denomination subculture hypothesis net of statistical controls. Models 2 and 3 examine the role of religious and spiritual capital on remarriage timing, respectively. Finally, Model 4 explores all the covariates used in this study. These models were estimated separately by gender to test the hypothesis based on the gendered institution thesis.

3. Results

Table 1. Descriptive Statistics.

	Women		Men
	n	Percent	n	Percent
Waiting Timing a	5.56	4.63	5.35	4.36
Not-Remarried	1256	63.5	757	62.2
Remarried	723	36.5	461	37.8
White (reference)	1142	57.7	737	60.5
Latino	428	21.6	228	18.7
African American	331	16.7	202	16.6
Other Race	78	3.9	51	4.2
Less Than High-School (reference)	423	21.4	286	23.5
High-School	617	31.2	391	32.1
College	939	47.4	541	44.4
Not Working (reference)	575	29.1	203	16.7
Working	1404	70.9	1015	83.3
Family Income a	8.28	3.98	9.74	3.72
Non-Biological Parents at Age 14	850	43.0	460	37.8
Biological Parents at Age 14	1129	57.0	758	62.2
Non-Poor (reference)	1152	58.2	950	78.0
Poor	827	41.8	268	22.0
Rural (reference)	389	19.7	249	20.4
Urban	1590	80.3	969	79.6
Post-Divorce Cohabitation a	0.34	0.75	0.15	0.45
Age at First Divorce a	26.84	5.73	29.35	5.81
Survey Year 2006 (reference)	508	25.7	296	24.3
Survey Year 2007	445	22.5	293	24.1
Survey Year 2008	505	25.5	311	25.5
Survey Year 2010	521	26.3	318	26.1
Unaffiliated	385	19.3	293	23.9
Catholic	420	21.3	235	19.3
Conservative Protestant (reference)	864	43.7	497	40.9
Mainline Protestant	171	8.7	109	9.0
Other Denominations	139	7.0	84	6.9
Worship Service Attendance a	3.59	2.11	3.26	2.09
Religion Unimportant (reference)	71	4.4	104	11.2
Religion Important	1526	95.6	824	88.8
N		1979		1218

^a Mean and SD.

features descriptive statistics for the variables in the present study. As can be seen from the table, about 37% of women and 38% of men were remarried following divorce in the pooled NSFG 2006–2010 data. This finding is not surprising because the data were collected from young Americans. It can be observed as well that, on average, women waited slightly longer than men to remarry after divorce. With reference to the focal covariates, namely, the religious variables, the majority of respondents under this study were conservative Protestants (43.7% for women and 40.9% for men) followed by Catholics (21.3% for women and 19.3% for men) and unaffiliated (19.3% for women and 23.9% for men). On average, both women and men attended worship services several times a month, whereas the vast majority of women and men deemed religion important in their lives (95.6% for women and 88.8% for men). As expected, women in this sample are more religious than men.

Model 1 reported in

Table 2. Log-logistic Survival Regressions to Predict Remarriage Timing among U.S. Women.

	Model 1		Model 2		Model 3		Model 4
Latina	0.243	*	0.330	***	0.306	***	0.281	**
African American	0.547	***	0.553	***	0.491	***	0.596	***
Other Race	−0.026		0.035		−0.009		0.012
High-School	−0.228	*	−0.227	*	−0.236	*	−0.220	*
College	−0.077		−0.043		−0.071		−0.050
Working	0.323	***	0.327	***	0.327	***	0.330	***
Family Income	−0.063	***	−0.062	***	−0.061	***	−0.063	***
Biological Parents at Age 14	0.008		0.031		0.025		0.016
Poor	0.212	**	0.223	**	0.218	**	0.219	**
Urban	0.179	*	0.195	*	0.195	*	0.174	*
Post-Divorce Cohabitation	0.418	***	0.414	***	0.434	***	0.410	***
Age at First Divorce	0.048	***	0.049	***	0.050	***	0.048	***
Survey Year 2007	−0.041		−0.053		−0.046		−0.056
Survey Year 2008	0.097		0.074		0.091		0.083
Survey Year 2010	0.099		0.083		0.087		0.095
Catholic	0.246	*					0.213	*
Mainline Protestant	0.192						0.169
Other Denominations	0.089						0.077
Unaffiliated	0.283	**					0.160
Worship Service Attendance			−0.060	***			−0.047	*
Religious Salience					−0.239		−0.142
Constant	0.783	***	1.026	***	1.036	***	1.104	***
N	1979		1979		1979		1979

Notes: * p < 0.05; ** p < 0.01; *** p < 0.001.

and

Table 3. Log-logistic Survival Regressions to Predict Remarriage Timing among U.S. Men.

	Model 1		Model 2		Model 3		Model 4
Latino	0.253	†	0.412	**	0.369	**	0.304	*
African American	0.347	**	0.404	**	0.287	*	0.442	***
Other Race	0.124		0.102		0.062		0.191
High-School	0.074		0.103		0.068		0.096
College	0.214	†	0.255	*	0.210	†	0.250	*
Working	−0.126		−0.126		−0.141		−0.112
Family Income	−0.097	***	−0.093	***	−0.094	***	−0.097	***
Biological Parents at Age 14	0.130		0.134		0.148		0.131
Poor	−0.240	*	−0.274	*	−0.246	*	−0.259	*
Urban	0.149		0.151		0.180		0.128
Post-Divorce Cohabitation	0.036		0.005		0.051		−0.012
Age at First Divorce	0.028	**	0.028	**	0.030	***	0.027	**
Survey Year 2007	0.002		0.005		−0.005		0.007
Survey Year 2008	−0.034		−0.051		−0.062		−0.046
Survey Year 2010	0.082		0.079		0.059		0.085
Catholic	0.373	**					0.290	*
Mainline Protestant	0.083						0.043
Other Denominations	−0.231						−0.256
Unaffiliated	0.349	**					0.125
Worship Service Attendance			−0.099	***			−0.081	***
Religious Salience					−0.262	†	−0.116
Constant	2.029	***	2.393	***	2.280	***	2.466	***
N	1218		1218		1218		1218

Notes: † p < 0.10; * p < 0.05; ** p < 0.01; *** p < 0.001.

tests the denominational subculture hypothesis (Hypothesis 1). The positive and significant survival regression coefficients for both male and female Catholics and unaffiliated support Hypothesis 1, which predicted shorter waiting times until remarriage for conservative Protestants when compared with Catholics and the unaffiliated. In other words, the expected waiting time to remarriage is 27.9% [(e^0.246 − 1) × 100] and 32.7% [(e^0.283 − 1) × 100] longer, respectively for female Catholics and unaffiliated. Likewise, the expected waiting time to remarriage is 45.2% [(e^0.373 − 1) × 100] and 41.8% [(e^0.349 − 1) × 100] longer, respectively, for male Catholics and unaffiliated persons. A set of ancillary survival regression analyses (not shown in the tables) reveal that men affiliated with other faith traditions remarried sooner than the unaffiliated as well. Overall, these regression findings show that, as hypothesized, conservative Protestants indeed remarry sooner than their Catholic and unaffiliated counterparts regardless of gender. These findings provide strong support for Hypothesis 1 based on the religious denominational subculture thesis in remarriage timing.

Model 2 in Table 2 and Table 3 tests Hypothesis 2, which anticipated that stocks of religious capital measured by worship service attendance would be inversely associated with remarriage timing. The negative survival regression coefficients reported in the tables show that, net of statistical controls, each unit increase in worship service attendance is associated with 5.8% [(e^−0.060 − 1) × 100] and 9.4% [(e^−0.099 − 1) × 100] decreases in expected time to remarry for women and men, respectively. Thus, Hypothesis 2 derived from the religious capital thesis is strongly supported.

Model 3 in the same tables tests Hypothesis 3, which predicted that stocks of spiritual capital measured by religious salience would be inversely associated with remarriage timing. Although the negative survival regression coefficients generally support this hypothesis, the finding is only marginally significant for men but insignificant for women. That is, net of statistical controls, the expected remarriage timing is 23% [(e^−0.262 − 1) × 100] shorter for men who deemed religion important in their lives than those who viewed religion unimportant. Given these inconsistent results, Hypothesis 3 based on the spiritual capital thesis is not supported.

Model 4 displayed in Table 2 and Table 3 combines all the focal covariates along with statistical controls. The survival regression results indicate that there are robust effects of Catholic affiliation and attendance at worship services. Both are statistically significant for women and men. In addition, men affiliated with other faith traditions also remarry sooner than the unaffiliated. Finally, religious salience remains statistically insignificant.

Last but not least, no support is evident for Hypothesis 4, which predicted gender variations in the effects of religion on remarriage timing. The significant effects exhibited for denominational affiliation and attendance were roughly similar for men and women. Moreover, the essentially null effects for religious salience were similarly evident for men and women as well.

4. Conclusions and Discussion

Analyzing data from a national probability sample of young Americans, significant religious effects in the timing of remarriage were observed. As hypothesized, a pattern of accelerated recoupling is most evident among conservative Protestants as compared with Catholics and the unaffiliated for both women and men. As such, Catholicism can be characterized by a pattern of delayed remarriage; that is, institutional policies such as required annulments that are designed to discourage divorce and prevent remarriage seem to have their intended effect. Remarriage timing is also shorter for men affiliated with other faith traditions than the unaffiliated. The most consistent and robust finding is that increased frequency of worship service attendance reduced the timing of remarriage, suggesting that congregations are not only houses of worship but may be institutional springboards into remarriage. However, religious salience exerts weak and insignificant effects on remarriage timing. Given these results, it is concluded that the denominational subculture thesis and the religious capital thesis are supported by empirical evidence in remarriage timing but there is no evidence to support the spiritual capital thesis. Contrary to expectations, the observed religious effects do not appear to be systematically different for men and women.

The findings summarized above have a number of important implications and suggest several fruitful directions for future research. First, conservative Protestant readings of the Bible place a premium on God’s “hatred” of divorce (Malachi 2:16), which is likely to be linked to the strongly anti-divorce sentiments found among adherents of this denominational family (Wilson and Musick 1995). In addition, there are scriptural passages that directly discourage adherents from marrying a divorced person and equate remarriage to adultery (Matthew 5:32). But the pattern of accelerated remarriage timing among conservative Protestants revealed in this study suggests otherwise, which implies an adaptive and pragmatic strategy developed by conservative Protestants to negotiate with the prevalent culture of divorce in today’s America. This pattern could be evidence of “redemptive remarriage”—a divorcee’s reaffirmed commitment to marriage through remarriage following divorce—that is hinted at in the lone previous study on this topic (Brown and Porter 2013). For its part, the Catholic Church has long imposed institutional impediments to remarriage. However, given debates that surfaced at the 2014 Third Extraordinary General Assembly of the Synod of Bishops (also called the Synod on the Family) at the Vatican, greater tolerance toward divorced and remarried Catholics may be on the horizon.

Second, survey research typically provides no clues for motivations of remarriage. However, qualitative interviews with conservative Protestant, Catholic, and nonreligious persons who have remarried or are considering doing so would address the subjective motivations for remarriage more directly. This methodological triangulation can vastly improve the findings reported in the present study. Surely, such research is needed to determine if personal motivations or organizational processes exert encouragement, or even pressure, to remarry in the wake of a divorce.

Third, space and data limitations did not permit us to explore possible selection effects and causal direction issues. It is possible that persons who feel very ashamed of their divorced or remarried status may relinquish their religious affiliation or might be less inclined to attend worship services. The ways in which divorce and remarriage may produce changes in religious affiliation or attendance patterns are not examined in this study largely because the NSFG 2006–2010 data are pooled cross-sectional data. Consequently, no respondents are followed up over time. Future research on this front would be beneficial, given preliminary evidence that remarried Catholics are especially likely to leave the Catholic Church (Hout 2000).

Finally, additional research is needed to explore how the quality of remarriages might vary by faith tradition. It is quite possible that the cultural and structural approaches to remarriage that vary by denomination yield distinctions in remarriage quality across faith traditions. If data permit, examining remarriage quality by comparing the affiliation and attendance patterns of the remarried couple (e.g., same-faith versus interfaith remarriages) should be a high priority. There is prior evidence that marital and family outcomes differ significantly for religiously homogamous and heterogamous couples (Bartkowski et al. 2008; Curtis and Ellison 2002). Until such research is pursued, our study provides additional evidence that religion exerts a strong influence on remarriage timing among American men and women.