SOCIAL SCIENCE RESEARCH 27,329-349 (1998) ARTICLE NO.S0980625 Windows on Divorce:Before and after Thomas L.Hanson University of California-Riverside Sara S.McLanahan Princeton University and Elizabeth Thomson University of Wisconsin-Madison This paper uses data from the National Survey of Families and Households to examine how parental divorce is related to three important types of family resources:economic resources,parental resources (socialization practices),and community resources.We pay special attention to the dynamic character of family resources,examining how resources change as parents move closer to divorce and whether the consequences of divorce for family resources are short term or long term.Consistent with other work in this area,we find that parental divorce has severe negative consequences for the economic well-being of mothers and children.Most of our results for parental resources suggest that marital disruption results in declines in effective parental practices,at least in the short term.Our results for community resources provide some evidence that parental divorce results in gains in extra-household sources of support-although in some areas,community re- sources decline after divorce as well.1998 Academic Press INTRODUCTION Divorce is both an event and a process.Many of the changes we associate with divorce can be observed well before and long after a couple lives apart or formally An earlier version of this paper was presented at the annual meeting of the Population Association of America,held May 9-11,1996,in New Orleans,LA.The research was supported by Grants HD29601 and HD19375 from the Center for Population Research,National Institute of Child Health and Human Development.The National Survey of Families and Households was funded by NIH Grant HD21009. Address correspondence and reprint requests to Thomas Hanson,Healthy Families Project-142, University of California-Riverside,Riverside,CA 92521.E-mail:tom.hanson@ucr.edu. 329 0049-089X98$25.00 Copyright 1998 by Academie Press All rights of reproduction in any form reserved
Windows on Divorce: Before and after Thomas L. Hanson University of California–Riverside Sara S. McLanahan Princeton University and Elizabeth Thomson University of Wisconsin—Madison This paper uses data from the National Survey of Families and Households to examine how parental divorce is related to three important types of family resources: economic resources, parental resources (socialization practices), and community resources. We pay special attention to the dynamic character of family resources, examining how resources change as parents move closer to divorce and whether the consequences of divorce for family resources are short term or long term. Consistent with other work in this area, we find that parental divorce has severe negative consequences for the economic well-being of mothers and children. Most of our results for parental resources suggest that marital disruption results in declines in effective parental practices, at least in the short term. Our results for community resources provide some evidence that parental divorce results in gains in extra-household sources of support—although in some areas, community resources decline after divorce as well. r 1998 Academic Press INTRODUCTION Divorce is both an event and a process. Many of the changes we associate with divorce can be observed well before and long after a couple lives apart or formally An earlier version of this paper was presented at the annual meeting of the Population Association of America, held May 9–11, 1996, in New Orleans, LA. The research was supported by Grants HD29601 and HD19375 from the Center for Population Research, National Institute of Child Health and Human Development. The National Survey of Families and Households was funded by NIH Grant HD21009. Address correspondence and reprint requests to Thomas Hanson, Healthy Families Project-142, University of California–Riverside, Riverside, CA 92521. E-mail: tom.hanson@ucr.edu. SOCIAL SCIENCE RESEARCH 27, 329–349 (1998) ARTICLE NO. S0980625 329 0049-089X/98 $25.00 Copyright r 1998 by Academic Press All rights of reproduction in any form reserved. SSR625 @xyserv1/disk4/CLS_jrnlkz/GRP_ssrj/JOB_ssrj27-3/DIV_231a04 debb No. of Pages—21 First page no.—329 Last page no.—349
330 HANSON,MCLANAHAN,AND THOMSON terminates their marriage.Identifying the timing of changes associated with divorce is particularly important for understanding how divorce influences the resources available to children and subsequently influences child welfare. We know that children whose parents divorcel are disadvantaged in a variety of ways as compared to children who grow up with both parents(see Amato and Keith,1991;Furstenberg and Cherlin,1991;McLanahan and Sandefur,1994). These findings have been replicated with different data sets,for children of different ages,and for children from different social class backgrounds and different race and ethnic groups(McLanahan and Sandefur,1994).We also know that divorce is associated with reduced family resources-economic,parental, and community-and that differences in resources can account for some of the associations between divorce and child welfare(McLanahan and Sandefur,1994). What is less clear is when and how differences in family resources emerge in the process of divorce.Are some differences clearly evident well before a marriage is in trouble?That is,are the resource differences between divorcing and intact families due in part to selection into divorce of families with fewer resources?Or do resource differences appear as the result of marital or family problems that eventually result in divorce?That is,do the true "effects"of divorce emerge during the process of marital separation rather than after the final event?Or is it the event of divorce itself that creates losses in economic,parental, or community resources for children?Finally,are resource losses associated with divorce recouped over time,particularly if a mother remarries or cohabits?The answers to these questions are necessary to understand how children are likely to fare in the process of their parents'divorce. In this article,we use data from the 1987-1988 and 1992-1994 waves of the National Survey of Families and Households to examine how divorce and family resources are related over time.We examine whether children experience signifi- cant reductions in family resources subsequent to divorce;whether resource losses,when they exist,precede the actual event of marital disruption;and whether divorce is associated with short-term or long-term losses of family resources,contingent on the mother's remarriage or repartnering. DIVORCE AND FAMILY RESOURCES Three types of resources have been identified as especially important in the divorce process:economic resources,parental resources,and community re- sources.Economic resources are the material resources available to the house- hold,including income and assets.A good deal of research has demonstrated that economic resources are positively related to child well-being(Elder,Van Nguyen, and Caspi,1985;McLoyd,1990;Huston,1991;Miller and Korenman,1993; Duncan and Brooks-Gunn,1997).Families with more income can purchase more goods and materials that enhance healthy child development.Economic resources 】Throughout the paper,.we use the terms“divorce,”“marital dissolution,”and“marital separa- tion”synonymously
terminates their marriage. Identifying the timing of changes associated with divorce is particularly important for understanding how divorce influences the resources available to children and subsequently influences child welfare. We know that children whose parents divorce1 are disadvantaged in a variety of ways as compared to children who grow up with both parents (see Amato and Keith, 1991; Furstenberg and Cherlin, 1991; McLanahan and Sandefur, 1994). These findings have been replicated with different data sets, for children of different ages, and for children from different social class backgrounds and different race and ethnic groups (McLanahan and Sandefur, 1994). We also know that divorce is associated with reduced family resources—economic, parental, and community—and that differences in resources can account for some of the associations between divorce and child welfare (McLanahan and Sandefur, 1994). What is less clear is when and how differences in family resources emerge in the process of divorce. Are some differences clearly evident well before a marriage is in trouble? That is, are the resource differences between divorcing and intact families due in part to selection into divorce of families with fewer resources? Or do resource differences appear as the result of marital or family problems that eventually result in divorce? That is, do the true ‘‘effects’’ of divorce emerge during the process of marital separation rather than after the final event? Or is it the event of divorce itself that creates losses in economic, parental, or community resources for children? Finally, are resource losses associated with divorce recouped over time, particularly if a mother remarries or cohabits? The answers to these questions are necessary to understand how children are likely to fare in the process of their parents’ divorce. In this article, we use data from the 1987–1988 and 1992–1994 waves of the National Survey of Families and Households to examine how divorce and family resources are related over time. We examine whether children experience signifi- cant reductions in family resources subsequent to divorce; whether resource losses, when they exist, precede the actual event of marital disruption; and whether divorce is associated with short-term or long-term losses of family resources, contingent on the mother’s remarriage or repartnering. DIVORCE AND FAMILY RESOURCES Three types of resources have been identified as especially important in the divorce process: economic resources, parental resources, and community resources. Economic resources are the material resources available to the household, including income and assets. A good deal of research has demonstrated that economic resources are positively related to child well-being (Elder, Van Nguyen, and Caspi, 1985; McLoyd, 1990; Huston, 1991; Miller and Korenman, 1993; Duncan and Brooks-Gunn, 1997). Families with more income can purchase more goods and materials that enhance healthy child development. Economic resources 1 Throughout the paper, we use the terms ‘‘divorce,’’ ‘‘marital dissolution,’’ and ‘‘marital separation’’ synonymously. 330 HANSON, MCLANAHAN, AND THOMSON SSR625 @xyserv1/disk4/CLS_jrnlkz/GRP_ssrj/JOB_ssrj27-3/DIV_231a04 debb
WINDOWS ON DIVORCE 331 can enhance parental and community resources as well.Economic stress can interfere with effective parenting practices(i.e.,McLeod and Shanahan,1993), and economic resources provide access to such community resources as safe neighborhoods and good schools. By parental resources,we mean high degrees of parental involvement,consis- tent parental supervision and discipline,and high and clearly stated expectations. A high degree of parental involvement provides directed learning opportunities for children and can also provide an important source of security and trust(Bradly et al.,1994).Consistent parental supervision and discipline can be effective deterrents of undesirable behavior (Patterson,1982).High and clearly stated parental expectations can provide a sense of direction and motivation to children. All of these types of parental practices have been found to be positively related to child well-being(Baumrind,1966;Patterson,1982;Maccoby and Martin,1983; Loeber and Stouthamer-Loeber,1986).2 The third and final dimension of family resources considered is community resources,including ties to kin,friends,and formal organizations.Community and kin contacts can provide social support as well as access to information and opportunities(Coleman,1988).Social integration can also benefit children by benefiting parents,enhancing their psychological well-being and providing sup- port for effective parenting(Bronfenbrenner,1986). Each type of resource is a potential source of marital stability;could be adversely affected by marital strains;could change with the event of divorce itself,and might continue to change with time after divorce or with remarriage or repartnering.Evidence that the process of divorce is related to the timing of resource changes is relatively meager,so we rely primarily on theoretical arguments in the following discussion. We know the most about the timing of divorce and reduced economic resources.Economic deprivation and declines in income do lead to divorce (Becker,Landes,and Michael,1977;U.S.Bureau of the Census,1991a;Mott and Moore,1979;Ross and Sawhill,1975).Divorce also has severe negative conse- quences for the economic well-being of custodial mothers and children (U.S. Bureau of the Census,1991a;Duncan and Hoffman,1985;Holden and Smock, 1991;U.S.Bureau of the Census,1991a),in part due to loss of economies of scale,but mostly because noncustodial fathers pay little or no child support(U.S. Bureau of the Census,1991b).The economic losses experienced by custodial mothers and children after divorce appear to be relatively long term,unless the 2 Of course,parental practices are not the only type of parental resources.A second type of parental resource that is not considered here is how well parents get along with each other-or the reverse of that:parental conflict.Parental conflict is a negative parental resource,and a long line of research has demonstrated that parental conflict is negatively related to child well-being (e.g,Cummings and Davies,1994;Emery,1988;Grych and Fincham,1990;Hanson,forthcoming).We do not examine parental conflict in this analysis.Rather,we focus exclusively on those types of parental resources bearing directly on parent-child interaction
can enhance parental and community resources as well. Economic stress can interfere with effective parenting practices (i.e., McLeod and Shanahan, 1993), and economic resources provide access to such community resources as safe neighborhoods and good schools. By parental resources, we mean high degrees of parental involvement, consistent parental supervision and discipline, and high and clearly stated expectations. A high degree of parental involvement provides directed learning opportunities for children and can also provide an important source of security and trust (Bradly et al., 1994). Consistent parental supervision and discipline can be effective deterrents of undesirable behavior (Patterson, 1982). High and clearly stated parental expectations can provide a sense of direction and motivation to children. All of these types of parental practices have been found to be positively related to child well-being (Baumrind, 1966; Patterson, 1982; Maccoby and Martin, 1983; Loeber and Stouthamer-Loeber, 1986).2 The third and final dimension of family resources considered is community resources, including ties to kin, friends, and formal organizations. Community and kin contacts can provide social support as well as access to information and opportunities (Coleman, 1988). Social integration can also benefit children by benefiting parents, enhancing their psychological well-being and providing support for effective parenting (Bronfenbrenner, 1986). Each type of resource is a potential source of marital stability; could be adversely affected by marital strains; could change with the event of divorce itself; and might continue to change with time after divorce or with remarriage or repartnering. Evidence that the process of divorce is related to the timing of resource changes is relatively meager, so we rely primarily on theoretical arguments in the following discussion. We know the most about the timing of divorce and reduced economic resources. Economic deprivation and declines in income do lead to divorce (Becker, Landes, and Michael, 1977; U.S. Bureau of the Census, 1991a; Mott and Moore, 1979; Ross and Sawhill, 1975). Divorce also has severe negative consequences for the economic well-being of custodial mothers and children (U.S. Bureau of the Census, 1991a; Duncan and Hoffman, 1985; Holden and Smock, 1991; U.S. Bureau of the Census, 1991a), in part due to loss of economies of scale, but mostly because noncustodial fathers pay little or no child support (U.S. Bureau of the Census, 1991b). The economic losses experienced by custodial mothers and children after divorce appear to be relatively long term, unless the 2 Of course, parental practices are not the only type of parental resources. A second type of parental resource that is not considered here is how well parents get along with each other—or the reverse of that: parental conflict. Parental conflict is a negative parental resource, and a long line of research has demonstrated that parental conflict is negatively related to child well-being (e.g., Cummings and Davies, 1994; Emery, 1988; Grych and Fincham, 1990; Hanson, forthcoming). We do not examine parental conflict in this analysis. Rather, we focus exclusively on those types of parental resources bearing directly on parent–child interaction. WINDOWS ON DIVORCE 331 SSR625 @xyserv1/disk4/CLS_jrnlkz/GRP_ssrj/JOB_ssrj27-3/DIV_231a04 debb
332 HANSON,MCLANAHAN,AND THOMSON mother remarries.Even with remarriage,it is not clear that children receive the full economic benefit of a stepfather's additional income or assets. Parenting resources have been investigated for the most part after divorce.We don't have evidence for the selection of less effective parents into divorce,but we can theorize that people who have difficulty in sustaining long-term emotional relationships with another adult may also be less effective in their relationships with children.More likely is that parenting difficulties observed prior to a divorce stem directly from the stress of a troubled marital relationship,particularly if the parents are no longer supportive of each other's parenting practices(Bronfenbren- ner,1986).The most obvious direct effect of divorce itself is,of course,the sharp reduction in contact between children and noncustodial parents(Furstenberg and Nord,1985;Seltzer,1991).Divorce may also interrupt bonds between custodial parents and children by preoccupying parents and decreasing their emotional availability (Hetherington and Camara,1984;Wallerstein and Kelly,1980).The income loss associated with divorce may require custodial parents to work more in the labor force and therefore to spend less time with their children.Some studies suggest that the adverse effects of divorce on effective parenting are short term.But when a custodial parent remarries,the new marital relationship may compete for the custodial parent's time with and attention to children. Finally,community resources can be a source of marital stability through support networks and normative pressure for a couple to remain together.Some research suggests that parents who subsequently divorce are less strongly tied to kin prior to divorce(Rossi and Rossi,1990).Residential mobility prior to divorce could weaken ties to friends and kin and reduce community resources available to sustain marriages and parent-child relationships.As a marriage becomes troubled, couples may withdraw from kin and community networks in which they have jointly participated and increasingly seek activities in which they are viewed as individuals rather than as a couple.After divorce,residential mobility is common for custodial parents,due primarily to reduced economic resources,and residen- tial mobility weakens community ties(McLanahan and Sandefur,1994).Parental divorce can also lead to a reduction of contact with old acquaintances- friendships that are more likely to have been based on couple-centered activities during marriage-because continued contact is uncomfortable (Milardo,1987). Over time,divorced persons are likely to develop new friendship and organiza- tional ties.Remarriage or repartnering can provide an entree into new kin and community networks. It is also possible that divorce and the process leading up to divorce may increase rather than decrease community resources for children.In the process and aftermath of divorce,parents'need for social support may strengthen their ties to extended family members and friends or increase their social participation (Albrecht,1980;Marks and McLanahan,1992).If so,changes in community resources associated with divorce could be a buffer for custodial parents and children against other resource losses
mother remarries. Even with remarriage, it is not clear that children receive the full economic benefit of a stepfather’s additional income or assets. Parenting resources have been investigated for the most part after divorce. We don’t have evidence for the selection of less effective parents into divorce, but we can theorize that people who have difficulty in sustaining long-term emotional relationships with another adult may also be less effective in their relationships with children. More likely is that parenting difficulties observed prior to a divorce stem directly from the stress of a troubled marital relationship, particularly if the parents are no longer supportive of each other’s parenting practices (Bronfenbrenner, 1986). The most obvious direct effect of divorce itself is, of course, the sharp reduction in contact between children and noncustodial parents (Furstenberg and Nord, 1985; Seltzer, 1991). Divorce may also interrupt bonds between custodial parents and children by preoccupying parents and decreasing their emotional availability (Hetherington and Camara, 1984; Wallerstein and Kelly, 1980). The income loss associated with divorce may require custodial parents to work more in the labor force and therefore to spend less time with their children. Some studies suggest that the adverse effects of divorce on effective parenting are short term. But when a custodial parent remarries, the new marital relationship may compete for the custodial parent’s time with and attention to children. Finally, community resources can be a source of marital stability through support networks and normative pressure for a couple to remain together. Some research suggests that parents who subsequently divorce are less strongly tied to kin prior to divorce (Rossi and Rossi, 1990). Residential mobility prior to divorce could weaken ties to friends and kin and reduce community resources available to sustain marriages and parent–child relationships. As a marriage becomes troubled, couples may withdraw from kin and community networks in which they have jointly participated and increasingly seek activities in which they are viewed as individuals rather than as a couple. After divorce, residential mobility is common for custodial parents, due primarily to reduced economic resources, and residential mobility weakens community ties (McLanahan and Sandefur, 1994). Parental divorce can also lead to a reduction of contact with old acquaintances— friendships that are more likely to have been based on couple-centered activities during marriage—because continued contact is uncomfortable (Milardo, 1987). Over time, divorced persons are likely to develop new friendship and organizational ties. Remarriage or repartnering can provide an entree into new kin and community networks. It is also possible that divorce and the process leading up to divorce may increase rather than decrease community resources for children. In the process and aftermath of divorce, parents’ need for social support may strengthen their ties to extended family members and friends or increase their social participation (Albrecht, 1980; Marks and McLanahan, 1992). If so, changes in community resources associated with divorce could be a buffer for custodial parents and children against other resource losses. 332 HANSON, MCLANAHAN, AND THOMSON SSR625 @xyserv1/disk4/CLS_jrnlkz/GRP_ssrj/JOB_ssrj27-3/DIV_231a04 debb
WINDOWS ON DIVORCE 333 DATA AND MEASURES Data for the analysis comes from the 1987-1988 and 1992-1994 National Survey of Families and Households(NSFH).The first wave ofthe NSFH included 13,008 individuals,including an oversample of minority families,single-parent families,families with stepchildren,cohabiting couples,and recently married persons (Sweet,Bumpass,and Call,1988).Interviews were conducted with a randomly selected adult in the household (primary respondent).Spouses and cohabiting partners of primary respondents(secondary respondents)were asked to complete self-administered questionnaires.The interview response rate for the first wave was 74%.At the second wave of the survey,an attempt was made to interview all primary and secondary respondents from wave 1,including second- ary respondents who no longer lived with primary respondents.A total of 10,008 primary respondents,5,635 wave 1 and new secondary respondents and 785 ex-spouses and ex-partners,were interviewed in the second wave(Sweet,Bump- ass,and Hansen 1995). The sample used in this analysis contains households in which a randomly selected child (focal child)lived with two married,biological (adopted)parents in 1987-1988.To be included in the sample,the focal child also must have lived with a biological mother and have been under the age of 19 at the time of the 1992-1994 survey.Single-father households were excluded from the analyses because they are relatively rare (n=63)and should be treated separately in an analysis of this type focusing on divorce and changes in family resources.The sample restrictions limit the potential sample size to 1813.3 Eleven percent (207/1813)of the children in the analytic sample experienced the breakup of their parents'marriage during the 5 years between 1987-1988 and 1992-1994.Almost half of those children also experienced the custodial mother's remarriage(48)or cohabitation(46)before the second wave. We include three measures of economic resources:(1)total household income, (2)income relative-to-needs and(3)home ownership.Total household income includes the annual income of all household members obtained from all sources (i.e.,wages and salaries,public assistance,child support).We converted house- hold income into 1992 dollars.The income-to-needs ratio is computed by dividing gross household income by the poverty thresholds established by the United States Bureau of the Census.4 An income-to-needs ratio at or below 1 indicates that the household is officially poor,while an income-to-needs ratio 3 Sample attrition was higher among nonwhites-particularly Hispanics-and among those with low levels of schooling.Separate analyses revealed few differences between respondents and nonrespondents on measures of family resources and expectations about divorce in 1987-1988. 4 These thresholds are based on the size of the household and the number of children under age 18. We used household size rather than family size as the basis for the poverty thresholds.Thus,we include the income and consumption of nonfamily members(i.e.,unmarried partners,grandmothers) in determining the standard of living
DATA AND MEASURES Data for the analysis comes from the 1987–1988 and 1992–1994 National Survey of Families and Households (NSFH). The first wave of the NSFH included 13,008 individuals, including an oversample of minority families, single-parent families, families with stepchildren, cohabiting couples, and recently married persons (Sweet, Bumpass, and Call, 1988). Interviews were conducted with a randomly selected adult in the household (primary respondent). Spouses and cohabiting partners of primary respondents (secondary respondents) were asked to complete self-administered questionnaires. The interview response rate for the first wave was 74%. At the second wave of the survey, an attempt was made to interview all primary and secondary respondents from wave 1, including secondary respondents who no longer lived with primary respondents. A total of 10,008 primary respondents, 5,635 wave 1 and new secondary respondents and 785 ex-spouses and ex-partners, were interviewed in the second wave (Sweet, Bumpass, and Hansen 1995). The sample used in this analysis contains households in which a randomly selected child (focal child) lived with two married, biological (adopted) parents in 1987–1988. To be included in the sample, the focal child also must have lived with a biological mother and have been under the age of 19 at the time of the 1992–1994 survey. Single-father households were excluded from the analyses because they are relatively rare (n 5 63) and should be treated separately in an analysis of this type focusing on divorce and changes in family resources. The sample restrictions limit the potential sample size to 1813.3 Eleven percent (207/1813) of the children in the analytic sample experienced the breakup of their parents’ marriage during the 5 years between 1987–1988 and 1992–1994. Almost half of those children also experienced the custodial mother’s remarriage (48) or cohabitation (46) before the second wave. We include three measures of economic resources: (1) total household income, (2) income relative-to-needs and (3) home ownership. Total household income includes the annual income of all household members obtained from all sources (i.e., wages and salaries, public assistance, child support). We converted household income into 1992 dollars. The income-to-needs ratio is computed by dividing gross household income by the poverty thresholds established by the United States Bureau of the Census.4 An income-to-needs ratio at or below 1 indicates that the household is officially poor, while an income-to-needs ratio 3 Sample attrition was higher among nonwhites—particularly Hispanics—and among those with low levels of schooling. Separate analyses revealed few differences between respondents and nonrespondents on measures of family resources and expectations about divorce in 1987–1988. 4 These thresholds are based on the size of the household and the number of children under age 18. We used household size rather than family size as the basis for the poverty thresholds. Thus, we include the income and consumption of nonfamily members (i.e., unmarried partners, grandmothers) in determining the standard of living. WINDOWS ON DIVORCE 333 SSR625 @xyserv1/disk4/CLS_jrnlkz/GRP_ssrj/JOB_ssrj27-3/DIV_231a04 debb
334 HANSON,MCLANAHAN,AND THOMSON above 1 indicates that the household is not poor.Home ownership is a dichoto- mous variable indicating whether the respondent owns her/his home.3 We include four measures of parental resources:(1)mother activities with children,(2)supervision in the home,(3)control,and(4)educational expecta- tions.Mother activities is a summary measure based on mother reports of the frequency of time spent with children sharing the following activities:weekly dinners(0-7),leisure activities away from home,working on a project or playing together at home,and helping with reading or homework ("never or rarely"to "almost everyday,"1-6).Supervision is measured by two items asking whether the focal child is allowed to be at home alone (1)in the afternoon and (2)at night, or overnight (0,1).Control is measured by two items asking about television restrictions:(1)whether the parent restricts the amount of television the child watches and(2)whether the parent restricts the type of programs watched(0,1). We created factor scores for mother activities,supervision,and control based on a confirmatory factor analysis of the items used to measure these constructs.The confirmatory factor analysis models were fitted using Muthen's(1988)methodol- ogy for factor analysis with dichotomous and/or ordinal variables and were standardized to have a mean of 0 and a standard deviation of 1.We measured educational expectations as a dichotomous variable,coded I if the mother reports that she expects the focal child to obtain more than a high school degree and 0 if otherwise. We use a variety of measures of community resources.First,we use two measures of residential mobility:a dichotomous variable indicating whether the parent moved during the 5 years prior to the survey and a continuous variable indicating the number of moves during that time period.We use these variables as proxies for ties to the community.Second,we include a variety of measures of relations between mothers and grandparents,including (1)frequency of contact (“not at all'"to“several times a week,."),(2)relationship quality(“very poor”to "excellent,"1-7,6(3)emotional support received,and (4)instrumental support received during the month prior to the survey(0,1).Instrumental support received includes help with childcare,transportation,and work around the house.Emo- tional support includes advice,encouragement,and moral or emotional support. Third,we examine relations between mothers and friends.We use a measure of the frequency of social time spent with friends(neighbors,coworkers,friends outside neighborhood,group recreation activities),and two dichotomous mea- sures indicating whether emotional support and instrumental support was re- ceived from friends.Finally,we include two very different measures of group participation:(1)religious social activities and (2)time spent in a bar or tavern. The measure of religious social activities is based on a question asking how often 5 Note that home ownership may also produce community resources,because home owners have a stake in maintaining community quality through community involvement,they also have longer tenure in a neighborhood or community and therefore should have stronger ties to their neighbors. 6 In the 1992-1994 round,this item contains 10 categories.We standardized the 1987-1988 item and the 1992-1994 items
above 1 indicates that the household is not poor. Home ownership is a dichotomous variable indicating whether the respondent owns her/his home.5 We include four measures of parental resources: (1) mother activities with children, (2) supervision in the home, (3) control, and (4) educational expectations. Mother activities is a summary measure based on mother reports of the frequency of time spent with children sharing the following activities: weekly dinners (0–7), leisure activities away from home, working on a project or playing together at home, and helping with reading or homework (‘‘never or rarely’’ to ‘‘almost everyday,’’ 1–6). Supervision is measured by two items asking whether the focal child is allowed to be at home alone (1) in the afternoon and (2) at night, or overnight (0, 1). Control is measured by two items asking about television restrictions: (1) whether the parent restricts the amount of television the child watches and (2) whether the parent restricts the type of programs watched (0, 1). We created factor scores for mother activities, supervision, and control based on a confirmatory factor analysis of the items used to measure these constructs. The confirmatory factor analysis models were fitted using Muthe´n’s (1988) methodology for factor analysis with dichotomous and/or ordinal variables and were standardized to have a mean of 0 and a standard deviation of 1. We measured educational expectations as a dichotomous variable, coded 1 if the mother reports that she expects the focal child to obtain more than a high school degree and 0 if otherwise. We use a variety of measures of community resources. First, we use two measures of residential mobility: a dichotomous variable indicating whether the parent moved during the 5 years prior to the survey and a continuous variable indicating the number of moves during that time period. We use these variables as proxies for ties to the community. Second, we include a variety of measures of relations between mothers and grandparents, including (1) frequency of contact (‘‘not at all’’ to ‘‘several times a week,’’), (2) relationship quality (‘‘very poor’’ to ‘‘excellent,’’ 1–7,6 (3) emotional support received, and (4) instrumental support received during the month prior to the survey (0, 1). Instrumental support received includes help with childcare, transportation, and work around the house. Emotional support includes advice, encouragement, and moral or emotional support. Third, we examine relations between mothers and friends. We use a measure of the frequency of social time spent with friends (neighbors, coworkers, friends outside neighborhood, group recreation activities), and two dichotomous measures indicating whether emotional support and instrumental support was received from friends. Finally, we include two very different measures of group participation: (1) religious social activities and (2) time spent in a bar or tavern. The measure of religious social activities is based on a question asking how often 5 Note that home ownership may also produce community resources, because home owners have a stake in maintaining community quality through community involvement; they also have longer tenure in a neighborhood or community and therefore should have stronger ties to their neighbors. 6 In the 1992–1994 round, this item contains 10 categories. We standardized the 1987–1988 item and the 1992–1994 items. 334 HANSON, MCLANAHAN, AND THOMSON SSR625 @xyserv1/disk4/CLS_jrnlkz/GRP_ssrj/JOB_ssrj27-3/DIV_231a04 debb
WINDOWS ON DIVORCE 335 the parent attends a social event at a church or a synagogue ("not at all"to "several times a week,").The bar/tavern measure is based on a similar item asking about time spent in a bar or tavern.We expect that religious groups provide support for effective parenting and access to community resources for children. On the other hand,social activities in bars or taverns are unlikely to provide access to child-centered community resources and may detract from parental resources for children.A"4 times per year"metric was used for contact with parents,time with friends,time with religious groups,and time in bars/taverns. Control Variables In the analyses that follow,we control for parental age(age of oldest resident parent),the number of children in the household,parental education(highest of two parents-no high school degree,high school degree,some college,college degree or more),and race/ethnicity (European-American,African-American, Hispanic,other).If the focal child is the referent to the outcome variable of interest(i.e.,supervision),we also control for the age and sex of the focal child. Otherwise,we control for the age of the youngest child and the number of female children in the household.All of the control variables come from the 1987-1988 survey. RESULTS Table 1 shows means(proportions)of demographic characteristics and the measures of family resources in 1987-1988 and 1992-1994.Columns 1 and 2 show means for families in which the parents were continuously married during the two time periods,while columns 3 and 4 show means for families in which the parents separated or divorced,including families in which there was a remarriage. A comparison of the numbers in column I and column 3 in panel 1 shows that parents who eventually divorced were younger and less educated than parents who stayed together,which underlies some of the gross differences in family resources observed at time 1. Gross Differences in Resources Do families that eventually divorce have fewer resources prior to divorce than families that stay together?Columns I and 3 show,first,that families with parents who eventually divorce have fewer economic resources-household income (p<.10),standard of living (p<10),home ownership-prior to divorce than families with parents who stay together.The results are mixed for parental resources.Mothers who eventually divorce engage in more activities with their children but have lower college expectations than mothers who remain married. There are no significant differences in supervision and television restrictions between the two groups.Results are also mixed for kin/community resources.On the positive side,mothers who eventually divorce report higher levels of contact with their own parents than mothers who stay married.On the negative side, they report lower quality relationships with their parents,less religious-related
the parent attends a social event at a church or a synagogue (‘‘not at all’’ to ‘‘several times a week,’’). The bar/tavern measure is based on a similar item asking about time spent in a bar or tavern. We expect that religious groups provide support for effective parenting and access to community resources for children. On the other hand, social activities in bars or taverns are unlikely to provide access to child-centered community resources and may detract from parental resources for children. A ‘‘4 times per year’’ metric was used for contact with parents, time with friends, time with religious groups, and time in bars/taverns. Control Variables In the analyses that follow, we control for parental age (age of oldest resident parent), the number of children in the household, parental education (highest of two parents—no high school degree, high school degree, some college, college degree or more), and race/ethnicity (European-American, African-American, Hispanic, other). If the focal child is the referent to the outcome variable of interest (i.e., supervision), we also control for the age and sex of the focal child. Otherwise, we control for the age of the youngest child and the number of female children in the household. All of the control variables come from the 1987–1988 survey. RESULTS Table 1 shows means (proportions) of demographic characteristics and the measures of family resources in 1987–1988 and 1992–1994. Columns 1 and 2 show means for families in which the parents were continuously married during the two time periods, while columns 3 and 4 show means for families in which the parents separated or divorced, including families in which there was a remarriage. A comparison of the numbers in column 1 and column 3 in panel 1 shows that parents who eventually divorced were younger and less educated than parents who stayed together, which underlies some of the gross differences in family resources observed at time 1. Gross Differences in Resources Do families that eventually divorce have fewer resources prior to divorce than families that stay together? Columns 1 and 3 show, first, that families with parents who eventually divorce have fewer economic resources—household income ( p , .10), standard of living ( p , 10), home ownership—prior to divorce than families with parents who stay together. The results are mixed for parental resources. Mothers who eventually divorce engage in more activities with their children but have lower college expectations than mothers who remain married. There are no significant differences in supervision and television restrictions between the two groups. Results are also mixed for kin/community resources. On the positive side, mothers who eventually divorce report higher levels of contact with their own parents than mothers who stay married. On the negative side, they report lower quality relationships with their parents, less religious-related WINDOWS ON DIVORCE 335 SSR625 @xyserv1/disk4/CLS_jrnlkz/GRP_ssrj/JOB_ssrj27-3/DIV_231a04 debb
336 HANSON,MCLANAHAN,AND THOMSON TABLE I Demographic Characteristics and Family Resources by Parental Divorce Continuously married Not continuously married 1987 1993 1987 1993 Panel 1:Demographic characteristics Parental age 33.82 30.87b Number of children 2.04 1.909 Education No high school degree 0.11 0.145 High school degree 0.35 0.49 Some college 0.25 0.19 College degree or more 0.29 0.18 Race/ethnicity Non-Hispanic white 0.82 一 0.76 African-American 0.07 0.11 Mexican-American 0.07 0.07 Other 0.04 0.06 Age of child 5.35 4.165 Female child 0.47 0.54d Age of youngest child 4.15 3.48 Number of female children 0.98 一 1.06 Panel 2:Economic resources Household income (1000s) 54.42 63.03c 48.374 35.46r Income/needs 3.93 4.24 3.48d 2.794r Home ownership 0.74 0.84e 0.62b 0.50,e Panel 3:Parental resources Supervision 0.03 -0.09 0.05 -0.02 Mother activities -0.01 -0.08c 0.22b 0.19 Television restrictions 0.02 -0.01 0.15 -0.01 College expectations 0.75 0.72 0.686 0.616e Panel 4:Kin/community resources: Residential move past 5 years 0.62 0.47e 0.735 0.84,e Number of moves(given any) 1.98 1.769 2.27 2.53h,e Contact with parents 58.36 57.85c 71.416 87.986 Relationship with grandmother 0.02 0.02 -0.245 -0.456,e Relationship with grandfather 0.06 0.04 -0.195 -0.195 Instrumental support from parents 0.47 0.52 0.49 0.51 Emotional support from parents 0.42 0.60e 0.41 0.65c Social time with friends 15.45 15.08 16.60 21.046e Instrumental support from friends 0.51 0.37 0.49 0.49b Emotional support from friends 0.51 0.48 0.45 0.564 Social time with religious group 19.02 16.17 8.17b 11.96 Social time in bars/taverns 3.20 2.59 4.16 7.705c a National survey of Families and Households;data are weighted. Significantly different from continuously married group(p <.05) Change between 1987 and 1993 is statistically significant(p<.05) d Significantly different from continuously married group(p<.10). Change between 1987 and 1993 is statistically significant(p <.10)
TABLE 1 Demographic Characteristics and Family Resources by Parental Divorcea Continuously married Not continuously married 1987 1993 1987 1993 Panel 1: Demographic characteristics Parental age 33.82 — 30.87b — Number of children 2.04 — 1.90d — Education No high school degree 0.11 — 0.14b — High school degree 0.35 — 0.49 — Some college 0.25 — 0.19 — College degree or more 0.29 — 0.18 — Race/ethnicity Non-Hispanic white 0.82 — 0.76 — African-American 0.07 — 0.11 — Mexican-American 0.07 — 0.07 — Other 0.04 — 0.06 — Age of child 5.35 — 4.16b — Female child 0.47 — 0.54d — Age of youngest child 4.15 — 3.48b — Number of female children 0.98 — 1.06 — Panel 2: Economic resources Household income (1000s) 54.42 63.03c 48.37d 35.46b,c Income/needs 3.93 4.24c 3.48d 2.79b,c Home ownership 0.74 0.84c 0.62b 0.50b,c Panel 3: Parental resources Supervision 0.03 20.09c 0.05 20.02 Mother activities 20.01 20.08c 0.22b 0.19b Television restrictions 0.02 20.01 0.15 20.01 College expectations 0.75 0.72c 0.68b 0.61b,e Panel 4: Kin/community resources: Residential move past 5 years 0.62 0.47c 0.73b 0.84b,c Number of moves (given any) 1.98 1.76c 2.27 2.53b,e Contact with parents 58.36 57.85c 71.41b 87.98b Relationship with grandmother 0.02 0.02 20.24b 20.45b,e Relationship with grandfather 0.06 0.04 20.19b 20.19b Instrumental support from parents 0.47 0.52c 0.49 0.51 Emotional support from parents 0.42 0.60c 0.41 0.65c Social time with friends 15.45 15.08 16.60 21.04b,e Instrumental support from friends 0.51 0.37c 0.49 0.49b Emotional support from friends 0.51 0.48 0.45 0.56d Social time with religious group 19.02 16.17c 8.17b 11.96 Social time in bars/taverns 3.20 2.59 4.16 7.70b,c a National survey of Families and Households; data are weighted. b Significantly different from continuously married group ( p , .05). c Change between 1987 and 1993 is statistically significant ( p , .05). d Significantly different from continuously married group ( p , .10). e Change between 1987 and 1993 is statistically significant ( p , .10). 336 HANSON, MCLANAHAN, AND THOMSON SSR625 @xyserv1/disk4/CLS_jrnlkz/GRP_ssrj/JOB_ssrj27-3/DIV_231a04 debb
WINDOWS ON DIVORCE 337 social activity,and greater residential mobility.Keep in mind that many of these differences may be spurious,due to parents'younger ages and lower education. Is parental divorce associated with a loss of family resources?Columns 3 and 4 in Table 1 show that parental divorce is associated with substantial declines in economic resources for the mother and child.After divorce,their household income is $13,000 lower,their standard of living is 20%lower,and home ownership drops by 12 percentage points.The drop in home ownership after divorce is illustrative of the loss of economic resources due to divorce.More than one-third of families that owned their homes prior to divorce did not own their home after divorce (not shown). Divorce does not appear to be associated with a loss of maternal resources, although college expectations drop by 7 percentage points.This compares with a 3-percentage-point drop among families that remained married throughout the two time periods.The other parental resources do not change significantly. Divorce is associated with both gains and losses in community resources.Divorce is associated with an increase in residential mobility,which should weaken community ties.It is also,however,associated with a more active social life, including time spent with friends and time spent in bars and taverns.Although both of the latter changes may represent gains for custodial mothers,time spent in bars and taverns is not likely to represent a community resource for children. As noted earlier,the preexisting differences in resource levels between dis- rupted and nondisrupted families as well as the changes in resource levels described above are gross differences.The numbers in Table 1 do not take into account differences in characteristics such as age and education between parents who do and do not divorce.In the remainder of the paper we focus on differences that are adjusted for predivorce differences in demographic characteristics. To investigate differences in family resources at different points in the divorce process,we take advantage of the fact that divorces or separations occurred at varying points between the two surveys,providing several different"windows" on divorce.We can identify families at the first survey who we subsequently discovered to have divorced the following year,families who divorced 2-3 years after the survey,and families who divorced 4-5 years after the first survey.We can compare families who were closer or farther from divorce to those who did not divorce between surveys in terms of their resources at the first survey.In this way, we can see whether resource differences were found prior to the likely beginning of marital troubles or how they emerge as a family moves closer to divorce.In a parallel manner,we compare families who divorced with those who did not at different points after divorce.At the second survey,we observe families who divorced within I year and 2-3 and 4-5 years prior to the survey.7 Resource differences between divorcing and nondivorcing families observed 7 In general,families who divorced soonest contribute to the longer-term divorces observed at the second survey,families who divorced latest to the shorter-term divorces.The time between surveys varied from 4-7 years,however,so that the correspondence is not exact
social activity, and greater residential mobility. Keep in mind that many of these differences may be spurious, due to parents’ younger ages and lower education. Is parental divorce associated with a loss of family resources? Columns 3 and 4 in Table 1 show that parental divorce is associated with substantial declines in economic resources for the mother and child. After divorce, their household income is $13,000 lower, their standard of living is 20% lower, and home ownership drops by 12 percentage points. The drop in home ownership after divorce is illustrative of the loss of economic resources due to divorce. More than one-third of families that owned their homes prior to divorce did not own their home after divorce (not shown). Divorce does not appear to be associated with a loss of maternal resources, although college expectations drop by 7 percentage points. This compares with a 3-percentage-point drop among families that remained married throughout the two time periods. The other parental resources do not change significantly. Divorce is associated with both gains and losses in community resources. Divorce is associated with an increase in residential mobility, which should weaken community ties. It is also, however, associated with a more active social life, including time spent with friends and time spent in bars and taverns. Although both of the latter changes may represent gains for custodial mothers, time spent in bars and taverns is not likely to represent a community resource for children. As noted earlier, the preexisting differences in resource levels between disrupted and nondisrupted families as well as the changes in resource levels described above are gross differences. The numbers in Table 1 do not take into account differences in characteristics such as age and education between parents who do and do not divorce. In the remainder of the paper we focus on differences that are adjusted for predivorce differences in demographic characteristics. To investigate differences in family resources at different points in the divorce process, we take advantage of the fact that divorces or separations occurred at varying points between the two surveys, providing several different ‘‘windows’’ on divorce. We can identify families at the first survey who we subsequently discovered to have divorced the following year, families who divorced 2–3 years after the survey, and families who divorced 4–5 years after the first survey. We can compare families who were closer or farther from divorce to those who did not divorce between surveys in terms of their resources at the first survey. In this way, we can see whether resource differences were found prior to the likely beginning of marital troubles or how they emerge as a family moves closer to divorce. In a parallel manner, we compare families who divorced with those who did not at different points after divorce. At the second survey, we observe families who divorced within 1 year and 2–3 and 4–5 years prior to the survey.7 Resource differences between divorcing and nondivorcing families observed 7 In general, families who divorced soonest contribute to the longer-term divorces observed at the second survey, families who divorced latest to the shorter-term divorces. The time between surveys varied from 4–7 years, however, so that the correspondence is not exact. WINDOWS ON DIVORCE 337 SSR625 @xyserv1/disk4/CLS_jrnlkz/GRP_ssrj/JOB_ssrj27-3/DIV_231a04 debb
338 HANSON,MCLANAHAN,AND THOMSON prior to divorce are estimated from regression models in which resources at the first survey are treated as the dependent variable.Similarly,resource differences between divorcing and nondivorcing families after divorce are based on regres- sion models in which resources at the second survey are the dependent variable. Both sets of models include control variables discussed earlier,all measured at the first survey.s We present the results for pre-and postdivorce differences together in graphical form.Full model parameters and statistics are available on request. In describing our results,we will use the word "loss"or "gain"to describe changes in economic,parental,and community resources associated with the divorce process.Unless otherwise noted,we are speaking of changes in resources relative to that of families with continuously married parents(the"difference"in the "difference").Thus a loss or gain need not reflect an absolute change in resources.For example,if the absolute level of family resources is stable for the divorced between 1987-1988 and 1992-1994,but rises for the continuously married,then the divorced experienced a relative loss of resources during the two time periods. Adjusted Differences in Resources The graphs in Figure I show the relationship between divorce and economic resources.Graph 1.1 shows the difference in household income between families who did and did not divorce.The bar on the left-hand side of the graph shows the income difference prior to divorce,while the bar on the right-hand side shows the income difference after divorce.Couples who divorce between waves I and 2 had about the same income in 1987 as families that remain married.Recall that prior to controlling for differences in background characteristics,the former had about $7,000 less income than the latter(Table 1,panel 2,columns 1 and 3).Differences in parental age and education are primarily responsible for the gross association between 1987-1988 household income and subsequent divorce(not shown). The bar on the right-hand side of graph 1.1 shows that income differences increase after divorce.Mothers and children experiencing divorce or separation by 1992-1994 have about $20,000 less income than couples who remain married. Both the postdivorce income difference and the change in the difference between 1987-1988 and 1992-1994 are statistically significant. Graph 1.2 simulates how household income differences change as parents move closer to and further from divorce.According to the graph,income differences are relatively small and stable prior to divorce,they increase sharply immediately after divorce,and they gradually decline in the postdivorce period. 8 To examine whether changes in resources associated with divorce are statistically significant,we also estimated models that included controls for the level of resources at Wave 1 ("static-score" model). 9 Because income is measured for the calendar year prior to the survey year,we do not observe the income of respondents at the time of separation even for many parents who are very close in time to separating at Wave 1.This could inhibit our ability to see whether economic resources decline as parents move closer in time to divorce.A more fine-grained analysis revealed that those who were
prior to divorce are estimated from regression models in which resources at the first survey are treated as the dependent variable. Similarly, resource differences between divorcing and nondivorcing families after divorce are based on regression models in which resources at the second survey are the dependent variable. Both sets of models include control variables discussed earlier, all measured at the first survey.8 We present the results for pre- and postdivorce differences together in graphical form. Full model parameters and statistics are available on request. In describing our results, we will use the word ‘‘loss’’ or ‘‘gain’’ to describe changes in economic, parental, and community resources associated with the divorce process. Unless otherwise noted, we are speaking of changes in resources relative to that of families with continuously married parents (the ‘‘difference’’ in the ‘‘difference’’). Thus a loss or gain need not reflect an absolute change in resources. For example, if the absolute level of family resources is stable for the divorced between 1987–1988 and 1992–1994, but rises for the continuously married, then the divorced experienced a relative loss of resources during the two time periods. Adjusted Differences in Resources The graphs in Figure 1 show the relationship between divorce and economic resources. Graph 1.1 shows the difference in household income between families who did and did not divorce. The bar on the left-hand side of the graph shows the income difference prior to divorce, while the bar on the right-hand side shows the income difference after divorce. Couples who divorce between waves 1 and 2 had about the same income in 1987 as families that remain married. Recall that prior to controlling for differences in background characteristics, the former had about $7,000 less income than the latter (Table 1, panel 2, columns 1 and 3). Differences in parental age and education are primarily responsible for the gross association between 1987–1988 household income and subsequent divorce (not shown). The bar on the right-hand side of graph 1.1 shows that income differences increase after divorce. Mothers and children experiencing divorce or separation by 1992–1994 have about $20,000 less income than couples who remain married. Both the postdivorce income difference and the change in the difference between 1987–1988 and 1992–1994 are statistically significant. Graph 1.2 simulates how household income differences change as parents move closer to and further from divorce. According to the graph, income differences are relatively small and stable prior to divorce, they increase sharply immediately after divorce, and they gradually decline in the postdivorce period.9 8 To examine whether changes in resources associated with divorce are statistically significant, we also estimated models that included controls for the level of resources at Wave 1 (‘‘static-score’’ model). 9 Because income is measured for the calendar year prior to the survey year, we do not observe the income of respondents at the time of separation even for many parents who are very close in time to separating at Wave 1. This could inhibit our ability to see whether economic resources decline as parents move closer in time to divorce. A more fine-grained analysis revealed that those who were 338 HANSON, MCLANAHAN, AND THOMSON SSR625 @xyserv1/disk4/CLS_jrnlkz/GRP_ssrj/JOB_ssrj27-3/DIV_231a04 debb
