"زناشویى بسته" تاریخ زناشویی بین افراد مطلقه و هرگز ازدواج نکرده در میان سفیدپوستان غیراسپانیایی

Abstract Whether and why the never married and the divorced marry partners of like marital history is not well known. The homogamous tendency on marital history may simply be a by-product of couples' homogamous tendencies on age, socioeconomic status, and parenthood status, in addition to the group size imbalance between the never married and the divorced. Alternatively, marital history homogamy may reflect spousal preferences for similarity in marital history that arises, in part, from continued ties of the divorced to their former marriage. To test hypotheses implied by these two perspectives, regression models are applied to unmarried non-Hispanic white men and women from the Panel Study of Income Dynamics, 1985–1997. Results are contrary to the hypothesis based on the by-product perspective. Specifically, in both the male and female subsamples, the tendency toward marital history homogamy is not removed by controlling for individuals' spousal choice on age, education, and parenthood status. However, results are consistent with the hypothesis based on the marital ties perspective. For example, in the subsample of women, the measure of ties to former marriage accounts for about a quarter of the tendency toward marital history homogamy. Results have implications for resource distribution within and across American families.

1. Introduction The need to better understand processes underlying social and economic inequality has long motivated demographers and sociologists to investigate marital pairing patterns (e.g., Dreachsler, 1920; Kennedy, 1944). The tendency for individuals to marry their likes on a specific dimension that affects resource distribution, such as race (e.g., Qian, 1997) and education (e.g., Mare, 1991), implies that individuals select partners in ways that widen inequality in a society.1 Recent evidence suggests that marriages between persons of like marital history (i.e., first marriage and remarriage) are linked to inequality among children. More specifically, recent studies indicate that children living with parents in a remarriage receive lower parental investment than do children living with parents in a first marriage (e.g., McLanahan and Sandefur, 1994), even when differentials attributable to biological versus step relations and levels of parental resources are removed (e.g., Hofferth and Anderson, 2003). Understanding sources of differentials between the never married and the divorced that underlie marital history homogamy would provide insights into components that construct inequality across these marital units.2 Although published statistics suggest a tendency toward marital history homogamy among the never married and the divorced (Lugaila, 1998),3 little is known about the sources of this tendency. Homogamous tendency of the never married and the divorced on marital history can be viewed as: (a) the by-product of constrained mating opportunities across marital history or (b) the result of individuals' preferences to marry a partner of like marital history due to the distinction between the two groups (Kalmijn, 1998). A by-product perspective posits that mating opportunities are constrained across marital histories by: the tendency toward homogamy on dimensions other than marital history, such as age, socioeconomic status (SES, typically measured with education in studies of assortative mating), and parenthood status4 (e.g., Atkinson and Glass, 1985; Kalmijn, 1991 and Kalmijn, 1994; Mare, 1991; Qian, 1998; see Kalmijn, 1998 for review); and the imbalance in group sizes of the never married and the divorced (Blau et al., 1982). Put simply, these conditions reduce the likelihood that the never married and the divorced meet in the process of mating. This perspective implies the hypothesis that when individuals' homogamous–heterogamous selection of spouses on age, education, and parenthood status are controlled in addition to group size imbalances of the never married and the divorced in the local marriage market, a substantial tendency toward marital history homogamy should be absent. The second perspective posits that homogamous preferences form among the never married and the divorced because the groups are differentiated, in part, by the ties to former marriages among the divorced. A theory of martial ties suggests that, when couples have heavily socialized each other toward compatibility and have made major emotional and financial investments in the former marriage, they tend to maintain social, psychological, and financial ties to their former marriage even when the marriage had dissolved. These ties contribute to a homogamous tendency on marital history by raising conflicts and emotions that suppress the benefits of a marital pairing across marital history. Some evidence suggests that these ties are particularly prevalent among divorced women (e.g., Gerstel, 1988). Other conditions that could differentiate the divorced from the never married and form homogamous preferences on marital history include the stigma of divorce (Gerstel, 1988), the selection process into divorce (Bumpass and Sweet, 1972), and the impact of divorce on personality and attitudes (Wolfinger, 2003). In this study, I focus on investigating the potential contribution of the ties to the former marriage because this factor has direct implications for resource distribution within and across families, and hence inequality among coresident family members, including children. The marital ties perspective implies the hypothesis that, even when individuals' selection of spouses of like age, education, and parenthood status as well as relative group size are controlled, a substantial tendency toward marital history homogamy will persist, a portion of which is attributable to ties to the former marriage. I test the contrasting hypotheses implied by these two perspectives by applying competing risks discrete-time event history models, logistic regression models, and log-linear models to non-Hispanic whites from the Panel Study of Income Dynamics (PSID), 1985–1997. Using subsamples defined by sex, I first assess whether divorced individuals are statistically significantly more likely than never married individuals to marry someone who has been divorced, as opposed to never married (controlling for relative group size of the never married and the divorced in the local marriage market). I then test whether the tendency: (a) becomes statistically non-significant when controlling for homogamous or heterogamous spousal pairing on age, education, and parenthood status in addition to the relative group size of the never married and the divorced in the local area or (b) is reduced notably in magnitude by adding a measure of (divorced women's) ties to the former marriage to the model, which statistically significantly influences spousal pairing toward marital history homogamy. I treat fertility as an indicator of parenthood status, a characteristic shared by the never married and the divorced, rather than as an indicator of ties arising from joint investments in the former marriage (e.g., Becker et al., 1977), to obtain a lower bound estimate of the contribution of marital ties. If results show that the tendency toward homogamy on marital history is explained by homogamous mate selection on age, education, and parenthood status, and relative group size (but not by the ties to former marriages that extend into the post-divorce period), I can infer that as long as the divorced have similar socioeconomic or demographic characteristics as the never married, they select mates as if they had never been previously married. The results would imply that inequalities across first marriage and remarriage are not a direct product of spouses' marital history but a symptom of other conditions that constrain mating opportunities across marital history. If, however, results show that the tendency toward marital history homogamy is explained at least in part by (divorced women's) ties to a former marriage, then I can infer that the differentiation of the divorced from the never married characterized partly by these ties contribute to the formation of first marriages and remarriages. This evidence would imply that remarriages are distinct from first marriages in that, at least for some, they contain a distributional link to family members from the former marriage maintained by the divorced. This distinction would help explain inequality in parental investments among coresident children across first marriage and remarriage (e.g., McLanahan and Sandefur, 1994), especially the portion that is unexplained by parental economic status (e.g., Cooksey, 1997).5

. Conclusion Very little is known about whether and why divorced and never married persons tend to marry others with like marital histories. Understanding sources of the group differences that yield a homogamous tendency on marital history would enhance the understanding of if and how first marriages and remarriages are distinct in ways that produce inequality among coresident family members (e.g., Hofferth and Anderson, 2003; McLanahan and Sandefur, 1994). A by-product perspective suggests that marital history homogamy is primarily a function of constrained mating opportunities across marital history that arises from homogamous preferences on dimensions other than marital history (i.e., age, education, and parenthood status) as well as group size imbalances between the never married and the divorced. This perspective implies the hypothesis that when homogamy–heterogamy of individuals on the bases of age, education, and parenthood status in addition to relative group size are held constant, a major tendency toward marital history homogamy should be absent. Alternatively, a marital ties perspective suggests that ties originating in the former marriages of the divorced are one of the major factors that differentiate the divorced from the never married and produce preferences toward marital history homogamy. Some scholars suggest that marital investments are more likely to convert to post-divorce ties for women than for men for two reasons: because women's investments tend to extend into roles and positions that are compatible with, or necessitate, continuing ties (Gerstel, 1988; Manning et al., 2003); and because some divorcing men avoid the pain of residential separation from their children by severing ties (Braver and Griffin, 2000; Kruk, 1991). This perspective implies the two-part hypothesis that: (a) the tendency toward marital history homogamy should persist after holding constant homogamous–heterogamous selection of spouses on age, education, and parenthood status and (b) ties to the former marriage (among divorced women) should explain part of the tendency toward marital history homogamy. The results reported here support the two-part hypothesis implied by the marital ties perspective and are contrary to the hypothesis implied by the by-product perspective. First, I find that a tendency toward marital history homogamy persists when homogamy–heterogamy on age, education and parenthood status, and relative group size of the never married and the divorced are held constant in the logistic regression analysis. This finding is confirmed by the log-linear analysis. Second, results from competing risks event history models suggest that ties to the former marriage among divorced women, as measured by an index combining measures of marital investment and alimony payment relations, doubles the chance that divorced women marry a divorced man as opposed to a never married man. Results suggest that about a quarter of the tendency toward marital history homogamy is accounted for by ties to the former marriage. These results show that marital history homogamy is not simply a function of individuals' homogamous tendencies on age, education, and parenthood, and that ties to the former marriage is a major contributor to marital history homogamy. The tendency toward marital history homogamy and the influence of marital ties to the former marriage found in this study suggest that divorce has after-effects for many Americans. These after-effects, on average, extend beyond the divorce stage and into the phase where new marriages are formed. The ties may be linked to the slow institutionalization of remarriages (Cherlin, 1978) and are expected to be particularly notable among divorced women who have heavily invested in their former marriage. The results have three implications for resource allocation across family arrangements in the US. First, remarriages are distinct from first marriages in that, at least for some, they contain a distributional channel to family members from the former marriage maintained by the divorced. If ties to former marriages persist into remarriages, they may constrain the level or quality of resources available for distribution to coresiding family members (e.g., children) in a remarriage relative to those in a first marriage. The constraint is likely to contribute to inequality in parental investments among coresident children across parental first marriage and remarriage (e.g., McLanahan and Sandefur, 1994). This implication poses a challenge for policy makers. On the one hand, distributional ties to non-resident family members are important to improve the well-being of the non-resident children of the dissolved marriage. On the other hand, these ties may constrain the resources available to improve the well-being of children who coreside with remarried parents, as suggested by recent research (e.g., Hofferth and Anderson, 2003). Second, because sources of inequality are rooted in processes that characterize the institution of marriage (e.g., joint investments) in addition to the socioeconomic and demographic conditions of the time, inequality across first marriage and remarriage are more persistent than the by-product hypothesis would imply. A theory of marital ties implies that fundamental changes in the institution of marriage, such as a substantial decline in spouses' level of joint investment, would need to take place for the inequality across first marriage and remarriage to become minor. Third, three-quarters of the divorced remarry (Martin and Bumpass, 1989), and when they remarry persons of like marital history, they end up distributing a disproportionate part of their resources to other divorced (as opposed to non-divorced) persons who are resident and non-resident family members. This aspect of resource reallocation that accompanies divorce is not often noted in the post-divorce support debate, which is focused almost exclusively on resource allocation to non-resident family members (e.g., Garfinkel, 1994). The current study focuses on non-Hispanic whites. An extension of this study may investigate how the tendency toward marital history homogamy among non-Hispanic whites compares to that of African Americans. Greater resource constraints imposed by discrimination and segregation among African Americans (Corcoran and Adams, 1997; Reskin, 2001) may reduce the extent to which they can engage in post-marital socialization and major joint investments. As a result, the constraints would diminish the distinction between the never married and the divorced among African Americans, and hence suppress their tendency toward marital history homogamy. However, any weaker tendency toward marital history homogamy found among African Americans may be in part attributable to the underreporting of divorces among African Americans in survey data (Bumpass et al., 1991). A future analysis of racial difference in marital history homogamy might combine survey with non-survey data (such as marriage certificates from the Vital Statistics data) to augment data limitations. Such an analysis would help elaborate why marriage patterns differ between African Americans and non-Hispanic whites by highlighting differentials in marital investment patterns across the racial groups. The current study also focuses on marital unions rather than cohabiting unions. Partners in cohabiting unions may have less joint investment in the union than do partners in marital unions (see Smock, 2000 for review). Thus, pairing across marital history may be more prevalent in cohabiting unions than in marital unions. This hypothesis may be tested with appropriate data sets in future studies. Results in this study show that post-marital ties help distinguish the divorced from the never married in ways that contribute to their marital history homogamy. Through marital history homogamy, this distinction between the divorced and the never married is transferred to the marital units that they subsequently form, namely remarriages and first marriages, respectively. More investigation is needed to better understand the extent to which these distinctions affect resource allocation among American families.