Adolescent Intergenerational Relationship Dynamics and Leaving and Returning to the Parental Home.

OBJECTIVE: Drawing on the life course perspective and theoretical models of intergenerational solidarity, this research explores how adolescent-parent relationships (i.e., parent-child closeness, parental attentiveness, family routines, and parenting styles) are associated with young adults' transitions to adulthood.
BACKGROUND: The study adds to the growing literature on adolescents' leaving and returning to the parental home by focusing on parent-child relationships and variations across gendered parent-child dyads.
METHOD: Based on data spanning nearly 2 decades from the National Longitudinal Survey of Youth 1997 (N = 5,201), event history analysis was employed to assess how intergenerational family dynamics correlate with young adults' risk of leaving (n = 4,519) and returning to (n = 2,749) the parental home.
RESULTS: The results indicate that, net of individual, household, and other contextual factors, parent-child closeness is significantly and positively associated with leaving the parental home. This suggests that close parent-child relationships can help launch children into adulthood. Looking at returns to the parental home, closeness becomes significant for daughters only and is moderated by parent gender. In addition, measures of parenting style indicate a significant and negative association between more-passive styles and children's return to the parental home.
CONCLUSION: These findings highlight the need to more closely consider the impact of gender and parent-child relationship dynamics in facilitating young adults' transition to adulthood.

Leaving the parental home is commonly understood to be a major milestone in the transition to adulthood. It signals the onset of young adults' social and financial independence and, in many cases, frames their housing careers, labor market participation, and other important life course trajectories. In the past decade, there has been a marked increase in the number of young adults residing in their parents' homes, and the period of parent–child coresidence has become protracted (Fry, 2017; Qian, 2012). Researchers have also been interested in the “reversibility” of the transition to adulthood, especially returns to the parental home after a period of autonomous living (e.g., Houle & Warner, 2017).

From a young adult's standpoint, decisions to move in and out of the parental household are based on a variety of individual, household, and other contextual factors—among them, the context and quality of their relationship with their parents. Previous research suggests family structure and relationships are associated with leaving and returning home. In stepfamilies, low‐quality relationships can lead to earlier departure from the parental home (Aquilino, 1991; Goldscheider & Goldscheider, 1999). Few studies, however, have drilled down to explore this relationship more explicitly—particularly whether and how intergenerational relationships facilitate or hinder young adults' achievement of independent adulthood.

By and large, studies of leaving and returning home tend to focus on parent‐level resources, household size and structure, and child characteristics. Few have undertaken a comprehensive analysis of whether and how intergenerational and family relationships are linked to leaving and returning to the parental home. Among those who have, the results have been mixed.

For example, Ward and Spitze (2007) drew on the first two waves of the National Survey of Families and Households (N = 765) and found that children with lower quality relationships with their parents at T1 were more likely to move back home by T2. However, South and Lei (2015) recently found in the Panel Study of Income Dynamics that stronger mother–child relationships are associated with a higher risk of moving out of the parental home, but did not find support for moving back. As such, it remains unclear whether and how parent–child relationships are associated with home leaving and returning. The primary objective of this study is to assess whether departing and returning to the parental home differ based on the parent–child relationship.

Using nationally representative panel data spanning nearly 2 decades covering the entire transition to adulthood for a cohort of young adults, this study provides the most thorough treatment of how parent–child and family dynamics facilitate or deter youth's initial departure from—and potential returns to—the parental home. Two important theoretical perspectives help provide a framework for thinking about the transition to adulthood in this context: the life course perspective and Bengtson's multidimensional model of intergenerational solidarity. Drawing on these frameworks and adding our own theories of the moderating effects of gender, we look beyond traditional individual, household, and other contextual variables to explore (a) whether and how early intergenerational relationships are associated with young adults' departure from the parental home, (b) whether the effects are similar for returns to the parental home, and (c) whether the effects of these intergenerational relationships are moderated by gender.

life course perspective

A key framework guiding this study is the life course perspective (Elder, 1998), which takes an integrative theoretical approach to the interactions between human agency, interpersonal, social, and historical contexts in shaping individuals' developmental trajectories (Elder, 1998; Elder, Johnson, & Crosnoe, 2003). This framework has been applied to a long tradition of research on leaving and returning home. The following two central themes from the life course perspective apply to the current study: variability and linked lives.

Variability underscores interindividual variation in individuals' social status, resources, and social roles (Shanahan, 2000). As a result, economic resources and other risk factors are unevenly distributed, which can also influence individuals' residential trajectories and the transition to adulthood. The concept of linked lives emphasizes that individuals' lives are “typically embedded in social relationships with kin and friends across the life span” (Elder, 1994, p. 6). It provides a basis for the notion that parent–child interactions and dynamics impact young adults' initial departure from the parental home.

Most researchers agree that the role of parents is central to children's development. In particular, parent–child relationships are closely linked to adolescent and young adult outcomes, such as delinquency, autonomy, and educational outcomes (e.g., Brenning, Soenens, Petegem, & Vansteenkiste, 2015; King, Boyd, & Pragg, 2018). For this reason, we expect that young adults' decisions about whether to leave, stay, or return will also be associated with the socioemotional context of the parent–child relationship. However, the mechanisms behind these associations are less clear, largely because of the multidimensional nature of parent–child relationships, which are reflected in models of intergenerational solidarity (Bengtson & Roberts, 1991).

intergenerational solidarity

Recent decades have seen longer life expectancies, high rates of divorce, and the protraction of emerging adulthood whereby children and parents rely on each other for longer periods of time (Furstenberg, 2010). As a result, intergenerational relationships, and parent–child relationships in particular, have become a more central aspect of family life. Therefore, we also employ the work of Bengtson (2001) and Bengtson and Roberts (1991) on the multidimensionality of intergenerational relationships to add nuance to the theoretical framework of the current study. Their model highlights several dimensions defining parent–child interactions, from which we draw primarily on the following three: affectual solidarity, associational solidarity, and normative–functional solidarity.

Affectual Solidarity

Affectual solidarity is rooted in the “sentiments and evaluations family members express about their relationship with other members” (Bengtson, 2001, p. 8). Affectual solidarity, or intergenerational feelings of emotional closeness, has been linked to geographic distance between parents and their children later in adulthood (Gillespie & Treas, 2017; Gillespie & van der Lippe, 2015). However, there are competing expectations regarding the ways affectual solidarity is associated with leaving and returning to the parental home. Some previous research finds adults retrospectively reporting leaving home earlier because of conflict with their parents (Cherlin, Kiernan, & Chase‐Lansdale, 1995). As Rogerson, Burr, and Lin (1997, p. 658) argued, the premise is that “early leaving could be an attempt for independence or a signal of frustrating cohabitation with parents.”

On the other hand, early leaving might result from positive parent–child relationships as well. A close, functional, and secure parent–child relationship might speed adolescent development by promoting skills for autonomy. Accordingly, supportive, active, and nurturing parent–child relationships might facilitate residential independence.

Associational Solidarity

Associational solidarity refers to the type and frequency of contact between intergenerational family members. Young adults might be less likely to leave the parental home when the family interacts frequently and engages in activities together. It stands to reason that young adults would appreciate—and capitalize on—living in a parental home when they experience a satisfying and active relationship with their coresident parents. At the same time, inactivity might motivate individuals to move out sooner, providing a corrective to unfriendly or distant parent–child interactions or ennui.

Normative–Functional Solidarity

A hybrid term, normative–functional solidarity refers to expectations regarding filial obligations and parental obligations and the degree of emotional and instrumental support exchanged among family members (Bengtson, 2001, p. 8). Specifically, parenting style is important because it is the basis for family roles and obligations that entail support and exchange between parents and their children. Indeed, a long line of research has shown that parenting styles are associated with adolescent development (Baumrind, 1967); they may affect young adults' departures and returns to the parental home as well.

Normative–functional solidarity may also be conceived as parental attentiveness, whereby parents obtain knowledge of their child's everyday activities and ingrain themselves in their children's lives. Regarding parental attentiveness, more is not always necessarily better. For example, excessive monitoring can hinder the development of autonomy and skills needed for the transition to adulthood (Fingerman et al., 2012).

Children with authoritarian or hypervigilant parents might be more likely to leave the parental home, and less likely to return, than those with other types of parents. On the other hand, overattentive and demanding parents might stifle young adults' personal autonomy, leading them to feel trapped or pressured to remain in the parental household. How the different dimensions are associated with leaving and returning to the parental home is not entirely clear; however, we do not necessarily expect the same mechanisms for leaving home to apply to returning.

Within contingency theory (Eggebeen & Davey, 1998), parents and grown children provide assistance to one another in response to specific needs. This perspective argues that intergenerational relationships are adaptive and parents will be responsive to difficult life circumstances regardless of the quality of the parent–child relationship. In this case, returning to the parental household might be warranted when it is triggered by necessity, such as housing issues, financial need, and separation or divorce (Fingerman, Miller, Birditt, & Zarit, 2009; Swartz, Kim, Uno, Mortimer, & Bengtson O'Brien, 2011; Smits 2010). Accordingly, we expect the relationship between intergenerational solidarity and leaving to differ from that between intergenerational solidarity and returning. However, we did not have a priori expectations about how these differences would be borne out.

gendered parent–child dyads

There are gender differences in leaving and returning to the parental home. On average, females leave the parental home sooner than males (Sandberg‐Thoma, Snyder, & Jang, 2015), partly due to gendered age differences for cohabitation and marriage. Some have speculated that another reason females leave home sooner might be because their autonomy is restricted by parental rules, regulations, and traditional expectations (White, 1994). Along the same lines, parental expectations regarding parent–child coresidence are different for sons than daughters. For example, Sassler, Ciambrone, and Benway (2008) found that parents attempted to control their coresidential daughters' social and dating behaviors but placed more emphasis on sons' financial contributions to the household.

Intergenerational relationships vary not only by daughters versus sons, but by mothers versus fathers. Females are generally higher on expressiveness, and males are higher on instrumentality (Bem, 1974). However, intergenerational relationships are rarely parsed out by parents' gender, with research usually focusing on the mother–child bond as the closer of the two. Still, child developmental research has long highlighted the importance of examining the unique parenting contributions of mothers and fathers individually (e.g., Zervides & Knowles, 2007) because each parent plays a separate yet significant role in shaping family interpersonal dynamics.

Furthermore, the gender composition of parent–child pairs may reveal further variations in patterns of intergenerational solidarity. Because the same‐sex intergenerational bond is reportedly closer than the cross‐sex parent–child bond (e.g., Hoeve et al., 2009), the relationship between parent–child relationships and leaving/returning home might be stronger for same sex parent–child relationships than cross‐sex relationships. Thus, in assessing how these characteristics are associated with departures and returns to the parental home, we also investigate potential differences in parent–child same‐gender and different‐gender dyads and the gender composition of the parent–child relationship. In other words, we assess whether gender moderates the association between the parent–child relationship and leaving or returning to the parental home across different dimensions of intergenerational solidarity.

control variables

Individual

There are important individual‐level differences in leaving and returning to the parental home. Foreign‐born young adults stay in the parental home longer than their native‐born counterparts (Van Hook & Glick, 2007). Young adults who experience emotional distress leave the parental household earlier (Sandberg‐Thoma et al., 2015). Linking back to the intergenerational solidarity perspectives, these researchers speculated that it “may be that distressed youth have difficulty maintaining a high‐quality relationship with parents” (Sandberg‐Thoma et al., 2015 p. 814). Physical health likely impacts individuals' initial departure from the parental home because the parental household might serve as a resource for needed assistance among youth in poor health.

Having a job allows young adults to live independently of their parental household and individuals with higher income are apt to move out sooner (Mulder & Clark, 2000). Employment and education are linked to leaving the home earlier and lower likelihood of returning (Sassler et al., 2008). There are also racial and ethnic differences in leaving and returning to the parental home (Lei & South, 2016). Young adults tend to leave the home after marriage to live with their partners or prior to marriage to cohabit (South & Lei, 2015). On the other hand, union dissolution is linked to returns to the parental home (Stone, Berrington, & Falkingham, 2014). Becoming a parent increases the likelihood of moving out of the parental home (Sandberg‐Thoma et al., 2015).

Parent and Household

Household and family‐level factors also play an important part in young adults' initial residential decisions. Resource constraints can tie individuals to their parental households, especially during times of personal economic hardships or large‐scale recession (Payne & Copp, 2013). Parent income creates a “feathered nest” and research has shown that, regardless of income, parents are more likely to provide support to coresidential than noncoresidential children (van den Berg, Kalmijn, & Leopold, 2018). Conversely, parent resources can also subsidize young adults' independent living (Iacovou, 2010).

Family structure is an important component, with two‐biological‐parent households often viewed as a social resource that provides a strong support system for children. Research has shown that children from blended families move out earlier and are less likely to return than those from families with two biological parents (e.g., Blaauboer & Mulder, 2009). Household size might impact individuals' decision to leave the parental home: Recent research found that the number of individuals coresiding in the parental household was negatively associated with risk of departure (Sandberg‐Thoma et al., 2015).

Geographic and Temporal Contexts

Within the life course perspective, life events and their consequences differ based on time, place, and developmental stage in which they take place (Elder, 1998). For example, young adults were less likely to leave home and more likely to return home during the Great Recession (Qian, 2012; South & Lei, 2015). Research has also indicated that young adults from the Midwest are slower to leave home than other regions of the United States (Goldscheider & Goldscheider, 1999).

hypotheses

Our hypotheses focus on the understudied potential relationship between intergenerational solidarity and young adults' leaving and returning to the parental home as well as the moderating effects of gender, net of more traditional individual‐ and household‐level effects.

Early intergenerational relationships, measured by affectual, associational, and normative–functional solidarity, will be associated with the timing of young adults' departure from the parental home.

Higher degrees of parent–child solidarity will be associated with delayed departure from the parental household (solidarity–retention hypothesis).

Higher degrees of parent–child solidarity will be associated with earlier departure from the parental household (solidarity–facilitation hypothesis).

The relationship between intergenerational solidarity and leaving will differ from that between intergenerational solidarity and returning.

The relationship between intergenerational solidarity and leaving and returning to the parental home will be conditioned by the gender composition of the parent–child relationship.

method

National Longitudinal Survey of Youth 1997 (NLSY97)

Analyses are based on the National Longitudinal Survey of Youth 1997 (https://www.nlsinfo.org/content/cohorts/nlsy97). The full sample of 8,984 adolescents were ages 12 to 17 in the initial 1997 wave (born between 1980 and 1984). The rich time‐series data capture adolescents' life events prior to and through the transition to adulthood and into adulthood. Collected from 1997 to 2015, the panel data allow analyses of factors in adolescence that predict the timing of departure from the parental home in the years that follow.

As of 2015, there have been 17 rounds of data collection that started when the birth cohort was 12 to 18 years old and were aged 30 to 36 in 2015. The first wave (1997) included measures for parent–child relationship dynamics, adolescent characteristics, and parent‐level and household‐level measures. Additional data collected in subsequent waves provide time‐varying information about leaving and contemporaneous independent variables. Given the complex design, a custom longitudinal weight ensured that the sample was nationally representative.

Sample

Owing to age restrictions on the variables of interest (the NLSY did not collect data on parent–child relationships from children older than age 15), this study focuses on only a subset of adolescents age 12 to 15, the youngest ages of the NLSY97 cohort (N = 5,402). We further removed all individuals who left the parental home before age 16 or prior to the initial interview (n = 201). Of those in the final sample (N = 5,201), 87% (n = 4,519) left the parental home between 1997 and 2015, starting at age 16 and ending at age 35. Of the 4,519 young adults eligible to return home, 2,749 (61%) did so. This yielded a full sample of 88,417 person‐year observations.

A chained multiple imputation procedure was used to handle missing data. This procedure uses the dependent and independent variables to impute missing data on the independent variables only (Allison, 2002). Descriptive statistics and parameter estimates were virtually identical for each imputed dataset.

measures

Dependent Variables

Departures from and returns to the parental home. Young adults' initial departure from the parental home is based on residential history data collected at each wave of the NLSY97, starting in 2003 and ending in 2015, the last currently available wave of data. Respondents who left the parental home prior to 2003 reported retrospectively on the date they left starting in 2003. Those who left the parental home were asked if and when they moved back into the parental household for at least 3 months. The information is then updated at each wave to capture departures and returns that occurred between subsequent waves. To reduce reporting bias associated with shorter term living arrangements (e.g., college students returning home for the summer), the respondents were prompted with the following definition of permanent living arrangements: “Sometimes people live in places temporarily while attending school or working a job or for some other reason, but they consider their permanent residence to be elsewhere. Do you consider the place you are currently living to be your permanent house?”

Independent Variables

Affective solidarity: parent–child closeness. Following Hair, Moore, Garrett, Ling, and Cleveland (2008), we use a validated, parent–child closeness scale familiar to youth outcome research (Day, Kaye, Hair, & Moore, 2009). Adolescents reported on feelings toward each coresidential parent separately in the first wave (1997). With responses from 0 (strongly disagree) to 4 (strongly agree), the respondents reported on the following statements: “I think highly of her,” “She is a person I want to be like,” and “I really enjoy spending time with her.” With possible responses from 0 (never) to 4 (always), the respondents also reported on the following: “How often does she criticize you or your ideas?” (reverse coded), “How often does she praise you for doing well?” “How often does she blame you for her problems?” (reverse coded), “How often does she make plans and cancel with you for no good reason?” (reverse coded), and “How often does she help you do things that are important to you?” The items are summed to create a 0 to 32 scale for parent–child emotional closeness (α = .73 for mothers and .83 for fathers). For both scales, higher scores indicate a closer parent–child relationship. We tap into the respondents' reported closeness to their father using the same validated eight‐question scale.

Family routines. Behavioral solidarity was a sum of the responses to how many days in a typical week the adolescent reported (a) having dinner with the family; (b) doing something fun as a family, such as playing a game, going to a sporting event, went swimming, and so forth; (c) doing something religious as a family, such as going to church, praying, or reading the scriptures together; or (d) getting the housework done when it is supposed to (e.g., cleaning up after dinner, doing dishes, or taking out the trash). The resulting index, which is high in predictive validity (Center for Human Resource Research, 2003) ranges from 0 to 28, with higher scores indicating more behavioral solidarity.

Attentiveness. A validated four‐item measure (Center for Human Resource Research, 2003), taken in 1997, assessed each coresidential parent's attentiveness by asking the child how much their parent knew about the following: (a) the child's close friends, (b) their child friends' parents, (c) who the child is with when the child is not at home, and (d) who the child's teachers are and what they are doing in school. Response options vary from 0 (knows nothing) to 4 (knows everything). The summative scale ranges from 0 to 16, with higher scores indicating more parental attentiveness (α = .75 for mothers and .82 for fathers).

Parenting style. Parenting style can be classified according to a prominent fourfold typology (Baumrind, 1967; Maccoby & Martin, 1983) that is widely used in parenting research. A categorical measure of parenting style, high in both construct and predictive validity (Center for Human Resource Research, 2003), assessed the interactive effect between each parent's demandingness and supportiveness (Bronte‐Tinkew, Scott, & Lilia, 2010; Maccoby & Martin, 1983).

In 1997, young adults responded to an item about whether they considered their mother and father “very supportive, somewhat supportive, or not very supportive” and a separate item asked whether they considered each parent “permissive or strict about making sure you did what you were supposed to do.” For responsiveness, “very supportive” responses are coded 1, else 0. For demandingness, “demanding” responses are coded 1, else 0. Combined, the variables create a two‐by‐two typology of parenting style: authoritative (demanding and supportive), authoritarian (demanding and not very supportive), permissive (nondemanding and very supportive), and uninvolved (nondemanding and not supportive).

Demographic. Individual‐level variables include the respondent's age and gender (“female” = 1, “male” = 0). A measure marked whether the respondent was born in the United States (0) or was foreign born (1). Based on NLSY97 measurement, race/ethnic dummy variables include mixed race, Black, Hispanic, or the omitted reference “non‐Black and non‐Hispanic” (hereafter referred to as “White”). The respondent's education is a time‐varying measure of highest grade completed. A time‐varying measure captures the respondent's enrollment status as (0) not enrolled, (1) enrolled in K‐12, or (2) enrolled in college or graduate school. Employment status is a time‐varying measure which indicates whether the individual was (0) unemployed or (1) employed for at least half of the year prior to the survey. A time‐varying dichotomous variable marks whether the young adult's income was at or below the poverty line in a given year (1), else (0).

Health. Respondent overall health is assessed with a time‐varying self‐reported ordered scale with options for the following: 1 = “poor,” 2 = “fair,” 3 = “good,” 4 = “very good,” 5 = “excellent.” Mental health was measured in 2000 with a validated five‐item scale. The respondent reported the frequency of symptoms with the response options 1 = “all of the time,” 2 = “most of the time,” 3 = “some of the time,” 4 = “none of the time.” “How much of the time during the last month have you…”: 1 = “been a very nervous person?” (reverse coded), 2= “felt calm and peaceful?” 3 = “felt downhearted and blue?” (reverse coded), 4 = “been a happy person?” 5 = “felt so down in the dumps that nothing could cheer you up?” (reverse coded). The summative scale ranges from 5 to 20, with higher scores indicating better mental health (α = .79).

Family life course transitions. Marital status change indicated whether there was (0) no marital change or the respondent (1) entered cohabitation, (2) got married, or (3) became unmarried. Change in parental status marked whether (0) no parental change occurred or (1) the respondent became a parent between waves.

Parent and household. Household structure in adolescence is a categorical variable distinguishing living with two biological parents (omitted reference), a biological parent and stepparent, a single mother or single father, or other. The “other” category includes those residing with foster parents, adoptive parents, grandparents, other relatives, or other persons in 1997. Household size is a time varying measure of the number of individuals living in the respondent's household in a given year. Parent‐level resources are measured with parent's household income level in 1997 (logged).

Geographic and temporal context. Urban residence was a time‐varying covariate indicating whether the adolescent lived in an urban or suburban area (1) as opposed to a rural area (0). Region consists of the following four classifications for U.S. region: (a) Northeast, (b) Midwest, (c) South, and (d) West. To tap into temporal context, a variable flagged whether or not a given survey year occurred between 2007 and 2009, the peak years of the Great Recession.

Analytic Strategy

Following recent research on exiting and returning to the parental home using the NLSY97 (Sandberg‐Thoma et al., 2015; Warner & Houle, 2018), the dependent variables—initial departure from the parental home and returns to the parental home—were modeled individually with Cox proportional hazard models. These models express the “hazard rate” of moving from the parental home when assumptions about proportional hazards have been violated.

The hazard represents the rate that a departure occurs at time t given that it had not yet occurred in prior waves. For the first series of models—initial departures, adolescents enter the “risk set” in 1997, the first year of NLSY97 interviews. The respondents then remain in the risk set until the final wave and are censored unless they move out of the parental home. Respondents who moved out of the parental home between 1997 and 2013 (n = 4,519) enter the risk set for returning to the parental home. The respondents remain in this subsequent risk set until 2015 unless they move back into the parental home before the observation period ends; otherwise, they are censored.

Model Diagnostics and Presentation

For all multivariate analyses, variance inflation factors indicated there was no severe multicollinearity in the models. Analysis of the correlation matrix (not shown) indicated that none of the observed relationships between the independent variables in the models were very strong. All results were weighted and corrected for the complex NLSY survey sampling design. (Descriptive statistics are provided as supplemental information files.)

The multivariate models proceed in three steps. Following recent research on intergenerational solidarity (Hank & Steinbach, 2018), we present all models separately for mothers and fathers. This obviates issues with multicollinearity, which occurred when all measures for both parents were included in a single model. However, it should be noted that this modeling strategy might also overestimate—or potentially even obscure—some of the effects of parent–child relationships on leaving or returning home.

Table 1, Model 1.1 presents Cox proportional hazard results on the relationship between adolescent mother–child relationship dynamics on leaving, controlling for individual‐level characteristics, family life course transitions, parent and household‐level characteristics, and geographic and temporal context. Model 1.2 includes interactions terms for family and mother–child relationship dynamics and gender. In the second panel of Table 1, the same analyses are presented for fathers in Models 2.1 and 2.2. Table 2 presents the results for the relationship between adolescent family and parent–child relationship dynamics on returns to the parental home for mother–child (Models 3.1–3.2) and father–child (4.1–4.2) pairs. To conserve space, the results for father relationships will be discussed in the text only when they differ in terms of statistical or practical significance from those for mothers. For added information about the statistically significant interaction effects, hazard ratios (HRs) are presented in the text but not the tables.

Table 1

Cox Proportional Hazard Model for Departures from the Parental Home (N = 5,201)

	Mother		Father
Variables	Model 1.1, controls	Model 1.2, interactions	Model 2.1, controls	Model 2.2, interactions
Individual
Female	0.09*	0.16	0.09*	0.03
Age	−0.13*	−0.13*	−0.18***	−0.19***
Foreign born	−0.24*	−0.24*	−0.25*	−0.25*
Race/ethnicity
White (reference)
Black	−0.17***	−0.17***	−0.14**	0.14**
Hispanic	−0.13*	−0.13*	−0.12*	−0.13*
Other	0.09	0.09	0.04	0.05
Highest grade	0.01	0.02	0.01	0.01
Enrollment status
Not enrolled (reference)
Enrolled in K–12	−1.54***	−1.54***	−1.65***	−1.65***
Enrolled in higher education	0.14***	0.14***	0.19***	0.19***
Employed	0.09*	0.09*	0.10*	0.10*
Poverty	0.26***	0.26***	0.23***	0.23***
Physical health	−0.04**	−0.04**	−0.05**	−0.05**
Mental health	−0.02**	−0.02**	−0.02*	−0.02*
Life course transitions
Change in marital status
No change (reference)
Started cohabiting	0.78***	0.78***	0.76***	0.76***
Got married	0.51**	0.51**	0.48*	0.48*
Stopped cohabiting/unmarried	0.30	0.30	0.38	0.38
Change in parental status	0.26*	0.27*	0.23	0.24
Parent and household
Household structure 1997
Both biological parents (reference)
Biological parent and stepparents	0.21***	0.21***	0.22***	0.22***
Single mother/father	−0.08†	−0.08†	−0.06	−0.05
Other living situation	−0.01	−0.01	0.07	0.07
Household size	−0.27***	−0.27***	−0.28***	−0.33***
Household income 1997	0.00	0.00	0.01	0.01
Geographic and temporal
Region
Northeast (reference)
Midwest	0.13*	0.13*	0.14*	0.14*
South	0.08†	0.08†	0.09†	0.09†
West	0.03	0.03	0.03	0.03
Urban	0.07	0.07	0.11*	0.08†
Recession	−0.11	−0.11	−0.17	−0.12
Parent–child relationship
Mother–child closeness	0.01*	0.01
Father–child closeness			0.01†	0.01
Mother attentiveness	0.01	0.01†
Father attentiveness			−0.00	0.01
Family routines	0.00	0.00	0.01	0.01
Mother parenting style
Authoritative (reference)
Uninvolved	−0.06	−0.08
Permissive	−0.04	−0.06
Authoritarian	0.03	0.04
Father parenting style
Authoritative (reference)
Uninvolved			0.04	0.07
Permissive			−0.11*	−0.10†
Authoritarian			0.01	0.04
Interactions
Mother–Child Closeness × Female		0.00
Father–Child Closeness × Female				0.01
Mother Attentiveness × Female		−0.02
Father Attentiveness × Female				−0.03*
Family Routines × Female		0.00		0.00
Mother parenting style
Authoritative (reference)
Uninvolved × Female		0.03
Permissive × Female		0.05
Authoritarian × Female		−0.03
Father Parenting Style × Female
Authoritative (reference)
Uninvolved × Female				−0.08
Permissive × Female				−0.01
Authoritarian × Female				−0.05
F	45.3***	41.7***	52.2***	46.5***

Source. National Longitudinal Survey of Youth 1997 (imputed and weighted data, corrected for design effects).

Note. Coefficients presented (not hazard ratios).

***p < .001. **p < .01. *p < .05. †p < .10.

Table 2

Cox Proportional Hazard Model for Returns to the Parental Home (N = 4,519)

Variables	Model 3.1, controls	Model 3.2, interactions	Model 4.1, controls	Model 4.2, interactions
Individual
Female	0.01	−0.38	0.02	−0.46†
Age	−0.24***	−0.25***	−0.30***	−0.30***
Foreign born	−0.21*	−0.20†	−0.24*	−0.24*
Race/ethnicity
White (reference)
Black	−0.10*	−0.10*	−0.10	−0.11
Hispanic	−0.19**	−0.20**	−0.20**	−0.20**
Other	0.18	0.19	0.20	0.22
Highest grade	−0.03†	−0.03†	−0.02	−0.01
Enrollment status
Not enrolled (reference)
Enrolled in K–12	−0.95***	−0.95***	−0.93***	−0.93***
Enrolled in higher education	0.20***	0.21***	0.22***	0.22***
Employed	−0.21***	−0.21***	−0.24***	−0.24***
Poverty	0.08	0.08	0.07	0.07
Physical health	−0.06***	−0.06***	−0.07**	−0.07**
Mental health	−0.03***	−0.03***	−0.02*	−0.02*
Life course transitions
Change in marital status
No change (reference)
Started cohabiting	0.15	−0.15	0.22	−0.23
Got married	−0.78†	−0.78†	−0.61	−0.61
Stopped cohabiting/unmarried	−0.37	−0.36	−0.22	−0.21
Change in parental status	0.04	0.04	−0.08	−0.07
Parent and household
Household structure 1997
Both biological parents (reference)
Biological and stepparents	0.01	0.01	0.03	0.04
Single mother/father	−0.07	−0.07	0.00	0.00
Other living situation	−0.37	−0.05	0.06	0.05
Household size	0.13***	0.13***	0.13***	0.13***
Household income 1997	0.00	0.00	0.00	−0.00
Geographic and temporal
Region
Northeast (reference)
Midwest	0.07	0.06	0.03	0.02
South	−0.09	−0.09	−0.12†	−0.13†
West	−0.12†	−0.12	−0.12	−0.13
Urban	−0.09†	−0.09†	−0.12*	−0.12*
Recession	−0.12	−0.12	−0.13	−0.12
Parent–child relationship
Mother–child closeness	0.01	−0.01
Father–child closeness			0.01	−0.01
Mother attentiveness	0.01	0.02*
Father attentiveness			0.00	0.01
Family routines	−0.01	−0.01	0.00	0.00
Mother parenting style
Authoritative (reference)
Uninvolved	−0.19*	−0.19†
Permissive	−0.09*	−0.09
Authoritarian	−0.06	−0.10
Father parenting style
Authoritative (reference)
Uninvolved			−0.20*	−0.13
Permissive			−0.15**	−0.10
Authoritarian			0.00	−0.06
Interactions
Mother–Child Closeness × Female		0.03**
Father–Child Closeness × Female				0.03**
Mother Attentiveness × Female		−0.04*
Father Attentiveness × Female				−0.01
Family Routines × Female		0.00		−0.01
Mother parenting style
Authoritative (reference)
Uninvolved × Female		0.00
Permissive × Female		−0.01
Authoritarian × Female		0.06
Father Parenting Style × Female
Authoritative (reference)
Uninvolved × Female				−0.11
Permissive × Female				−0.10
Authoritarian × Female				0.08
F	10.1***	9.37***	10.2***	10.1***

Note. Coefficients presented (not hazard ratios). National Longitudinal Survey of Youth 1997 (imputed/weighted data, corrected).

***p < .001. **p < .01. *p < .05. †p < .10.

results

Departures From the Parental Home

Cox proportional hazard models in Table 1 lend some support to theories of intergenerational solidarity and leaving the parental home. Models 1.1 and 2.1 indicated an association between youths' affectual solidarity (i.e., closeness) with their mother and father, respectively, and earlier exit from the parental home, controlling for a wide range of individual, parent, and household covariates. Normative–functional solidarity was a significant dimension for fathers only, with permissive fathers associated with lower likelihoods of leaving. Only one of the two normative–functional measures was significant in the gender interactions model (Model 2.2), with father's attentiveness negatively associated with leaving—the association is significantly different for male than female children. This provides some, albeit only slight, evidence of moderating effects of gender on solidarity and leaving outcomes.

Individual‐level characteristics were associated with leaving in the expected directions. Family life course transitions were also associated with the initial move out of the parental household. Youth who enter into cohabitation arrangements (b = 0.78, p < .001) or marriage (b = 0.51, p < .01) were more likely to move out than those without a change in marital/cohabitation status. Those who have their first child (b = 0.26, p < .05) were somewhat more likely to leave the parental home than those without a change in parental status.

A number of parent and household‐level factors were significantly associated with leaving the parental home. Consistent with past research (e.g., Iacovou, 2010), when compared with young adults who lived with both biological parents in 1997, those who lived with a parent and stepparent were more likely to leave the parental home (b = 0.21, p < .001). In addition, those residing in more‐populated households reported delayed departures from home (b = −0.27, p < .001).

Geographic context—namely U.S. region of residence—was also associated with departures from the parental home. Youth residing in the Midwest were more likely to move from the parental home than those from the Northeast (b = 0.13, p < .05). The results presented in Table 1, Model 2.2 provide some support for Hypothesis 3. Gender composition of the parent–child dyad moderates the association between intergenerational relationships and leaving, specifically for father–daughter dyads and attentiveness (HR = 0.97, p < .05).

Returns to the Parental Home

Table 2 provides the results of the Cox proportional hazard models for returning to the parental home. Again, the results lend some support for both theories of intergenerational solidarity and life course, indicating some expected moderate associations not only with leaving but also returning to the home. When the dependent variable was risk of returning to the parental home, measures of affectual solidarity were no longer salient; rather, normative–functional solidarity, operationalized as parenting style, exhibited a significant association, controlling for all individual, parent, and household controls. Specifically, uninvolved parenting style became a salient influence, whether characterizing mothers (Model 3.1, b = −0.19, p < .05) or fathers (Model 4.1, b = −0.20, p < .05). The results also indicated that returns were somewhat less common among young adults with permissive parents (mothers in Model 3.1, b = −0.09, p < .05; fathers in Model 4.1, b = −0.15, p < .01) when compared with authoritative parents.

In Models 3.2 and 4.2—the interaction models for gender and the mother–child and father–child relationships, respectively—there was a significant and positive coefficient for either parent's affectual solidarity with their child (HRs = 1.03, ps < .01) as well as a negative coefficient for mother's attentiveness (HR = 0.96, p < .05). These results thus lend some additional support to our second and third hypotheses, the moderating influence of child's gender on solidarity covariates, which was not observed in Table 1 on leaving.

Auxiliary and Sensitivity Analyses

Auxiliary analyses included interactions between each of the parent–child relationship variables but were excluded from the main analyses because the results were neither statistically nor practically significant. Sensitivity analyses were run using discrete‐time event history models, which have a different series of assumptions about the data. These models have been used in recent research on the topic (Lei & South, 2016; South & Lei, 2015). The substantive results for parent–child relationships and leaving and returning to the parental home did not differ substantially from the Cox Models presented here.

discussion

Leaving Home

The first set of Cox proportional hazard models, focused on youth's first departures from the parental home, provided support for Hypothesis 1 in general by showing that measures of intergenerational solidarity are significantly associated with earlier leaving, net of all control variables. Specifically, the results showed some support for Hypothesis 1.2 (solidarity–facilitation hypothesis), which suggests higher levels of solidarity may instill feelings and capabilities of autonomy that better equip youths to achieve independent residence and are thus associated with earlier departures from the parental home. One explanation for these findings might be linked to self‐determination theory.

Within self‐determination theory (Deci & Ryan, 2000), children act on their interests and values as a direct result of parental support of their autonomy. This suggests that in interdependent, intergenerational social contexts, young adults will be more likely to engage in autonomous behaviors if their parents show an interest in their attitudes and values (see Brenning et al., 2015; Villacorta, Koestner, & Lekes, 2003). Thus, close parent–child relationships can help launch children into adulthood.

The results of this first set of Cox models also demonstrate that in predicting the timing of leaving, affectual dimensions of parent–child solidarity (i.e., closeness) may be more salient than associational (i.e., family routines) or normative–functional (i.e., attentiveness and parenting styles). Although closeness with either parent was associated with earlier leaving, the normative–functional measure of parenting style was significant for fathers only; furthermore, the measure of attentiveness was significant only for father–daughter dyads. These findings highlight the importance of the more‐nuanced view of intergenerational relationships brought to the current study. Parent–child relationships are associated with departures from the parental home—and normative–functional solidarity (parenting style) is particularly important for father–child relationships.

Gender interactions exhibited one significant relationship between solidarity and leaving, thus lending some support to our third hypothesis. Although gender interactions yielded no new associations for mother–daughter dyads, the interaction term for father–daughter parental attentiveness introduced a significant negative association with leaving. Thus, gender composition does seem to play some role in decision‐making about residence, but this role is perhaps more nuanced than originally expected.

Returning Home

The second set of models, in which the focus switched to returns to the parental home, provided support for Hypothesis 2, as different dimensions of intergenerational solidarity came into play. Whereas affectual solidarity and to a lesser extent normative–functional solidarity (measured as parent–child closeness and parent attentiveness or parenting style, respectively) were significantly associated with leaving, no significant associations appeared between these measures and the phenomenon of returning to the parental home (with the exception of mother's attentiveness, with a significant positive association in Model 1.2 only). This falls in line with contingency theory, which proposes that, regardless of the parent–child relationship, parents may act as safety nets for their young adult children by offering aid that minimizes the negative effects of housing instability, financial crises, or other life difficulties.

However, normative–functional measures of parenting style took on significance, specifically in associating uninvolved and permissive parenting styles with decreased likelihood of return to the parental home when compared with youth that have authoritative parents. Returners may thus have been discouraged by a perceived lack of structure in the parent–child relationship that countered potential benefits of return, such as parents' help in compensating for low employment or health problems experienced by their children. This support for Hypothesis 2 only points to a need to further investigate the complex array of intergenerational solidarity effects that may or may not complement more traditional predictors of returns to the parental home.

Table 2 models also show some support for Hypotheses 2 and 3, showing a few significant associations between solidarity–gender interactions and return to the parental home, associations that did not appear salient to leaving decisions. In particular, mother–daughter and father–daughter closeness (our measures of affectual solidarity) exhibited significant associations with returning (interestingly, in different directions: negative for mother–daughter pairs and positive for father–daughter pairs). This tempers the earlier conclusion that affectual solidarity (i.e., parent–child closeness) fades from the picture of children's return to the parental home, suggesting support that the measure retains importance for daughters' decisions to return.

This finding may reflect an accord with traditional gender role ideology, by which adulthood means different things for males versus females: Adulthood for men focuses on the public sphere and work; for women, the focus has historically been the private sphere of managing family life (Hogan & Astone, 1986). Indeed, recent work finds that childbearing increases the risk of leaving the home more for women than it does for men (Goldscheider, Hofferth, & Curtin, 2014). Returning to the parental home is often conceived as a way of ameliorating lost resources, typically financial, but as womanhood is constructed around profession and family, a return to the family home may be a way of supplementing a family life that falls short of the perceived norm. The more women feel strong affectual ties toward their parents, the more likely it will make returning a viable substitute to a delayed or failed family life outside the parental home. Although these assertions are only speculative, this would be an interesting avenue for future research to build on this study.

In addition, there is some support for the idea that the same “mother attentiveness” measure of normative–functional solidarity associated with earlier leaving may also be associated with decreased likelihood of returning among female children. This falls in line with the previous suggestion of return as accomplishing women's adulthood obligations as family managers: A return to the parental home is only a viable supplement for lacking family life if the young woman is accorded a new role of “cohead of household,” not the old role of “daughter,” which mother attentiveness implies. Father attentiveness may seem acceptable to a greater degree, due simply to traditional patriarchal gender norms, and thus not factor into a young woman's return, but being monitored by one's mother may deny young women the desired and expected experience of adult family life, and thus make the return less likely. Such varied findings across dimensions of solidarity, leaving and returning outcomes, and parent–child gender compositions all point to the value of this study's more‐nuanced view of the increasingly complex process of youths' transitions to adulthood and decision‐making about (co)residence.

Finally, the significant associations of many individual, parent, and household characteristics with leaving and returning outcomes only serve to highlight the complexity of these phenomena and the ongoing profusion of alternative pathways to adulthood. Most were consistent with expectations based on previous research. Overall, these results suggest the need for a more thorough understanding of how parent–child relationships, and in particular gendered parent–child relationships, fit into this array of significant factors rather than a limited focus on individual and household‐level characteristics.

Limitations and Directions for Future Research

These results might still reflect some selection bias. Psychological factors, such as being a well‐adjusted child, might select children into close parent–child relationships as well as early leaving. In support of this, research has shown that parents are more likely to invest in children they classify as “deserving” (Fingerman et al., 2009). We did control for young adult mental health in 2000, but this measure was not optimal because it tapped into the mental health status of youth who had already left the parental home (between 1997 and 2000).

One reason for the null results for the family routines measure of associational solidarity is that the measure might not actually include either or both parents because the items refer generally to the entire family unit. We do not have information on whether the young adults' father or mother were involved in any or all of the routines. Another reason might be that individuals who engage in frequent activities with their families can still do so if they remain nearby when they do leave home. An interesting avenue for future research would be to explore the specific ways in which families enact behavioral solidarity and what this means for young adults' relocation decisions.

This is a single birth‐cohort only. The results might differ across different cohorts. In addition, we were only able to draw on time‐invariant childhood family characteristics, such as parent resources and other information about parents, which were measured in the first round of parent interviews. For example, we treated parenting style as fixed, but research indicates that it can change in relation to stressful events (Gillespie, 2015), and therefore our measures might not capture parenting immediately prior to the young adult's departure from the parental home. This possibility aligns with recent models of domain‐specific parenting, which argue that parenting styles are not constant but situationally determined, where parents enact different parenting methods in different situations and contexts (Smetana, 2017). Of course, parenting styles might also change over time as a direct reaction to children's prolonged coresidence in the parental home.

conclusion

The process of transitioning to adulthood is becoming more complex for contemporary youths than in previous generations, involving more heterogeneous pathways as well as multiple levels of influencing factors, from individual to household to social and historical characteristics coming to bear on the way young people navigate leaving and, increasingly, returning to the parental home. Recent research has begun to tap into many of these characteristics, but there is still much to add to this picture.

This study provides empirical support for a deeper investigation of two sets of factors that may be linked to children's decisions to leave and return to the parental home. First, we find support for a multidimensional view of intergenerational solidarity in the significant associations between some solidarity variables from each of the solidarity categories (affectual, associational, and normative–functional) and either leaving, returning, or both. Second, these findings also support a gendered view of intergenerational solidarity, specifically as pathways appear to diverge for women versus men in models of both leaving and returning to the parental home.

Our multidimensional view of intergenerational solidarity suggests that some aspects of solidarity are associated with leaving and returning decisions more than others. Measures of affectual solidarity were significantly associated with leaving, suggesting that parent–child closeness prepares children to leave the parental home earlier. On the other hand, measures of normative–functional solidarity were significant in models of returning, suggesting that children are less likely to return to households with permissive or uninvolved parents. This lends support and nuance to the idea that the factors that drive children out of the parental home may not be the same factors that bring them back (Goldscheider, Goldscheider, Clair, & Hodges, 1999); we elaborate on this idea by showing that specific dimensions of intergenerational solidarity that encourage leaving may not be the same that encourage return.

Our gendered view of solidarity indicates some significant differences between women and men in terms of leaving and returning to the parental home as well as across different parent–child gender dyads. Only in father–daughter dyads is attentiveness significant, with more attentiveness associated with a lower likelihood of leaving the parental home. Only for female children is returning still positively associated with parent–child closeness and also negatively associated with mother attentiveness. Females thus appear to apply different rubrics in deciding to return home than their male counterparts, continuing to value solidarity, albeit in different ways depending on which parent is in question. These findings only underscore the complexity involved in decisions to leave and return to the parental home and the varying effects of solidarity for female and male children and parents. Future research might elaborate on the nuanced models presented here to gain a deeper understanding of how parent–child solidarity and gender interact to shape children's trajectories out of, and sometimes back to, the parental home.

Note

The author thanks Clara Mulder, Stefan Malizia, and Janet Lever for their helpful advice and comments on earlier drafts of the manuscript. This article is based on research funded by the European Research Council under the European Union's Horizon 2020 research and innovation program (Grant 740113).

Table S1. Sample Characteristics: Mean (SD) or Proportion (n = 4,519)

Click here for additional data file.