Abstract
This article addresses a debate about the relationship of singlehood and informal ties—singlehood as isolating versus integrative—and evaluates structural explanations for this relationship, focusing on life course characteristics and socioeconomic resources. Using the National Survey of Families and Households (1992–1994) and the General Social Survey (2000, 2004, 2006, 2012), we examine ties to relatives, neighbors, and friends among U.S. adults. We find that single individuals are more likely to frequently stay in touch with, provide help to, and receive help from parents, siblings, neighbors, and friends than the married. These differences between the single and the married are more prominent for the never married than for the previously married, suggesting that marriage extends its reach after it ends. Being single increases the social connections of both women and men. Overall, much of the positive relationship between singlehood and social ties remains even when we take into account structural explanations. We conclude that instead of promoting marriage, policy should acknowledge the social constraints associated with marriage and recognize that single individuals have greater involvement with the broader community.
Keywords
Many Americans today lament the breakdown of community, and others emphasize the decline of marriage and the rise of singlehood, although few consider the connections between marital and community ties. Those social theorists who have considered the link between singlehood and broader social connections make competing claims about this relationship. These lay the ground for what we term the singlehood as isolating versus singlehood as integrative debate. On the one side, theorists view marriage as the primary building block of community, an institution fostering social integration, and suggest that those who are single are isolated, lonely, and have limited social ties. On the other side, scholars argue that marriage is privatizing, as it competes with, even undermines, other relationships, while being single promotes these relationships. Few scholars, however, empirically examine these claims or theorize the processes that may account for the link between singlehood and social ties. In this article, we use national representative U.S. data to assess this link as well as develop and test theoretical models to explain it.
We conceptualize social integration as multidimensional, including contact and help exchanges with friends and neighbors as well as with relatives, such as aging parents and siblings (Cornwell, Laumann, & Schumm, 2008; Putnam, 2000; Utz, Carr, Nesse, & Wortman, 2002). Although prior research examines the relationship of marital status to particular social ties, no studies have looked across these dimensions to consider the two sides of this debate.
Arguments made on each side of the debate suggest that the never married and the previously married might differ from the married in distinct ways; therefore, we separate these two groups of singles. Because numerous scholars have argued that social ties are shaped by gender (Dahlin, Kelly, & Moen, 2008; Liebler & Sandefur, 2002; Rossi & Rossi, 1990), we examine the relationship of singlehood to social ties separately for women and men.
After empirically establishing the links between singlehood and social involvement, we ask: How can we explain these links? Many sociologists adopt a structural perspective to explain social connections (Cornwell et al., 2008; Fischer, 1982; Moore, 1990; Putnam, 2000). Such an explanation would suggest that the issue is not singlehood itself but rather that the single and the married differ because of discrepant constraints and opportunities rooted in their distinctive life course positions as well as socioeconomic resources. Once we take into account these external conditions, the structural explanation suggests, singlehood itself is neutralized and will yield little independent effect on social integration.
The article includes three parts. We begin with a literature review discussing theories underpinning the two sides of the debate, prior research examining links between singlehood and social connections, and research on the structural correlates of marriage and singlehood associated with these connections. Next, we turn to empirical analyses of two national U.S. data sets—the General Social Survey (GSS, 2000, 2004, 2006, 2012) and the National Survey of Families and Households (NSFH, 1992–1994). Our analyses, first, examine the relationship of singlehood to social integration and, second, assess whether that relationship can be explained by differences in life course positions and resources. We conclude with a discussion of the theoretical and social implications of our findings.
Singlehood as isolating versus integrative
On one side, theorists beginning with Durkheim (1897/1966) have connected singlehood to social isolation as they viewed marriage as fostering attachment to wider communities and heightened investment in social ties (see Campbell & Lee, 1992). Slater (1963), in his classic piece on dyadic withdrawal, argued that marriage is the building block of community and that marital rituals provide an institutional solution to a couple’s natural inclination to withdraw. Rossi (2001) discussed the reduced “social embeddedness” of the single compared to the married. Fischer, Sollie, Sorell, and Green (1989) have argued that marriage strengthens social ties by providing opportunities and resources for relationships, and Wilcox (2005) suggested that the single receive less support than the married because others believe that singles are less responsible.
On the other side, scholars argue singlehood encourages other relationships, while modern marriage competes with, even undermines, them. Discussing the spread of “companionship” marriage in the early 20th century, Burgess and Locke (1945) emphasized its inward turn: Wives and husbands were expected to rely primarily on each other. Combining a structural and normative argument about that shift, Parsons and Bales (1960) posited that the modern economy demands such a marital relationship as the nuclear family “must stand on its own” (p. 10). They elaborated, “in particular, then, spouses are thrown upon each other, and their ties with members of their own family of orientation, notably parents and adult siblings, are correspondingly weakened” (p. 19). Bengtson (2001) argued that the spread of marital instability and lengthened lives result in longer periods that people spend being single, which in turn makes extended kin increasingly central in people’s lives. Feminists Barrett and McIntosh (1982) went further, suggesting marriage was a “trap” or a “prison”—an exclusive relationship that undermined other ties.
Whether asserting that singlehood is isolating or integrative, scholars tend to suggest that these processes differ for women and men. On one side, some integrationists argue that it is particularly men who become socially disengaged when single and that they are linked into communities by marriage because of its “civilizing power” (Nock, 1998, 2005; Popenoe, 1996). As Nock (2005, p. 20) writes, “Marriage integrates men into social groups of like-minded others and by doing so establishes acceptable boundaries around their behavior.” Researchers also argue that gender might shape the effects of marital status on social integration because of the conventional division of labor, although these arguments are contradictory. Some say marriage turns men (but not women) toward extra-domestic concerns that enhance social connections (Rotolo, 2000). Others agree that marriage is more integrative (and singlehood more isolating) for men than for women but suggest that wives as the families’ primary caregivers pull their husbands into caregiving and therefore enhance men’s social ties (Gerstel & Gallagher, 2001). Still others postulate a reversal of gender differences, suggesting that wives as primary kinkeepers are more integrated than husbands (Di Leonardo, 1987).
On the other side, many emphasize that women experience the integrative benefits of singlehood (and the isolating effects of marriage) more than men. Coser (1974) developed the concept of “greedy marriage” to characterize the lives of women who were housewives with little time for people besides their husbands and children. So, too, Barrett and McIntosh (1982), DePaulo (2006), and Trimberger (2006) were more concerned with the “entrapping” effects of marriage on women than men and viewed the liberating effects of remaining single as more pronounced for women. In contrast, some have argued that men are more constrained by marriage and hence obtain more benefits from singlehood than women. Single men may be under less financial pressure and work fewer hours, leaving them with more time to socialize, while husbands’ breadwinning role often demands long hours of paid work (Gerstel & Clawson, 2014; Moen & Sweet, 2006), which may come at the expense of social ties (Townsend, 2002).
Given these theoretical arguments, we might expect not only to find differences between those currently married and single but also between the never married and previously married. Three alternatives are possible. First, if singlehood is isolating, those never married may be least integrated because the previously married may maintain some (but not all) of the advantage provided by marriage. Second, if singlehood is integrative—and marriage, in contrast, involves a retreat from social ties—it may be difficult for the previously married to regain all ties they had as singles. Finally, the previously married may be the least integrated of all three groups: They may lose not only a spouse but also connections formed with that spouse without reclaiming ties they had while single. To empirically assess the theories, we compare each of the two groups of singles to the married.
Existing research on singlehood, marriage, and social integration
Empirical research on the question of whether it is being single or married that leads to social integration is limited. Most research examines only one aspect of integration, making no sustained argument about the relationship of singlehood and marriage to a range of connections. Furthermore, many studies use marital status as one of a series of controls, which makes findings difficult to compare and interpret. When they do examine marital status, most studies distinguish the single from the married without separating the previously married from the never married. Few compare women and men. Given the variation in conceptualization, measurement, and methods, it is not surprising that the findings are mixed.
For friends and neighbors, most research suggests that the single have more friends on average (Gillespie, Lever, Frederick, & Royce, 2014) and are more likely to socialize (Ha, 2008; Fischer, 1982; Kalmijn, 2003; Moore, 1990) and exchange help (Leibler & Sandefur, 2002) than the married. Putnam (2000), who expected to find that marriage would increase interactions with friends and neighbors, was surprised to discover the reverse and wrote, “Married people tend to be homebodies” (p. 278). The few studies that look at the relationship of friendship and marital status separately for women and men do not find gender differences (Fischer, 1982; Gillespie et al., 2014; Leibler & Sandefur, 2002; Moore, 1990).
Examining ties to parents, some research suggests that singles exchange more support with their parents than those married (Laditka & Laditka, 2001; Sarkisian & Gerstel, 2004, 2012; Suitor & Pillemer, 2006), although some find no such differences by marital status in support (Campbell & Martin-Matthews, 2003; Johnson & Lo Sasso, 2000) or contact (Sechrist, Suitor, Henderson, Cline, & Steinhour, 2007). Furthermore, a few studies suggest that marriage separates women from their parents more than it separates men (Ettner, 1996; Laditka & Laditka, 2001; Rossi & Rossi, 1990).
Looking at siblings, research finds that the married have less contact (Connidis & Campbell, 2001) and support than the single (Campbell, Connidis, & Davis, 1999; White & Reidmann, 1992). Examining the transitions, White (2001) finds that marital formation is associated with reduced sibling contact and receipt of help, whereas marital dissolution is associated with greater sibling contact. No research examines gender differences in the effects of marital status on sibling ties.
In sum, a fair amount of research, albeit not consistently, suggests that singlehood involves stronger ties to kin, neighbors, and friends. Little research examines gender differences, but it appears that marriage might reduce the ties to parents for women more than for men. Almost no research differentiates among the single and compares the married separately to the never married and previously married.
Structural explanations
Two structural dimensions may help explain differences in integration by marital status. The first suggests that marital status differences rest not on the institution of marriage itself but instead on other life course differences associated with it. The second explanation suggests that differences in integration are a result of the unequal resources of the single and the married.
Life course
Whereas the shift from singlehood to marriage itself is a life course transition and as such is the main focus of this study, it is also associated with other life course factors that might explain its relationship to social integration. Specifically, aging as well as the other life course transitions associated with it, such as parenthood, employment, and retirement, may help explain gaps in ties between those single and those married (Cornwell et al., 2008; Elder, Johnson, & Crosnoe, 2003; Rossi & Rossi, 1990; Shanahan, 2000). Evidence regarding the relationship of aging to social engagement is contradictory: Some find that aging intensifies (Cornwell et al., 2008; Miller & Nakamura, 1996), while others show that aging reduces social ties (Ajrouch, Blandon, & Antonucci, 2005), and still others report a quadratic relationship (Sarkisian & Gerstel, 2008). Although mixed findings make precise predictions difficult, age differences may account for some of the differences between the social ties of the single and the married.
Because the married are more likely to have children than singles, parenthood may also be a life course transition that accounts for what initially appear to be the effects of marital status. Children, needing time and affection, may cause parents to cut back on other relationships and activities. Some studies have suggested that the presence of children reduces the contact with (Rossi & Rossi, 1990) and assistance to (Pezzin & Schone, 1999; White & Peterson, 1995) aging parents and even the support received from them (Cooney & Uhlenberg, 1992). Several researchers have also found that parenthood leads to a reduction in the size of adults’ networks (Munch, McPherson, & Smith-Lovin, 1997). These effects might be gendered, with parenthood particularly restrictive of wives’ personal networks (Munch et al., 1997). But a number of scholars have found exactly the opposite. They show that children increase involvement of both mothers and fathers—whether with their aging parents (Cooney & Uhlenberg, 1992; Gerstel & Gallagher, 2001; White & Peterson, 1995), siblings (White, 2001), or neighbors (Moore, 1990). Given these mixed findings, parenthood could account for either the socially constraining or integrating effects of marriage.
Finally, employment is a life course transition that is often tied to marriage (Smock, Manning, & Porter, 2005) and may therefore serve as an explanation for its relationship to social integration, although the link might differ by gender (Townsend, 2002). Studies of paid work and kin ties are again mixed: Some research finds that employment is associated with less caregiving to kin (Sarkisian & Gerstel, 2004), others find the employed give more care to kin (Silverstein & Waite, 1993), while still others find no difference between those with and without jobs (Cooney & Uhlenberg, 1992; Roschelle, 1997). As Putnam (2000, p. 194) argues, “Getting a job outside the home has two opposing effects on community involvement—it increases opportunities for making new connections and getting involved, while at the same time it decreases time available for exploring those opportunities.” Thus, just like the differential rates of parenthood, differences in the rates of employment between the single and the married could potentially account for either the separating or connecting effects of singlehood and marriage.
Resources
A second set of structural explanations focuses on material conditions. The married tend to have higher household incomes and more education than the single (Waite & Gallagher, 2000), but it is unclear how these are related to social integration. Researchers disagree whether income and education increase (Cooney & Uhlenberg, 1992; Roschelle, 1997; Silverstein & Waite, 1993), decrease (Laditka & Laditka, 2001), or exhibit a quadratic relationship (Sarkisian & Gerstel, 2008) with kin integration. Research suggests that higher income and education are tied to more connections with friends and neighbors (Fischer, 1982; Moore, 1990).
In sum, although the findings are somewhat mixed, life course theory and the resources explanation suggest that, once we take into account these structural characteristics, the relationship of marital status to social integration will likely be at least attenuated.
Controls
In addition to these two main structural explanations, the relationship between singlehood and integration could be due to a number of compositional differences by marital status, especially in race/ethnicity (Lye, 1996) and health (Waite & Gallagher, 2001) as well as differences in kin composition and proximity (Cooney & Uhlenberg, 1992; Rossi & Rossi, 1990; Silverstein, Parrott, & Bengston, 1995). We use these demographic and kin characteristics as controls.
Research questions and hypotheses
We use two U.S. national data sets to address three research questions and test the hypotheses.
What are the differences in social integration between married individuals and those never married and previously married in the U.S.? We hypothesize that the married are, overall, less connected than the never and previously married. Further, we hypothesize that the differences will be particularly strong between the married and the never married but will also exist between the previously married and the married.
Does the relationship between marriage and social integration vary by gender? We expect few gender differences, with the exception of ties to parents, where we expect singlehood will be more integrative for women than men.
Can the relationship of marital status to social integration be explained by differences in life course position or socioeconomic resources? We hypothesize that the differences in social ties related to marital status can be explained, at least in part, by differences in these structural characteristics.
Method
Data
To empirically assess the debate, we use four waves of GSS data (2000, 2004, 2006, 2012; Davis, Smith, & Marsden, 2009) and the second wave of NSFH (1992–1994; Sweet & Bumpass, 1996). In each data set, our analyses focused on those respondents who were married, never married, or previously married but excluded those cohabiting because their smaller numbers precluded separate analyses.
First, we used GSS to evaluate socializing with friends and neighbors. We combined four waves in order to increase the sample size. These waves were selected because they included a question asking whether respondents were cohabiting with someone; that allowed us to avoid including cohabiters among the singles. GSS data are collected biennially from English-speaking U.S. adults aged 18 and older who are selected using stratified multistage area probability sampling to the block level; at the block level, quotas based on sex, age, and employment status are used. Our dependent variables are based on questions about socializing that, at each wave, were asked of a randomly selected subset of the total sample.
Second, we used NSFH to explore respondents’ contact and help transfers with parents, siblings, and friends and neighbors. Unfortunately, more recent data sources are not available to examine these outcomes for a nationally representative sample of adults. Other data sets that contain information on help transfers utilize age-restricted samples (e.g., the Health and Retirement Study and the third wave of NSFH) or do not include sufficient numbers of never married and previously married women and men for separate analyses (e.g., the second wave of the Midlife in the U.S. study). The second wave of NSFH collected data in 1992–1994 from 10,005 main respondents aged 23 and older who were selected during the first wave (1987–1988) using a stratified, multistage area probability sample of U.S. adults. Our analyses of ties to parents and siblings are restricted to those with at least one living parent (n = 6,745) or one living sibling (n = 9,200), respectively.
Dependent variables
Friends and neighbors (GSS, NSFH)
Our measures of contact with friends and neighbors are two dichotomies, indicating whether respondents socialized with their neighbors and friends at least several times a month, based on GSS questions asking respondents to report on a 7-point scale how often they (a) spend a social evening with someone who lives in their neighborhood and (b) spend a social evening with friends who live outside their neighborhood. Our measures of giving and receiving help to/from friends and neighbors are dichotomies based on NSFH questions asking whether, in the last month, the respondent has (a) given to and (b) received from their friends, neighbors, or coworkers any advice, encouragement, and moral or emotional support; help with shopping, errands, or transportation; help with housework, yard work, car repairs, or other work around the house; or help with child care.
Parents and siblings (NSFH)
Our measures of contact with parents and siblings are dichotomies, indicating whether, during the last 12 months, respondents saw their mother, father, and siblings at least once a week (based on questions that used a 6-point scale). To measure both the provision and the receipt of help, we generated four dichotomies, indicating whether respondents (a) gave help to parents, (b) received help from parents, (c) gave help to siblings, and (d) received help from siblings. These dichotomies relied on questions asking whether, in the last month, the respondent has (a) given to and (b) received from their parents and/or siblings any advice, encouragement, moral or emotional support, help with shopping, errands, or transportation, help with housework, yard work, car repairs, or other work around the house and/or help with childcare. For parents, two separate childcare questions were posed instead of the single question used for siblings: help with childcare for children aged 12 or younger while respondent (or wife/husband/partner) was working and help with childcare for any minor children at times other than when respondent (or wife/husband/partner) was working. These 2 items were only asked about help received from parents but not help given to parents.
Independent variables
Marital status
Respondent’s marital status in each data set is represented by two dichotomies, never married and previously married (including divorced, separated, and widowed), with married as the omitted category. Single respondents who were currently cohabiting were excluded from our analyses, as there were not enough cases to examine them separately. We did not consider it appropriate to combine them with either the married or single individuals because a fair amount of research suggests that marriage and cohabitation differ in numerous ways (Eggebeen, 2005; Nock, 1995). In the NSFH, cohabitation status was ascertained for all respondents using the household roster. In GSS, only half of the sample (randomly selected) responded to a question about living with a partner in 2004 and 2006, while everyone was asked this question in 2000 and 2012. We used multiple imputation to determine cohabitation status for those singles who did not answer that question, and based on that imputation, individuals were excluded from the sample (as a result, sample size varied slightly across imputed data sets).
Life course
In both data sets, we included respondent’s age (mean centered), as well as age squared (divided by 100 to simplify presentation), to control for potential quadratic relationships. We also include a parental status dichotomy indicating whether the respondent has minor children. In NSFH, that variable reflects whether the respondent lives with his/her own or spouse’s minor children (biological, adopted, or foster and under 19 years of age). In GSS, this variable indicates whether any children under 18 are in the household. Finally, our measure of employment status is a dichotomy indicating whether the respondent was currently employed.
Resources
To measure socioeconomic resources, we use a household income variable divided by the household size, then logged to improve the distributional properties, mean centered and used in the models along with its squared term to allow for quadratic relationships. In NSFH, the household income variable, constructed by NSFH staff, indicates the combined income of all household members in the past year. In GSS, it indicates total family income (using categories) from all sources in the past year, and we recoded categories to the midpoint of each interval. Our education variable indicates the number of years of completed education, bottom coded at 7 (i.e., values 0–6 recoded to 7 to reduce the impact of outliers). In both data sets, this variable was constructed by the survey staff on the basis of respondents’ reports of their degrees and schooling experiences.
Controls
For respondents’ race/ethnicity, we use three variables: Black/African American, Latino/a, and Other minority (White is the omitted category). We control for respondents’ health using a dichotomy indicating good or excellent self-reported health. For the analysis of parents and siblings using NSFH, we included four variables to control for the characteristics of parents and siblings: (a) an indicator of whether respondents had both of their parents still living, (b) the number of respondents’ siblings (full, half, and step), log transformed to improve its distributional properties; and (c) two dichotomies indicating whether respondents live with or within 25 miles of any of their parents and siblings.
Analytic strategy
Our analysis, conducted using Stata 13, consists of two parts. First, our bivariate analyses compared the never and previously married to the married in terms of their involvement in the three sets of ties (parents, siblings, friends and neighbors), separately for women and men. Second, we constructed logit models (separate for women and men) for each of the outcomes in order to examine whether life course characteristics, resources, and controls explained the differences by marital status.
To obtain generalizable results, bivariate analyses used weights adjusting for oversampling, nonresponse, and attrition as well as aligning selected demographic characteristics of the sample to those in the population. All analyses for NSFH adjusted for the clustered and stratified sample. To handle missing data problems, in NSFH, for income and education, we relied on the constructed variables generated and imputed by NSFH staff. For consistency, we used single regression imputation based on customized multivariate models for the remaining independent variables (such imputations were applied to less than 1% of the total number of data points). In the GSS, we performed multiple imputation by chained equations (Royston & White, 2011), subsequently deleting the imputed values of the dependent variables (von Hippel, 2007). Our estimates were based on 20 imputed data sets and incorporated the uncertainty that resulted from imputation into the standard errors.
Findings
To begin, we use bivariate analyses to examine whether the never and previously married differ from the married. Table 1 indicates that most of the differences are in the expected direction. Starting with friends and neighbors, we find that never married women and men are more likely to socialize, give, and receive help than the married. Previously married men are more involved with friends and neighbors than married men, although there is no difference in giving help, while previously married women are more likely to socialize with neighbors but less likely to provide help than the married.
Social integration by marital status.
Note. GSS = General Social Survey; NSFH = National Survey of Families and Households. For GSS data, marital status groups varied in size slightly across imputations, and the lowest sample size for each group is reported.
*p < .05; **p < .01; ***p < .001 (two-tailed): Significant differences (comparison group = married).
Singles are significantly more likely than the married to frequently see their parents and siblings and to give help to and receive help from parents. So, too, never married women and men are significantly more likely than those married to give help to and receive help from their siblings. Previously married men and women, however, are less likely than the married to give help to their siblings and equally likely to receive help from them.
Thus, answering our first two questions, we find that singlehood is associated with more social ties for both women and men, but especially for men. As we expected, the marital status gaps are consistently larger for the never married than for the previously married.
Evaluating structural explanations
We hypothesized that two sets of structural factors—life course characteristics and resources—might explain the link between marital status and social integration. Table 2 presents bivariate results for these structural variables as well as controls, demonstrating that the single and the married differ in their structural characteristics. On average, the married are older than the never married but younger than the previously married. The married tend to have more minor children and higher incomes, and are less likely to be Black, than both groups of singles. Previously married women and men tend to have lower employment rates and less education than the married in both data sets, while for the never married, the results differ by data set. The married, especially women, tend to have better health than the previously married, but they are similar to the never married. The never married are more likely and the previously married are less likely than the married to have both parents still living. Finally, the married are also less likely to live in close proximity to their parents and siblings than the never married.
Independent variables by marital status.
Note. GSS = General Social Survey; NSFH = National Survey of Families and Households. For GSS data, marital status groups varied in size slightly across imputations, and the lowest sample size for each group is reported.
*p < .05; **p < .01; ***p < .001 (two-tailed): Significant differences (comparison group = married).
To test whether these differences in structural factors and controls can explain the link between marital status and integration, we turn to multivariate analyses (Tables 3 to 5). The first two rows of coefficients in each table represent the odds ratios for marital status without any controls (except for year dummies in GSS). The next two rows provide comparisons by marital status using the full models. By comparing these two sets of coefficients, we can determine whether life course, resources, and controls explain the relationship of marital status to integration.
Logistic regression models for involvement with friends and neighbors: Odds ratios.
Note. GSS models also included year dummies as controls.
*p < .05; **p < .01; ***p < .001 (two-tailed): Significant odds ratios.
Logistic regression models for involvement with parents: Odds ratios.
*p < .05; **p < .01; ***p < .001 (two-tailed tests): Significant odds ratios.
Logistic regression models for involvement with siblings: Odds ratios.
*p < .05; **p < .01; ***p < .001 (two-tailed tests): Significant odds ratios.
Presenting explanatory models for ties to friends and neighbors, Table 3 shows that in the full model, both groups of singles are consistently more likely than the married to socialize and give and receive help. Some of the effects have strengthened with controls, and those comparisons that indicated differences in reverse or no significant differences in the bivariate model (socializing with friends and giving and receiving help for previously married women and giving help for previously married men) now demonstrate a singlehood advantage as well. Overall, after taking into account the structural explanations, we see that singlehood is integrative with regard to friends and neighbors.
Turning to parents in Table 4, we find that even with life course characteristics, resources, and controls in the models, the rates of contact, help given, and help received remain significantly higher among the never and previously married than among those married for both women and men. The only exception is the difference between married and previously married women in frequent contact, which is no longer significant (it was mostly due to closer proximity to parents and siblings among the previously married than the married). Overall, the structural explanations do not account for most marriage gaps in relations with parents.
Table 5 turns to siblings and shows that most marital status differences persist even when we take life course characteristics, resources, and controls into account. In fact, the power of marriage as a deterrent of sibling ties becomes even clearer. Whereas in the bivariate analyses, we found no differences between the previously married and the married in terms of receiving help from siblings, once we control for structural factors, especially education and age, the previously married are significantly more likely to receive help than the married. Moreover, previously married women’s and men’s lower rates of giving help—the only comparison in sibling ties that favored the married in the bivariate analyses—are explained by the structural factors, primarily by married individuals’ higher levels of education and younger age which boost their rates of giving help. Overall, the positive links between singlehood and sibling ties persist and, in some cases, are even strengthened, when we take into account structural explanations.
In sum, our multivariate analyses provide little support for structural explanations of the link between marital status and integration. Even when we take structural differences into account, singlehood continues to promote—and marriage continues to restrain—many connections.
Discussion
Many believe that singlehood brings loneliness and isolation because marriage is the fundamental building block of society, an institution that broadens social ties and ensures community. Others argue that singlehood bolsters, while marriage diminishes social involvement. To assess these contrasting perspectives, we compared empirically the social integration of those single and married.
Singlehood, marriage, and social integration
Supporting the view that singlehood increases integration, we find that being single is associated with greater social involvement. Compared to those married, the single are more likely to contact and receive help from their parents or siblings. Single people are more likely than those married to give help to, get help from, or socialize with neighbors or friends. The impact of marriage, as we expected, extends to those whose marriages end: The differences between the never married and married are larger than the differences between the previously married and married.
We also hypothesized no gender gaps in the effects of marital status, with one exception: We expected that marriage would diminish women’s ties to parents more than men’s. Our analysis shows, however, that marriage separates both women and men, and in fact, some reductions in ties are greater for husbands than for wives. The reasons marital status affects women’s and men’s ties may differ. Given theories about women’s greater emphasis on social support and caregiving (Cancian & Oliker, 2000; Sarkisian & Gerstel, 2012) and men’s greater marital power (Connell, 1987; Pyke, 1996), we might speculate that the diminution of wives’ ties results more from their husbands’ needs and desires than from their own. Future research should explore the role of such preferences and marital power.
Explanations
Overall, how do we explain why singles have more social ties than the married? We found little support for structural explanations as they did not fully account for any of the integrating effects of singlehood. Overall, even when the married, never married, and previously married are at the same point in their life course and have equal resources, the single still have more connections to family, neighbors, and friends. Two other theoretical perspectives—universal and cultural (Johnson & Leslie, 1982)—could provide alternative explanations for the links between marital status and social integration.
Universal explanations suggest that the link between marriage and community ties is invariant, as it is based in the very nature of human coupling. Such explanations would, for example, suggest that the married quite naturally become soul mates, drawn to and focused on each other in such a way that excludes personal ties to others. It would suggest the married are naturally and necessarily homebodies. In contrast, single individuals would naturally feel lonely if they stayed at home, so they get out more, perhaps trying to find a marital partner along the way.
Alternatively, cultural explanations suggest that the relationship of marital status to social ties reaches beyond its structural correlates, but are far from universal: They vary across time and place according to the norms and expectations surrounding marriage. For example, the greediness of marriage could be linked to the cultural expectation of romantic love (Goode, 1959) and conjugal self-sufficiency (Bellah, Madsen, Sullivan, Swidler, & Tipton, 1985). That is, Americans believe couples should be able to make it on their own—both practically and emotionally, and spouses are expected to rely on each other for their day-to-day needs and to act as each other’s soul mates and confidants (Cherlin, 2009; McPherson, Smith-Lovin, & Brashears, 2006) and, as a result, may turn less to kin, friends, and neighbors than their single counterparts. Such disengagement would be far less common in those “traditional” societies, where extended kinship is the primary source of support and marriage is primarily a linkage forged between groups rather than between two individuals who marry (Levi-Strauss, 1969).
Like much of the theoretical and empirical work in this area, we focus on the U.S. in our analysis, although we would like to emphasize that it is very likely that different societies (and even different groups within the same society) may organize marriage in different ways, with distinct consequences for social integration.
Limitations and future directions
This study has some limitations that suggest directions for further work in this area. First, whereas we cannot empirically assess universal and cultural explanations with the data used here, future research could profitably utilize cross-national data to test them. An examination of group differences within the U.S. could also be helpful in that regard: For example, a recent article by Campos, Rojas Perez, and Guardino (2014) found that extended familism is linked to higher quality romantic relationships in a Latino/a sample but not among Whites or Asian Americans. Likewise, singlehood may have different implications for individuals of diverse racial/ethnic and social class backgrounds.
Second, due to data availability constraints, some of our analyses relied on data from the early 1990s (NSFH II data), but the role and nature of marriage can shift and evolve over time. Future research should collect and utilize similar data for the current decade to evaluate whether the integrative effects of singlehood persist. Moreover, while we controlled for age in this study, future research should carefully differentiate the influence of age, period, and cohort, and singlehood may have different effects depending on individuals’ age and on specific time period, and there can be some cohort shifts in such effects as well. Future research should utilize longitudinal data to explore and disentangle these effects.
Third, our study controls for proximity to parents and siblings when examining the effects of marriage and singlehood on these ties; however, we had no measures of proximity to friends. Future research should examine such proximity (or residential mobility as its proxy) to test whether it is responsible for some of the observed differences between the single and the married. Moreover, while we focused on individuals, singlehood and marriage can also create patterns within the entire networks—that is, not only one’s individual marital status may matter but also the marital status of one’s friends and relatives. Future research could examine such effects by focusing on friendship dyads, examining networks (including their density and homogamy by marital status), as well as by using neighborhood composition measures. Moreover, it would be important to conduct such analyses separately by age because network and neighborhood composition by marital status likely shift over the life course.
Fourth, future research should also compare both the married and cohabiters to singles who live alone to distinguish the effects of the institutional properties of marriage from those of partnership per se, as some scholars (e.g., Slater, 1963) have argued that these can have opposite effects on social ties, while others (Kalmijn, 2003) have suggested that both those married and cohabiting have fewer social connections than singles.
Fifth, even though we examine multiple dimensions of social integration, our analysis focuses on informal kin and community ties. Future research should compare whether singlehood and marriage have similar effects on more formal aspects of social integration, as well as explore more intangible aspects, such as perceived sense of community and engagement as well as the meanings that people attach to their social ties.
Finally, some might argue that our findings result from self-selection. It is possible that those lacking strong ties are especially likely to marry: For them, marriage might serve as a substitute for other relationships, while those who already have strong ties may have less need to marry because these ties provide much of what they need. It is also possible that those deeply immersed in social ties are less likely to get married because, to prevent the weakening of their relationship, those close to them may discourage them from getting married. Thus, the differences between the single and the married may result both from self-selection into marriage and from the nature of marriage itself. Future research should use longitudinal data to examine these processes.
Conclusion
Social commentators frequently assail two types of disengagement: Some mourn the lack of community involvement, and others lament the spread of singlehood and the lack of commitment to marriage. At times, they connect the two concerns, arguing that marriage serves as a gateway to other social ties and therefore singlehood and marital instability are responsible for disengagement. In contrast, we find that marriage constrains community ties.
To be sure, we should neither romanticize nor underestimate the value of such ties. Putnam (2000) noted that there can be a “dark side of social capital,” including gangs and criminal networks. The reduction of relationships outside marriage may mean a withdrawal of the married, especially men, from dangerous networks. Indeed, as Sampson, Laub, and Wimer (2006) find, marriage reduces criminal activities and incarceration, at least for White men. Thus, the isolating effects of marriage may be, in some ways, beneficial.
There are also negative consequences. Marriage may deprive aging siblings and elderly parents of some of the help they want and need. Marriage can also generate excessive burdens on those who are single, as they may be expected to provide the care that their married siblings do not (Laditka & Laditka, 2001). Reduced ties may even put a strain on marriage itself. Even Parsons who argued that the isolation of the couple was functionally necessary recognized that such isolation could leave the marital pair in a “structurally unsupported” situation: “Neither party has any other adult kin on whom they have a right to ‘lean for support’ in a sense closely comparable to the position of the spouse” (Parsons & Bales, 1960, p. 20). As marriage becomes the only place where individuals look for and find support and comfort, marital unions may become more fragile.
We do not want to overstate: A growing proportion of Americans think marriage is becoming obsolete, and almost half of those not married say they have no plans to marry (Pew Research Center, 2010). Nevertheless, there remains significant ideological support for marriage. Such support is further aided by the state as it invests in marriage promotion programs (Avishai, Heath, & Randles, 2012). Similarly, relationship researchers tend to emphasize, if only implicitly, the favorable consequences of marriage, whether in their support of gay and lesbian marriage or their emphasis on the negative consequences of marital separation and divorce. To be sure, some sociologists did document the favorable effects of marriage for adults and children—in terms of economic well-being, physical health, and mental health (although others show that those benefits at least partially result from self-selection into marriage and the married individuals’ greater access to resources). It is important to acknowledge that most Americans—never married, previously married and married—do maintain some kinship and community ties. Nevertheless, the overwhelming emphasis on the benefits of marriage and the disadvantages of singlehood, we argue, is based on an inadequate understanding and analysis of negative implications of marriage and the positive contributions of singlehood. Although such understanding was a key part of the second-wave feminist agenda, it has fallen out of fashion. More recent discussions overwhelmingly tend to overlook the ways that single individuals contribute to social welfare more than the married as well as the costs that marriage imposes.
Based on our findings, we concur with Cherlin (2009) who warns against a family policy that relies too heavily on marriage, suggesting policy makers should craft a balance of marriage-focused and marriage-neutral programs to support children. Our analysis suggests a more balanced policy would also help change expectations associated with marriage, encourage community-friendly families and partnerships, and support those who are single. After all, single individuals, as we show, are in many ways the lifeline of the community that social policy should seek to sustain.
Footnotes
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.
References
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
