|Home | About | Journals | Submit | Contact Us | Français|
We know little about children who have two living nonresident biological parents. Using data from the National Longitudinal Study of Adolescent Health, this study examines the diverse living arrangements of U.S. adolescents in this situation, the kinds of relationships they have with each of their nonresident parents, and how these living arrangements are associated with adolescent well-being. Differences between these adolescents (N = 502) and those who have one nonresident biological parent (N = 4746) are also examined. Results point to certain groups of adolescents with two nonresident parents who are at particular risk of exhibiting higher levels of behavior problems (those living alone or with an aunt and uncle) or who, alternatively, are faring comparatively better (those living with biological relative caregivers or two nonbiological parent figures).
Continued high rates of divorce and nonmarital childbearing have contributed to large numbers of American children growing up in households with only one biological parent present, usually the mother (Teachman, Tedrow, & Crowder, 2000). Concerns about the consequences of this trend for child well-being have prompted an increase in research focusing on nonresident biological fathers and their children, particularly over the past two decades (Marsiglio, Amato, Day, & Lamb, 2000). More recently, some limited attention has turned toward examining nonresident biological mothers as well (e.g., King, 2007; Stewart, 1999). In most of these studies, the focus is on children who are residing with their other biological parent. As a consequence, we know little about children who have two living nonresident biological parents. Yet such children may be at particular risk of low levels of nonresident parent involvement (Harris & Ryan, 2004) and child well-being (Sun, 2003).
This study seeks to provide a better understanding of adolescents with two living nonresident biological parents who, from a research standpoint, remain largely invisible. Using nationally representative data from the National Longitudinal Study of Adolescent Health (Add Health), we address three key questions: (a) with whom do these adolescents live? (b) how are these living arrangements related to adolescents’ relationships with each of their nonresident parents in terms of contact and closeness? (c) how are these living arrangements related to adolescent well-being? Two important indicators of adolescent well-being are considered: internalizing problems and externalizing problems. Adolescence may be a particularly challenging time for many nonresident parents to maintain an active presence in their children’s lives although such involvement may be crucial as offspring navigate adolescence and the transition to adulthood.
To further shed light on adolescents who have two nonresident biological parents, we compare them to adolescents who have one nonresident biological parent and live with the other biological parent, distinguishing between those who live with their biological mother (and have a nonresident biological father) and those who live with their biological father (and have a nonresident biological mother). We examine differences among these three family structures in levels of nonresident parent involvement and adolescent outcomes.
It is unknown exactly how many U.S. children have two nonresident biological parents. Almost 3 million children (or 4% of all children) were living in nonparent households in 2001 (U.S. Census, 2001). Not all of these children had two living nonresident biological parents, however. Indeed, one of the reasons children live in nonbiological parent households is because one or both of their biological parents is deceased. Nevertheless, the number of children with two nonresident biological parents is not trivial and including such children in our studies will provide a more accurate picture of nonresident parenting.
Existing studies suggest a variety of reasons why children might live in nonbiological parent households, including a parent’s incarceration, mental illness, substance abuse, economic hardship, and child abuse or neglect (Billing, Ehrle, & Kortenkamp, 2002; Sun, 2003). Most of these children live with relatives and usually do so as a result of private arrangements made within a family, not as a result of involvement by a child welfare agency (Hynes & Dunifon, 2007). Even children in the foster care system are increasingly being placed with relatives rather than unrelated foster parents (Hynes & Dunifon).
Children with two nonresident biological parents do not necessarily become separated from them at the same time. As we will document, many children experience the departure of the father (if he was ever co-resident in the child’s household to begin with) sometimes years before the separation from the mother, a likely outcome of divorce or nonmarital childbearing. In addition, there is much variation in the duration of nonparental living arrangements, with some lasting only a few months and others lasting for years (Hynes & Dunifon, 2007). Both of these factors likely have implications for children’s relationships with their nonresident parents.
We know of no prior research that provides detailed information on whom children with two nonresident biological parents live with, although studies of nonparent households in general (in which children may have one or two deceased parents) are relevant in this regard. The vast majority of children in nonparent households live with relatives, with grandparents topping the list, a pattern that likely holds for children with two nonresident parents (U.S. Census, 2001). Information on nonparent households is often limited to details on only a few subcategories or focuses on select subgroups of children, such as those living with relatives or those in foster care (Hynes & Dunifon, 2007). We include all adolescents under 18 years old who report having two nonresident biological parents and provide a more detailed analysis of household types that reveals more of the diversity and complexity of children’s living arrangements.
An adolescent’s living arrangement may be differentially associated with the involvement of nonresident biological parents. Certain caregivers may be more likely to function as or become substitute parents. This is especially likely to be the case for parent figures who have adopted the child.1 An adolescent’s living arrangement may also have different implications for relationships with nonresident mothers compared to nonresident fathers. To the extent that kin caregivers are more likely to be biologically related to the child’s mother than to the child’s father (White & Riedmann, 1992), they may be more likely to support and facilitate the mother’s access to her children. We focus on two important aspects of the nonresident parent-child relationship: contact and closeness. Contact is essential for nonresident parents to maintain a strong presence in their child’s life. Although not a guarantee, more frequent contact appears necessary for nonresident parents to maintain close, high quality relationships with their children (King & Sobolewski, 2006). Closeness of the parent-child bond is a particularly salient dimension of the nonresident parent-child relationship that is associated with better outcomes for children (Amato & Gilbreth, 1999; King, 2007).
We consider two important indicators of adolescent well-being: internalizing problems and externalizing problems. Although prior research suggests that children in nonbiological parent households tend to be at particular risk of lower levels of well-being (Jeynes, 1999; Sun, 2003), these findings are based on studies that combine all children in nonbiological parent households and make comparisons to other family structures that include at least one biological parent. We know of no research examining child well-being across diverse subgroups of children within nonparental households. Do children fare better in some of these arrangements than others? Or perhaps the generally more difficult circumstances of children in nonparental households result in few differences between them.
There are several reasons to expect adolescent outcomes to vary by living arrangements, although different theories offer alternative predictions of where those distinctions may exist. Biological or evolutionary based theories (e.g., Daly & Wilson, 1983) suggest that adolescents who live with biologically related relatives will do better than those who do not because caretakers are predisposed to offer more resources and care to biologically related individuals. This biological perspective predicts that adolescents living with relatives such as grandparents, aunts and/or uncles, and siblings will have fewer internalizing and externalizing problems than those living with nonrelatives, partners, alone, and unrelated parent figures.
Theories that focus on the amount of resources available to children (e.g., McLanahan, 1985; Muller, 1993) suggest that adolescents who live with coupled adults will do better than those who have single caretakers or who live alone. Not only do coupled adult households tend to have higher incomes on average, but the availability of two caretakers implies greater levels of monitoring, supervision, and support of children. This resource perspective predicts that adolescents living with coupled adults such as two parent figures, two grandparents, and an aunt and uncle will have fewer internalizing and externalizing problems than those living with single parent figures, one grandparent, a single aunt or uncle, and those living alone.
Other theories, such as compensation theories (Hamilton, Powell, & Cheng, 2007) or role theory (Biddle, 1986), suggest that adolescents who are adopted or live with other parent figures who have chosen to raise nonbiological children may fare better. This may result from these kinds of caregivers consciously selecting into the caretaking role rather than having it pressed upon them, having an intensified commitment to and more positive view of the child, feeling a greater need to respond to the negative experiences these children faced prior to adoption or face currently as their adopted children, or having greater motivation to be a good parent and to prove themselves as such by actively trying to compensate for the lack of biological ties (Hamilton et al.). Formally adopting the role of a parent may provide clearer norms and expectations for these caregivers to follow than those who unexpectedly find themselves thrust into a parenting role (Biddle). Consistent with this premise, research suggests that single fathers raising children who wanted custody are more satisfied with parenting and better prepared for it than fathers who had custody forced upon them (Greif, 1985). This role perspective predicts that adolescents living with parent figures will have fewer internalizing and externalizing problems than those living with other types of caregivers.
It is important to note that selection into different living arrangements may also play a role in any association between living arrangements and adolescent well-being in that children with higher levels of problem behaviors may be more likely to wind up in certain types of living arrangements (e.g., living alone).
The few studies to compare nonresident biological mothers and nonresident biological fathers suggest that nonresident mothers have more contact with (Stewart, 1999) and closer relationships to (King, 2007) their children than do nonresident fathers. What is less clear is how children with two nonresident biological parents may differ from those who have only one nonresident biological parent. Nonresident father involvement appears to be higher when adolescents live with their biological mothers than when they live with neither parent (Harris & Ryan, 2004). Less is known regarding how nonresident mother involvement varies by whether the child is living with the biological father or not. Do resident fathers facilitate the nonresident mother’s involvement as resident mothers appear to do for nonresident fathers, suggesting that nonresident mother involvement will be greater for adolescents living with their fathers than those living with neither parent? Or perhaps nonparent caregivers, who are more likely to be related to the nonresident mother than the nonresident father, are more or just as likely to support the child’s ties to the nonresident mother, suggesting that nonresident mother involvement will not be different or will be greater when adolescents do not live with their biological father. A recent study by Hawkins, Amato, and King (2006) finds evidence for this latter possibility; patterns of nonresident mother involvement were similar regardless of whether adolescents lived with their biological fathers or not.
Research on child well-being by family structure generally finds that children residing with two biological parents have higher levels of well-being on average than children residing with only one or no biological parents (Amato, 2000; Sun, 2003). Studies comparing children living with only one biological parent are mixed. Some studies report few or small differences in child well-being by whether the resident biological parent is a father or a mother, but when differences are found they generally suggest that children fare worse in father-resident than mother-resident households (Demuth & Brown, 2004; Downey, 1994; Downey, Ainsworth-Darnell, & Dufur, 1998; Hoffman & Johnson, 1998). The more limited studies that include nonparental households suggest that children may fare worse in these households than those in which one biological parent is present (Hollist & McBroom, 2006; Jeynes, 1999), especially if it is the mother who is present (Sun, 2003). The fact that nonparental households are generally more disadvantaged and have even lower levels of nonresident father involvement also lead us to hypothesize that adolescent well-being will be lowest in these households.
The different living arrangements of adolescents with two nonresident biological parents are likely linked in important ways to differential access to resources, both material and in terms of access to the time and attention of caregivers and nonresident biological parents, as well as to other demographic characteristics. For example, adolescents living with coupled adults are likely to live in households with higher average levels of income than those who live with single caretakers or by themselves. Adolescents with two nonresident biological parents likely live in more disadvantaged households on average (e.g., lower income) than either group of adolescents with one nonresident biological parent, but especially in comparison to those living with their biological fathers, who tend to fare better economically (Downey, 1994). In our analyses we control for several factors likely to be related to living arrangements, the involvement of nonresident biological parents, and adolescent well-being (King, 2006, 2007). These factors include characteristics of the adolescent (race, gender, and age), the household (household income and size), and the nonresident biological parents (education and years since sharing a residence with the adolescent).
Data for this study come from the first wave of the National Longitudinal Study of Adolescent Health (Add Health). The full sample includes 20,745 high school and middle school students in 1995. When appropriate sample weights are used, these data are a nationally representative sample of adolescents in grades 7 through 12 in the United States. A parent or parent figure of each adolescent also was asked to complete a questionnaire (n = 17,670; see Bearman, Jones, & Udry, 1997, for a detailed description of the data).
The analysis sample for this study was restricted to adolescents with valid sample weights who were 17 years old or younger and who reported either: (a) having both a living nonresident biological mother and nonresident biological father (n = 502), or (b) living with their biological mother and having a nonresident biological father living elsewhere (n = 4029), or (c) living with their biological father and having a nonresident biological mother living elsewhere (n = 717). Missing data were rare (1% of the sample or less) for most of the variables in the analyses. The two exceptions were household income (24% missing) and the nonresident parent’s education (13% missing). The estimation maximization algorithm in SPSS 14.0 was used to impute missing values. This process produces more reliable estimates than mean substitution or listwise deletion when up to 50% of the cases are missing (Acock, 2005; Allison, 2001).
All analyses are conducted using the Wave 1 sample weight to correct for the differential probabilities of sample selection. The survey (SVY) procedures in Stata (Stata Corporation, 2005) are used to adjust the standard errors of the model estimates for the weighted, clustered, and stratified design of Add Health (Chantala & Tabor, 1999).
All measures come from adolescent reports with the exception of household income, which was reported by the parent-figure.
Adolescents were asked to identify everyone who lived in their household and what their relationship was to them (e.g., father, grandmother, uncle, brother, cousin, nonrelative). Follow-up questions were asked when adolescents identified a mother or father, brother or sister, or son or daughter to determine legal and biological relatedness. For example, if adolescents identified a “father” in the household, they were asked which description best fit this relationship: biological father, stepfather, adoptive father, step/adoptive father, foster father, or other. Further information, such as whether the adoptive parent might have formerly been a foster parent or whether an adoptive or foster parent is biologically related to the child, is unfortunately not available.
Using the full list of household members, we categorized adolescents into different living arrangements on the basis of who was most likely to be the primary caretaker(s) or head of the household (adolescents were not asked to identify the household head or primary caretaker). Some cases proved to be more ambiguous than others in this regard (e.g., a child living only with a grandparent vs. a child living with a grandparent and an aunt). We use two different categorizations of household living arrangements. The first includes a more extended set of categories that we present for descriptive purposes. Although comprehensive, many of the household types were too small to analyze separately. To make the subsequent analyses more manageable, we further combined the household types into a smaller set of 12 categories (see Table 1), which are represented as a set of dummy variables in the regression analyses (alternating the omitted category allowed us to test differences between all 12 categories).
Although different categorizations are possible, we attempted to create a set of meaningful categories that would capture the diversity in living arrangements while also being of sufficient size to allow comparisons between them. A few of the 12 categories remain relatively small so some caution is warranted in interpreting results; one consequence is that comparisons with these smaller categories will be less likely to yield significant differences. The final 12 categories are 1) two parent figures (those who report living with two step, adoptive, foster, or other parents); 2) one parent figure (those living with one step, adoptive, foster, or other parent); 3) two grandparents; 4) one grandparent; 5) aunt and uncle; 6) aunt or uncle; 7) a combination of grandparents and aunts or uncles (including at least one grandparent and one aunt or uncle); 8) siblings; 9) a spouse or partner; 10) alone; 11) other nonrelatives; and 12) all others (including responses of other, other relatives, and the few living only with a child).
Our approach for assigning adolescents to the 12 categories was as follows. When adolescents reported living with a step, adoptive, foster, or other mother and/or father we categorized them as living with a parent figure regardless of who else was living in the household under the assumption that the parent figure(s) is likely the child’s primary caretaker. If no parent figure lived in the household, we next looked at whether they lived with grandparents or aunts or uncles (regardless of other household members). It is particularly unclear who the primary caretaker is in families that include both grandparent(s) and aunt(s)/uncle(s) so we combined such households into a single category that is of less interest to us relative to the living arrangements that distinguish between living with grandparents (no aunts or uncles present) and living with an aunt and/or uncle (no grandparents present). If no parent figure, grandparent, aunt, or uncle was in the household we grouped together adolescents who were living with siblings (including only siblings or siblings with partners and/or children; in most cases at least one sibling was age 18 or older but five adolescents reported living with siblings who were all under age 18). Another group identifies adolescents who were living with a spouse or partner. Some lived only with the spouse/partner or with the spouse/partner and their children. Others, however, were living in households with other relatives or nonrelatives, some of whom may be the household head or may serve as a caretaker figure (e.g., grandparents, uncles, older siblings, or in-laws). Those who are classified as living alone did not report anyone else in their household. The nonrelatives category includes those who live only with nonrelatives. We have no further information on who these nonrelatives are, but they likely include roommates or friends. The final other group consists of all remaining adolescents including those who report ambiguous responses of “other relative,” “other,” and the few living only with a child – a unique and interesting group but one that is too small to analyze separately. This final group is also of less interest to us in terms of examining subgroup differences because of its heterogeneity and ambiguity, but having them in the analysis allows us to include all of the adolescents with two nonresident parents.
In separate questions, adolescents reported how close they felt (1 = not at all close, 5 = extremely close) to each of their nonresident biological parents. Contact with each nonresident biological parent is the average of two items indicating how often in the past 12 months (0 = not at all, 5 = more than once a week) the adolescent has stayed overnight with the nonresident parent, and how often the adolescent talked to the nonresident parent in person or on the telephone or received a letter from them. In models comparing adolescents with two nonresident parents to adolescents with only one nonresident parent, we used control variables pertaining to the nonresident parent that the adolescent reported being closest to (more often the mother, 48%, than the father, 16%) for the sample of adolescents with two nonresident parents. For adolescents who reported identical levels of closeness to their nonresident mother and nonresident father (36%), we used the information pertaining to the nonresident mother. Research suggests that closeness to nonresident mothers may be more strongly associated with adolescent well-being than closeness to nonresident fathers (King, 2007), a finding confirmed in this study.
Two child outcomes, internalizing problems and externalizing problems, were created from adolescent reports. For each outcome, a factor score was calculated based on a factor analysis of three standardized scales. For internalizing problems, the scales were depressive symptoms, negative outlook, and low self-esteem (all three factor loadings were above .7). Depressive symptoms is the average of seven items (α = .83) tapping feelings in the past week (0 = never or rarely, 2 = a lot, most, or all of the time) such as feeling lonely, feeling sad, and being unable to shake off the blues. Negative outlook is the average of four items (α = .70) tapping the absence of positive feelings in the past week including feeling as good as other people and feeling hopeful about the future (0 = most or all of the time, 3 = never or rarely). Low self-esteem is the average of six items (α = .82) regarding disagreement with statements about the self such as having a lot of good qualities and liking oneself (1 = strongly agree, 4 = disagree or strongly disagree).
The scales that form the externalizing problems factor are nonviolent delinquency, violence, and substance use (all three factor loadings were above .7). Nonviolent delinquency is the average of 10 items (α = .79) regarding whether adolescents engaged in certain delinquent behaviors in the past 12 months (0 = never, 2 = 3 or more times), including stealing, lying, and general antisocial behavior. Violence is the average of eight items (α = .82) referring to fighting (0 = never, 2 = 3 or more times) and using weapons (0 = never, 2 = more than once) in the past 12 months. Substance use is the average of six dichotomous items (α = .84) tapping moderate to heavy use (as opposed to no use, or very infrequent/very modest use or experimentation that is fairly common in adolescence but that often abates by young adulthood; Hetherington & Kelly, 2002) of tobacco (in the past month), alcohol (in the past year), and marijuana (in the past month). These same items in the scales that form internalizing and externalizing problems have been used in prior research using Add Health data (e.g., Hawkins, Amato, & King, 2007).
Race-ethnicity is measured as a set of dummy variables that includes non-Hispanic Whites (omitted reference group), non-Hispanic Blacks, Hispanics, and all others. The number of Asians, Native Americans, and other groups were too small to analyze separately. Adolescent’s gender is a dichotomous variable (1 = female, 0 = male). Adolescent’s age is a continuous variable ranging from 11 to 17 years. Income is a continuous variable reported in the parent figure survey that refers to the income in thousands of dollars of the household in which the adolescent lives. The log of this variable is used in the regression analysis to minimize skewness. Household size refers to the total number of people residing in the adolescent’s household. Each nonresident biological parent’s education is an ordinal variable ranging from 1 = eighth grade or less to 8 = postgraduate training. Years since lived with the nonresident father or the nonresident mother refers to the number of years since the adolescent lived with the biological father or biological mother. For adolescents who never lived with the biological parent, this variable corresponds to their age.2
Table 1 reveals the great diversity of living arrangements among adolescents who have two nonresident biological parents while also highlighting some of the most common ones. (The percentages reported in the table are weighted.) Just over 13% of the adolescents report living with parent figures – adoptive, step, foster, or other parent(s). Of these adolescents, almost twice as many live with a single parent figure, who is overwhelmingly likely to be a mother figure, than live with two parent figures. Living with grandparents or aunts and/or uncles is also common. The most common living arrangement (18%) is living with two grandparents (no parent figures, aunts or uncles present). Over 12% of the households included both a grandparent(s) and aunt(s)/uncle(s). Unlike what we found for parent figures, adolescents living with grandparents and/or aunts/uncles are more likely to experience living with a couple than a single individual.
Most of the remaining categories of living arrangements are relatively uncommon but cumulatively comprise a significant minority of living arrangements. These include living with siblings, living with a spouse or partner with or without others present, and living alone. Several living arrangements are ambiguous, with adolescents responding that they are living with “other relatives” only (less than 2%), “other nonrelatives” only (11%), or “others” (3%). The last column of Table 1 indicates how the more extensive set of living arrangement categories were combined to create a more manageable set of 12 household types for subsequent analyses (e.g., Tables 2 and and44).
The most consistent finding revealed in Table 2 is the lower levels of nonresident parent involvement for adolescents who are living with two parent figures. Compared with most other arrangements, adolescents living with two parent figures have the lowest levels of contact with their nonresident fathers and nonresident mothers, and on average they report being less close to their nonresident fathers. Interestingly, despite lower levels of contact, adolescents living with two parent figures do not report being significantly less close to their nonresident mothers. Indeed, there is little variation in levels of closeness to nonresident mothers with the single exception of adolescents who are living with a spouse or partner who report being closer to their nonresident mothers than adolescents in a few other living arrangements, a pattern that is also seen in their somewhat higher levels of closeness to nonresident fathers and more frequent contact with both nonresident parents than a few other groups. Perhaps these adolescents married early with their parents’ blessing.
The only other discernable pattern to emerge was the comparatively lower level of nonresident father contact for adolescents living in households with an aunt and uncle. Overall, it appears that adolescents living with two parent figures are least likely to maintain strong connections to their nonresident biological parents, although closeness to nonresident mothers is relatively unaffected by the adolescent’s living arrangements.
Relationships to nonresident biological parents are further elucidated by the comparisons shown in Table 3 between adolescents with two nonresident biological parents and those with only one nonresident biological parent. The first comparison only includes children with one nonresident parent and compares those who live with their biological mothers (and have a nonresident father) to those who live with their biological fathers (and have a nonresident mother). The findings replicate prior studies that suggest that children have more contact with and are closer to nonresident mothers than nonresident fathers.3
The second set of models examines children with nonresident fathers and compares those who also have a nonresident mother to those who live with their mother. When adolescents have two nonresident parents they have lower levels of contact with their nonresident fathers on average than adolescents who live with their mothers, suggesting that the resident mother may be facilitating the child’s contact with the father in ways that other caretakers do not or perhaps can not. Despite differences in contact, however, the two groups of adolescents do not differ in levels of closeness to their nonresident fathers. The final comparison involves children with nonresident mothers and compares those who also have a nonresident father to those who live with their fathers. Levels of contact and closeness to the nonresident mother do not vary by whether the adolescent is living with the father or has two nonresident parents.
In sum, although adolescents’ relationships to nonresident mothers and nonresident fathers differ, with higher levels of contact and closeness to nonresident mothers than to nonresident fathers, there are fewer differences in adolescents relationships to nonresident parents by whether the other biological parent is resident or not. The one exception is the somewhat lower levels of nonresident father contact experienced by adolescents who also have a nonresident mother compared to adolescents who reside with their mothers.
Levels of internalizing problems significantly differ between several of the subgroups of adolescents who have two nonresident parents (see Table 4). Adolescents living with an aunt and uncle or who live with other nonrelatives exhibit higher levels of internalizing problems than several other groups including adolescents living with two parent figures, one or both grandparents, or siblings. Indeed, adolescents living with one or both grandparents or with siblings exhibit the lowest levels of internalizing problems. Other groups that show higher levels of internalizing problems on average include adolescents living with only one parent figure, adolescents living with a spouse or partner, and adolescents who are living alone. These results lend some support for the importance of biological relative caregivers, especially grandparents and siblings, although not unequivocally since those living with an aunt and uncle have relatively higher levels of internalizing problems. The higher levels of internalizing problems among adolescents living with aunts and uncles may reflect the fact that only one of these adults is biologically related to the adolescent, or other family dynamics associated with these households (e.g., competition between the aunt and uncle’s own children and the adolescent).
Similar to what was found for internalizing problems, adolescents living with an aunt and uncle show higher levels of externalizing problems than several other groups of adolescents (e.g., adolescents living with two parent figures, with a single aunt or uncle, with grandparent(s) and aunt(s)/uncle(s)). Higher levels of externalizing problems are also exhibited by adolescents living alone or with siblings, although differences only reach significance in comparison to adolescents living with a single aunt or uncle. Taking results for internalizing and externalizing problems together, it appears that adolescents living with an aunt and uncle and those living alone are at particular risk of experiencing elevated levels of problem behaviors.
Results from these models (not shown in Table 4)4 also revealed that closeness to the nonresident biological mother was associated with both fewer internalizing problems (b = −.15, p < .05) and externalizing problems (b = −.14, p < .01), suggesting that the continuing involvement of nonresident mothers can have a positive influence on adolescent outcomes if a high quality relationship can be maintained. Closeness to the nonresident biological father was not significantly associated with either outcome.
Table 5 compares all adolescents with two nonresident parents to those who have only one nonresident parent. Adolescents living with their mothers who have a nonresident father exhibit significantly fewer internalizing problems than the two groups of adolescents with a nonresident mother (i.e., those living with their fathers who have a nonresident mother and those who have two nonresident parents). Adolescents living with their resident fathers do not significantly differ on internalizing problems from adolescents with two nonresident parents. There are no significant differences between adolescents in these three family structures in levels of externalizing problems although the trend for the lowest average levels among adolescents living with their mothers is in the same direction as was found for internalizing problems. Overall, adolescent well-being appears higher on average when the mother is resident.
Results from these models (not shown) also revealed that regardless of family structure, closeness to a nonresident parent was associated with both fewer internalizing problems (b = −.09, p < .001) and externalizing problems (b = −.09, p < .001).
There is great diversity in the living arrangements of adolescents with two nonresident biological parents, which is largely masked in prior research on children in nonparental households that either considers all of these adolescents as a homogeneous group or that makes simple distinctions between a few subgroups such as living with relatives vs. nonrelatives. To be sure, some of these living arrangements are more common than others. Similar to prior research on nonbiological parent households, we find that living with grandparents is the most common arrangement. Other adolescents, however, are cared for by aunts and/or uncles, adult siblings, or other parent figures. A few live on their own, with a partner, or with other nonrelatives.
Adolescents’ relationships to nonresident mothers and nonresident fathers differ, with higher levels of contact and closeness to nonresident mothers than to nonresident fathers. This gender difference is apparent whether we examine children who have only one nonresident parent, as prior research has done (King, 2007), or we examine children who have two nonresident parents. Although not being resident poses obstacles for parents in maintaining ties to their children, mothers are more effective at maintaining these relationships from a distance. A fruitful area for future research is to explore why this gender difference exists. It may reflect gendered patterns in children’s relationships to parents that existed before the parents became nonresident; nonresident mother’s greater effort at maintaining ties after separation; or in the case of adolescents with two nonresident parents, the influence of caregivers who may support the child’s ties to the mother more readily than those to the father. This may reflect in part the greater likelihood that kin caregivers are more likely to be biologically related to the child’s mother than to the child’s father (White & Riedmann, 1992). Also relevant is the fact that the separation of the child from the father often occurred prior to the separation from the mother, three years earlier on average.5
There are fewer differences in adolescents’ relationships to nonresident parents by whether the other biological parent is resident or not. Adolescents with two nonresident parents do not differ from those who live with their biological father and have a nonresident mother in levels of contact or closeness to the nonresident mother. Adolescents with two nonresident parents also do not differ from those with who live with their biological mother and have a nonresident father in levels of closeness to the nonresident father. The one exception is the somewhat lower levels of nonresident father contact experienced by adolescents who also have a nonresident mother compared to adolescents who live with their mothers. It may be that resident mothers try to more actively facilitate the child’s contact with the father in ways that other caretakers do not or perhaps cannot.
The living arrangements of adolescents with two nonresident parents have some influence on patterns of parent-child involvement, although again, many living arrangements did not differ from one another in this regard. The most consistent finding is the lower levels of nonresident parent involvement for adolescents living with two parent figures. Compared with adolescents in most other living arrangements, these adolescents report the lowest average levels of contact with their nonresident fathers and nonresident mothers, and on average they report being less close to their nonresident fathers. Levels of nonresident father contact were also notably lower for adolescents living in households with an aunt and uncle. It appears that in these households, the two parent figures or aunt and uncle more likely serve as “replacement” or substitute parents than other caretakers.
Adolescents living with two parent figures have also been living apart from their biological parents for a much longer period of time, often several years on average, than adolescents in other living arrangements, which may have facilitated the development of closer bonds to the caretaker. Concurrently, the passage of time and younger age of the child at the point of parental departure may have eroded bonds to the nonresident parent. The passage of many years since both parents lived with the child also suggests a more permanent situation than a temporary one that may exist in other types of arrangements (e.g., a grandparent taking in a grandchild until a parent can take them back) or a situation that represents a child’s early transition into adult roles (e.g., leaving the parental home to get married). Finally, some of these parent figures have also legally adopted the child, which in effect terminates the parental rights of the biological parents. Nevertheless, a majority of the adopted adolescents had contact with at least one of their nonresident biological parents in the past year, including 60% of the adopted adolescents living with two parent figures and 75% of the adopted adolescents living with one parent figure. As noted previously, these adolescents are a selective subgroup of adopted children compared with the more typical pattern of closed adoption at birth where ties to the biological parents are not usually maintained.
Another notable finding, and an exception to the above, is the lack of variation in levels of closeness to nonresident mothers. Even adolescents living with two parent figures, who report levels of contact with nonresident mothers that are lower on average than other groups of adolescents, do not report being less close to their nonresident mother. Again, the nonresident mother-child relationship appears to be somewhat more resilient, whether this is due to greater motivation to preserve this relationship on the part of the nonresident mother, the adolescent, or among various caregivers. Further, our findings suggest that maintaining close ties to the nonresident biological mother is associated with better adolescent outcomes. Both caregivers and agencies working with these families may want to encourage and facilitate close ties between children and their nonresident parents when they exist.
Our findings also suggest that some adolescents in certain living arrangements appear to be at higher risk of experiencing elevated levels of internalizing and externalizing problems, including those living with an aunt and uncle and those living alone. The higher levels of problem behavior among the latter group may reflect the lack of material resources as well as the time, supervision, and commitment of resident caregivers. Adolescents living with an aunt and uncle, however, are in households with relatively higher incomes and two adult caretakers, yet this does not translate into better outcomes. This may reflect the fact that only one of these adults is biologically related to the child. The unrelated adult (most likely the uncle) may feel less connection to the child and indeed may be more likely to be a reluctant care provider or even resent the adolescent’s presence in the household. Certainly research on stepfamilies suggests that such processes may occur with some stepfathers being disengaged or reluctant caregivers who compete with the child for the mother’s time and attention, which may be one reason that children in stepfamilies do not do as well as children living with both biological parents and often do not do any better than children living with single parents (McLanahan & Sandefur, 1994).
In aunt and uncle households, even the related adult may be a reluctant care provider. Consistent with the high frequency of grandparent caregivers, individuals are more likely to turn first to parents for help; sibling ties tend to be more peripheral and individuals with living parents are less likely to report that they would ask a sibling for help, suggesting that they may do so more reluctantly (White & Riedmann, 1992). In addition, there may be more competition for resources in aunt and uncle households where couples are concurrently raising their own biological children.6 This situation may also lead to friction or tension between the adolescent and the biological children of the couple. Finally, selection into different living arrangements may also play a role in that children with higher levels of problem behaviors may be more likely to wind up living alone or with aunts and uncles.
Adolescents exhibiting fewer internalizing problems on average included those living with grandparent(s) or siblings. Adolescents with fewer externalizing problems on average included those living with two parent figures, two grandparents, an aunt or an uncle. Overall, these results lend some support for the importance of biological relative caregivers, but not unequivocally. The higher incidence of internalizing and externalizing problems among adolescents living with an aunt and uncle does not fit easily with biologically based theories. Evidence for other theories, however, is even more mixed. The higher incidence of problems among adolescents living with an aunt and uncle also does not support resource or coupled adult theories and overall there were few differences between single and coupled adult households (e.g., adolescents living with one grandparent were not significantly different on levels of internalizing or externalizing problems than those living with two grandparents). With respect to compensation or role theories, adolescents living with two parent figures exhibited fewer problems than some other groups, but at the same time were not doing any better than many others (e.g., adolescents living with grandparents).
Overall, our findings suggest that simple distinctions between biological and nonbiological caregivers or between coupled and single caregivers do not adequately identify adolescents who are doing better or worse, and future research should consider ways to go beyond them. Indeed, categorizations such as these may miss a group of children who are at particular risk of exhibiting higher levels of behavior problems – namely those living with an aunt and an uncle – despite living with a coupled biological caregiver.
Future research would benefit from examining whether the higher levels of behavior problems found in this study among adolescents living with an aunt and uncle extend to other child outcomes or other samples of children, such as all children in nonbiological parent households (not just those with two living nonresident biological parents). An important unanswered question for future research is why adolescents living with an aunt and uncle appear to be at risk for poorer outcomes. For example, does competition for resources or tension between the adolescent and the biological children of the aunt and uncle play a role? A greater understanding of why such adolescents appear to be at risk would allow agencies working with these families to better assist them.
Comparing all adolescents with two nonresident parents to those who have only a nonresident mother or only a nonresident father reveals that adolescent well-being is higher on average when the mother is resident. Adolescents living with their mothers who have a nonresident father exhibit significantly fewer internalizing problems than the two groups of adolescents with a nonresident mother (these two latter groups do not differ significantly from each other). There are no significant differences between adolescents in these three family structures in levels of externalizing problems, although the trend for the lowest average levels among adolescents living with their mothers is in the same direction as was found for internalizing problems. These findings are consistent with studies suggesting that the disadvantages faced by children in single parent and stepfamilies may be more pronounced when the resident biological parent is a father rather than a mother (Hoffman & Johnson, 1998), as well as studies suggesting that adolescents who live with neither parent have particularly low levels of well-being (Jeynes, 1999; Sun, 2003). Again, such differences may also reflect that the route to father residence or nonparental residence may be different from, and potentially more traumatic than, the more common occurrence of mother residence after parental separation.
This study is limited by the lack of information in Add Health on the circumstances responsible for the child being separated from the biological parents and the route by which they came to be in their current living arrangement. This study is also limited by examining these adolescents at a single point in time. The living arrangements of these adolescents may be unstable (Hynes & Dunifon, 2007), with some having experienced multiple transitions along the way. Some may eventually return to living with the nonresident parent(s) but many others likely will not. All of these factors (reasons for the parental separation and its permanence, instability of living arrangements, and circumstances leading to the current arrangement) likely affect both adolescents’ relationships with nonresident parents as well as their levels of well-being. Future research would benefit from a more thorough investigation of these issues.
Finally, the findings reported here regarding differences between adolescents in different living arrangements may reflect in part the choices we made in assigning adolescents to these groups, particularly given that it was not possible to definitively identify who was serving as the primary caretaker(s) of the adolescent. Different categorizations could lead to different results. Some may also view certain categories of adolescents (e.g., those living with a spouse) as representing a qualitatively different kind of living arrangement that they would not include in a study of children with two nonresident biological parents.7 We chose to examine the living arrangements of all children under 18 who reported having two nonresident biological parents in order to represent the full diversity of circumstances that exist.
We know little about children who have two living nonresident biological parents. This study makes important contributions toward better understanding the living arrangements of U.S. adolescents in this situation, the kinds of relationships they have with each of their biological parents, and how these living arrangements are associated with adolescent well-being. Although these adolescents may generally be at risk, there is also much diversity in their circumstances. Our study points to two groups of adolescents who may be at particular risk for poor outcomes, those living alone and those living with an aunt and an uncle. Our findings also suggest that adolescents who fare comparatively better are those who live with biological relative caregivers, such as grandparents, or with two nonbiological parent figures.
This research was supported by funding from the Eunice Kennedy Shriver National Institute of Child Health and Human Development (NICHD) to Valarie King, principal investigator (R01 HD43384), and from core funding to the Population Research Institute, The Pennsylvania State University (R24 HD41025). This research uses data from Add Health, a program project designed by J. Richard Udry, Peter S. Bearman, and Kathleen Mullan Harris, and funded by a grant P01 HD31921 from NICHD, with cooperative funding from 17 other agencies. Special acknowledgment is due Ronald R. Rindfuss and Barbara Entwisle for assistance in the original design. Persons interested in obtaining data files from Add Health should contact Add Health, Carolina Population Center, 123 W. Franklin Street, Chapel Hill, NC 27516-2524 (ude.cnu@htlaehdda). No direct support was received from grant P01 Hd31921 for this analysis. We thank Paul Amato, Alan Booth, Cassandra Dorius, Daphne Hernandez, Bryndl Hohmann-Marriott, Catherine Meyers, Jennifer Pearce-Morris, Mindy Scott, Jinsook Helen Seo, and Christina Wolfe for their helpful comments on previous drafts. We also thank Jeanne Spicer of the Population Research Institute for expert programming assistance.
1It should be noted that the adopted adolescents in our study (n = 60) are a selective subgroup of adopted children compared with the more traditional pattern of closed adoption at birth, where ties to the biological parents are usually not maintained. All of the adopted adolescents in our study reported having both a living nonresident biological mother and a living nonresident biological father that they knew something about, a situation that is likely very uncommon for children who had a traditional adoption at birth or a very young age. A majority of these adolescents, 75%, further reported that they had lived with at least one of their biological parents in the past and a majority, 65%, also reported that they had contact with at least one of their nonresident biological parents in the past year.
2Tables with descriptive information (e.g., frequencies, means, and standard deviations) for all study variables for the full sample, the main three subsamples (two nonresident parents, nonresident mother only, nonresident father only), and the 12 subgroups of adolescents with two nonresident parents are available from the first author upon request. As expected, the different living arrangements were linked to the availability of different resources. For example, adolescents with two nonresident biological parents were generally living in more disadvantaged circumstances (e.g., lower household income, lower parental education) than adolescents who only had one nonresident biological parent. Adolescents with two nonresident parents also varied among themselves in terms of resources with coupled adult households having higher incomes on average than their single adult household counterparts.
3This gender difference is also found within our sample of adolescents with two nonresident parents: nonresident mother contact M = 2.76 vs. nonresident father contact M = 1.83, p < .001; nonresident mother closeness M = 3.72 vs. nonresident father closeness M = 3.01, p < .001.
4All results referred to and not shown are available upon request.
5Further analyses revealed that 9% (n = 43) of the adolescents reported never residing with either biological parent (the majority of these adolescents were currently residing with relatives). Another 2% (n = 10) of the adolescents report never residing with the nonresident mother (but previously resided with the nonresident father), and 20% (n = 102) report never residing with the nonresident father (but previously resided with the nonresident mother). Directly comparing the adolescent’s report on the number of years since he or she lived with each biological parent revealed that 42% experienced separation from both parents simultaneously (or within the same year), whereas 48% report that the separation of the father occurred earlier than the separation from the mother; only 11% indicated that the separation from the mother occurred first.
6Among the 45 adolescents who were living with their aunt and uncle, 78% reported having at least one cousin in the household (67% had two or more cousins), and in 26% of these families, the adolescent had a sibling present in the household as well.
7In our analyses comparing adolescents with two nonresident parents to those with one nonresident parent (i.e., Tables 3 and and5),5), we tested models excluding several of these subgroups of adolescents with two nonresident parents (e.g., those living with a spouse, those living alone, and those in the diverse other category). Results and conclusions remained the same, suggesting robust findings across different specifications of the two nonresident parent’s sample.
Valarie King, Department of Sociology, 211 Oswald Tower, The Pennsylvania State University, University Park, PA, 16802. Email: ude.usp.pop@gnikv. Office: (814) 863-8716. Fax: (814) 863-7216. Home: (814) 466-7226.
Katherine C. Stamps, Department of Sociology, 211 Oswald Tower, The Pennsylvania State University, University Park, PA, 16802. Email: ude.usp.pop@spmatsk. Office: (770) 633-3627. Fax: (814) 863-7216. Home: (770) 633-3627.
Daniel N. Hawkins, Department of Sociology/Anthropology, 6001 Dodge St, University of Nebraska at Omaha, Omaha, NE 68182. Email: ude.ahamonu@snikwahd. Office: (402) 554-4935. Fax: (402) 554-3786. Home: (814) 571-5485.