1
The Transition to Adulthood among the Forgotten Half: Home Leaving and Living Arrangement
among Less Educated Young Adults in the United States
Anastasia R. Snyder
Sara E. Mernitz
The Ohio State University
Joy Bohyun Jang
University of Michigan
2
Abstract
College experiences during emerging adulthood are associated with various life course transitions for young adults, including leaving the family home and living independently, but little is known about how these patterns differ for those who do not attend college. Using data from the National Longitudinal Survey of Youth 1997, we examined the timing of first parental home-leaving and the type of first living arrangement for young adults who do not pursue post- secondary education, comparing young men and women. Life table estimates found that young women exit the family home sooner after finishing high school, and most exit to live with a romantic partner, motivated in large part by early pregnancy. Discrete-time competing risk models estimated separately by sex and found that, controlling for individual, family and contextual factors, female young adults exited earlier to live with a partner. Individual
characteristics and family context are significantly associated with the timing and type of home leaving.
Keywords
Educational difference, emerging adulthood, home leaving, living arrangement, National
Longitudinal Survey of Youth 1997
3
Introduction
The theory of emerging adulthood posits that socioeconomic and demographic changes in
recent decades have resulted in a new stage of human development – emerging adulthood—
which ranges from approximately age 18 until the late 20s (Arnett, 2000). During this period,
young adults spend an extended period of time self-focused as they explore their identity and
prepare for adulthood (Arnett, 2000). Yet, critics of this theory argue that structural forces limit
opportunities to explore and prepare for adulthood for many young adults, especially those from
lower socioeconomic classes (e.g. Bynner, 2005; Côté, & Bynner, 2008). Arnett (2016)
responded to these critiques concerning social class, operationalized by mother’s educational
attainment, by showing similarities across class in emerging adult perceptions of the time period,
views of love, sex, and marriage, and views of education and work. However, these conclusions may be premature given the reliance on cross-sectional data (Furstenberg, 2016) and empirical
evidence using longitudinal data from the Youth Development Study found different trajectories
by emerging adult educational status on the pathway to adulthood (Mitchell & Syed, 2015).
One area that is linked closely with the theory of emerging adulthood is home-leaving, or exiting the family home. Home-leaving is a key life course transition for young adults that has important implications for young adult independence and psychological development, family relationships, housing market dynamics, and population distribution (Arnett, 2006; Garasky,
Haurin, & Haurin, 2001; Kins & Beyers, 2010; Mulder & Clark, 2002; Sassler, Cramborne &
Benway, 2008). Broadly, the growing constraints in labor and housing markets in the U.S. have contributed to observed declines in young adult independent living, where the young adult is either the head of the household or the spouse, and a greater prevalence of young adults co-residing with parents (Qian, 2012; Yelowitz, 2007). However, the home-leaving process has yet to be examined among U.S. young adults who do not pursue a college education, or the ‘forgotten half’ (Halperin, 1998). Because these youth have different employment and 4 family formation experiences than those pursuing a college education during the young adult years (Arnett & Tanner, 2011), their home-leaving experiences likely differ as well.
Using longitudinal and geocode data from the National Longitudinal Survey of Youth
1997, we examined the timing of home-leaving and the type of living arrangement entered into upon first exit from the family home for those that do not attend post-secondary schooling. We address the following research questions:
1. What are the home-leaving and living arrangement patterns among young adults in
the United States who do not attend college after exiting the family home?
2. How do specific barriers faced by these youth impact the timing of home leaving
and type of living arrangements?
3. Do the home-leaving and living arrangement patterns for these youth differ by
gender? Do these barriers differ by gender?
Theory of Emerging Adulthood
The theory of emerging adulthood posits that macro-level social and economic changes have influenced human development leading to a longer transition to adulthood and a distinct developmental period termed ‘emerging adulthood’ (Arnett, 2000). Emerging adulthood begins after adolescence (age 18) and last until age 25 (Arnett, 2000), although recent demographic changes have increased the upper age boundary until the late 20s (Arnett, 2006). Five features are used to distinguish this developmental period from adolescence or young adulthood, including identity exploration, instability, self-focus, possibilities, and feeling in-between
(Arnett, 2006). These features of emerging adulthood align closely with pursuing a college education, which is now considered necessary to be competitive in the labor market (Danziger &
Ratner, 2010). In college, emerging adults are able to spend more self-focused time building their human capital and exploring their individual identities as they slowly transition into adulthood
(Arnette, 2006). This extended transition delays other life course events that were historic indicators of adulthood, such as marriage (Cherlin, 2010; Goldstein & Kenny, 2001) or home 5 ownership (Mulder & Clark, 2001). These historic indicators of adulthood are more permanent, and delaying these events provides emerging adults with opportunities to try new experiences
(Arnett 2006; 2007). Further, because these emerging adults have not achieved adult status, they
can still rely on their families for support, such as helping to fund the cost of independent living
(Aquilino, 2006).
A key criticism of emerging adulthood theory is that these features of emerging
adulthood are only relevant for more privileged groups, specifically those that are able to attend
college (Bynner, 2005; Côté, 2014; Côté & Bynner, 2008; Hendy & Kloep, 2007). Indeed,
increased time spend pursuing higher educational attainment is a key factor motivating the life
changes that have in theory resulted in the new emerging adult stage of life. Those that do not
attend college may not have opportunities to experience a longer transition to adulthood, or
explore their identity, preferences, and life goals in a setting that aids in that development;
rather, these youth assume adult roles and responsibilities earlier because of their lack of
economic resources (e.g. Silva, 2013). Empirical evidence has identified groups of youth in the
U.S. who do not seem to experience the emerging adult developmental stage both quantitatively
(Osgood, Ruth, Eccles, Jacobs, & Barber, 2005; Silva, 2012) and qualitatively (Carr & Kefalas,
2011).
While those in college can postpone adulthood for exploration purposes (Arnett, 2000;
Schwartz, Côté & Arnett, 2005), any delayed transition to adulthood observed for those with no
college attendance is likely due to structural barriers, such as constraints in the labor market and
lack of economic opportunities (Hamilton & Hamilton, 2006). While the longer transition may
be observed for both college attending and non-college attending emerging adults, those with no
college attendance experience this adult postponement without the hope of future possibilities
(Hamilton & Hamilton, 2006), which is a key distinction that likely has developmental
implications. Further, traditional pathways to adulthood that were previously available to non-
college bound youth are now blocked, due to limited economic opportunities and the 6 disappearance of social norms and structures that have guided this process in the past (Côté &
Bynner, 2008). Thus, youth who do not attend college may be reacting to structural barriers rather than delaying adulthood due to individual agency.
Home-leaving Patterns
Educational attainment is considered a key factor influencing the pathway to adulthood for young adults (Hamilton & Hamilton, 2006; Sandefur, Eggerling-Boeck & Park, 2005;
Turner, 2004). Compared to those who attend college, young adults who do not pursue post- secondary education typically form families earlier (Arnett & Tanner, 2011), enter into the
workforce earlier and earn less income by age 30 (Mitchell & Syed, 2015), and have more
difficulty securing stable jobs (Danziger & Ratner, 2010). However, home-leaving for this
population has yet to be empirically examined in the United States. European studies provide
some evidence that non-college attending youth may have distinct home-leaving patterns from
their peers. Cross-sectional data from young adults in the United Kingdom suggests that a
higher percent of young adults with less education live in the parental home, and that economic
insecurity is positively associated with living with parents (Stone et al., 2009; 2011). Using
longitudinal data from Europe, Iacovou (2010) found that the income of young adults is
positively associated with home leaving to all types of living arrangements considered.
Because non-college attending young adults may find it difficult to support themselves on
income earned from low-skilled employment (Danziger & Ratner, 2010), they may live in the
family home longer. Overall, the existing evidence suggests that young adults who do not
attend college will remain in the family home longer in part because they will find it difficult
to support independent living.
Non-college attending youth may also differ from other youth in the type of living
arrangement they enter into after exiting the family home. With the economic downturn and the
housing market collapse during the early 21st century, it is much harder for all young adults to
become homeowners right away (Clark, 2013). Perceptions of a tight housing market have been 7 found to delay home leaving in the Netherlands (Billari & Liefboer, 2007). Instead, young
adults live in diverse types of living arrangements such as living with roommates or living with a cohabiting partner (Clark & Mulder, 2000; Goldscheide & DaVanzo, 1989). Because non- college attending young adults may have been disproportionally affected, they may be unlikely to live independently upon first exit from the family home. Indeed, greater economic constraints were associated with a greater likelihood of living with others (Danziger & Ratner, 2010; Sironi
& Furstenberg, 2012) when young adults did exit the family home. Recent research found that emerging adults with less education were more likely to form cohabiting unions (Manning,
Brown, & Payne, 2014; Jang & Snyder, 2014), suggesting that non-college attending young adults may be more likely to cohabit with a romantic partner as a first living arrangement.
Barriers to Independent Living
Many young adults continue to be financially dependent on their parents, or other family figures (i.e. grandparents) during the transition to adulthood and family socioeconomic resources are significantly associated with home-leaving patterns (DeMarco & Berzin, 2008; Garasky et
al., 2001; Goldscheider & DaVanzo, 1989; Iacovou, 2010; Mulder & Clark, 2002). Further, the
parent’s ability to help subsidize the cost of living for their young adult children helps determine living arrangement options upon first exit (Goldscheider & Goldscheider, 1998; Iacovou, 2010;
Mulder & Clark, 2002). Because youth who do not attend post-secondary education are disproportionally more likely to come from lower socioeconomic backgrounds (Karen, 2002), family financial resources may be limited for this group. Thus, although family members may still be motivated to help their young adult offspring, they may be unable to help financially.
The lack of family socioeconomic resources available for young adults without a post- secondary education may shape home-leaving patterns. Fewer socioeconomic resources may be associated with earlier exits from the family home because families cannot afford to financially support a young adult. Because living with adult family members, such as parents or grandparents, can be an adaptive strategy for young adults in that it allows them to financially 8 prepare for independent living (Settersten & Ray, 2010), youth whose families cannot
financially support them into young adulthood may be forced to exit the family home early although they might not be prepared to do so. Indeed, early exits from the parental home are associated with young adult poverty (Rich, 2010). Further, upon exiting the home, families with
few socioeconomic resources may be less able to financially support their young adult’s
independent living, and young adults may be more likely to live with others.
Similar to parental socioeconomic resources, the socioeconomic resources available in a geographic area may also alter home-leaving patterns. For instance, nonmetropolitan areas of the
United States face worse economic conditions compared to metropolitan areas (Lichter, Parisi &
Taquino, 2012; Lichter & Johnson, 2007; Tickamyer & Duncan, 1990) and access to employment and livable wages differ between metropolitan and nonmetropolitan areas (Slack,
2010). Emerging adults who do not attend postsecondary education fare worse in the job market
(Danziger & Ratner, 2010), which may be magnified in rural, nonmetropolitan areas. Given these economic constraints, young adults in rural areas may be more likely to exit the parental home and live with a romantic partner or roommate, rather than live alone.
Gender Differences
Prior studies recognize the gendered nature of the young adult life course and have focused attention on gender differences in home leaving and returning (Berrington, Stone & Falkingham,
2009; Garasky, Haurin & Haurin, 2001; Stong, Berrington & Falkingham, 2011, 2014).
Regardless of educational attainment, women consistently exit the family home earlier than men
(Garasky, 2002), often for family formation purposes (Chiuri & Del Boca, 2010; Kleider & Ellis,
2011). When type of living arrangement upon first exit was considered, however, gender differences in the timing of home leaving decrease (Garasky et al., 2001). The gender gap in the timing of home leaving has also narrowed in recent years due to a larger share of young adult
women pursuing post-secondary education (Sironi & Furstenbers, 2012). When considering non
college bound young adults, gendered trends in post-secondary educational attainment may be 9 less relevant for home leaving patterns than economic factors such as housing costs and local
wages. Economic factors overall, especially housing costs, are associated with young adult
women’s, but not men’s, home leaving patterns, especially so as young women age (Garasky et
al., 2001). Although gender differences in these associations for home leaving patterns of non-
college-bound young men and women are unknown. We focus on gender differences in the type
and timing of first home leaving among non-college bound young adults, with the goal of filling
some of these gaps in the research literature.
Confounding Variables
Demographic characteristics. Decisions to exit from the family home differ by age,
race/ethnicity, high school dropout status, and family formation. Age is an important factor
related to the home-leaving process (Garasky et al., 2001; Setterson, 1998) as the desire to live independently and resources to be independent grow with age (Sironi & Furstenberg, 2012).
Racial/ethnic groups tend to reveal different life course trajectories (Glick, Ruf, White, &
Goldscheider, 2006); Black and Hispanic emerging adults are more likely to co-reside with parents or in other arrangements compared to White emerging adults (White, 1994), although
Hispanics are more likely, and African Americans notably less likely, to cohabit with a
nonmarital partner (Manning et al., 2014). Given the role of education in influencing the
pathway to adulthood (Hamilton & Hamilton, 2006; Sandefur, et al, 2005; Turner, 2004),
emerging adults who drop out of high school may be the most likely to remain in the family
home longer than those with greater education. Lastly, early childbearing is one pathway that
restricts opportunities for self-focused identity exploration that is central to the emerging
adulthood period of development (Arnett, 2006). Having children decreases the likelihood of
home-leaving to a group because of privacy (Avery et al. 1992), but may increase the likelihood
of exiting to live alone or with a partner.
Family characteristics. Emerging adult’s home-leaving process is influenced by their
family circumstances, including family structure and household size (Clark & Mulder, 2000). 10 Young adults with stepparents are more likely to leave the parental home for independent residence whereas those with both biological parents stay in the parental home longer than those with other family types (Avery et al., 1992; Goldscheider & Goldscheider 1998). Studies also find that while young adults from single parent households are less likely to live alone, they are more likely to live in a group setting or with a partner, due to economic constraints of both parents and children (Garasky et al., 2001; Manning & Lamb, 2003). Household size is another important factor associated with young adult’s home-leaving and living arrangements (Mulder &
Clark, 2002). It is harder for individuals to have privacy in a crowded household and resources are likely diluted for children in large families (Downey, 1995). Because privacy becomes increasingly more valued during the transition to adulthood, emerging adults from larger families may exit earlier than those from smaller families.
Current Study
The current study builds on prior research by examining the timing of home-leaving and type of living arrangement upon first exit for the ‘forgotten half’ or those who do not attend post-secondary education or training. We propose the following hypotheses:
H1: The timing of exits from the parental home for non-college bound young adults
will be later than existing patterns observed for all young adults; exits to live alone
will be less common than to live with a romantic partner, and will happen later.
H2: Personal and contextual socioeconomic barriers to home-leaving will delay
home-leaving from the family home; family socioeconomic barriers to home-leaving
will be associated with earlier exits from the parental home for households with
romantic partners.
H3: Women will be more likely to exit the family home earlier than men; women will
also be more likely to exit and live with romantic partners than men. Socioeconomic
barriers to home-leaving will be more pronounced in men than women.
Method 11 Data
We used data from the National Longitudinal Survey of Youth 1997 (NLSY97) public
and geocode data (n = 8,984). The NLSY97 public dataset was designed to examine the family
formation patterns, labor market behavior, and family and community backgrounds of youth.
The geocode restricted data provides supplemental detailed information on the geographic
residence of the respondent. Participants were born in the United States between 1980 and 1984
and interviewed annually beginning in 1997 (aged 12 to18 years old) until 2011, and again in
2013 (aged 28 to 34 years old). Out of an initial sample of 8,984 respondents, we restricted the
sample to those who do not enroll in any further education within one year following 1) receipt
of a high school diploma, 2) receipt of a general education development (GED) or high school
equivalency certificate, or 3) exiting school without a high school diploma or GED (n = 3,924).
We further restricted the sample to those with information on the date they received their high
school diploma, GED, or left school (n = 3,603).
Dependent Variable
Home-leaving. The home-leaving categorical variable was measured from household roster data available at each survey year; home-leaving was captured from the century months (converted into years) each respondent finished secondary school (via receipt of diploma, GED, or leaving) until the year they exited a family home. Only first exits from the family home were captured in the current analyses. Participants were given a ‘0’ when they remained in the family home
(reference), a ‘1’ when they first exited the family home to live alone (with or without children) or a ‘2’ when they first exited the family home to live with a spouse or partner (with or without children). Family home was used rather than parental home because of the diversity of living arrangements experienced by this sample.
Ten percent (n= 367) of respondents reported exiting the family home in the same year they finished secondary school. Monthly data was available for the secondary school end date and monthly information was available for respondent’s interview dates. Because home leaving was 12 captured by household roster information, or who the respondent was living with at the time of the interview, we were able to differentiate between cases where respondents left home prior to leaving school and those who left home after finishing school. Home leaving was lagged to the following year for 365 cases where the interview date occurred after the date respondents finished school (the remaining 2 participants who exited the family home prior to finishing school were not included in the analyses).
Independent Variables
Individual, parental, and regional socioeconomic characteristics were the independent
variables in each analysis; for analyses by gender, gender was measured where 1 = female and 0
= male. The current educational and/or employment status were time-varying dichotomous
indicators of working only, going to school only, and both working and going to school
(reference). Work history was taken from the average hours worked per week and the number of weeks worked employment history arrays; any work (including part-time) was counted as work.
Yearly current enrollment was used to measure schooling; any enrollment in any type of school was counted. Parental education was measured as a continuous variable in 1997 of the highest grade completed by the mother; if mother’s educational attainment was unavailable, father’s highest grade completed was used.
Regional characteristics included the classification of geographic area is urban or rural, county unemployment rates, and census regions. Urban or rural locations were time-varying measures at each survey year where 1 = urban and 0 = rural; participants living outside of the
United States were considered missing. We imported county unemployment rates from the
Local Area Unemployment Statistics (LAUS) of the Current Population Survey (CPS) (Bureau of Labor Statistics, 2013). The Federal Information Processing Standards (FIPS) codes of
respondents’ residence from the geocode file were used to merge with the Urban Influence
Codes and the county unemployment rates from LAUS. These county unemployment rates were
only available up to 2010. Census regions were measured by four time-varying dichotomous 13 indicators: Northeast region (reference), North central region, South region, and West region. To
help address issues relating to causality, all independent variables were lagged one year to
ensure that they occurred prior to exiting the parental home.
Control Variables
Control variables included age, race/ethnicity, experiencing pregnancy, having children,
dropping out of high school, family structure, and household size. Age was included as a
continuous variable; age-squared was also included to test for a nonlinear effect of time.
Race/ethnicity variables were dichotomous indicators of Black, non-Hispanic, Hispanic, and non-
Hispanic, non-Black (reference). Pregnancy was a time-varying dichotomous variable of any
pregnancy resulting in a live birth. Children were a time-varying dichotomous variable where 1 =
any biological child under age 18 in the household and 0 = no biological child. High school
dropout status was a dichotomous variable where 1 = enrolled in high school, but never received
a high school diploma or completed a GED. Family structure was measured by dichotomous
time-varying variables at each survey year: biological/adoptive two parent family (reference),
single parent family, step mother, and step father. Household size was a continuous time-varying
variable of the number of individuals who were included in the household roster at each survey.
Similar to the independent variables, all control variables were lagged one year to ensure that
they occurred prior to exiting the parental home.
Analytic Plan
To test our research questions, we used survival estimates and discrete-time multinomial logistic regression models. In our data, the onset of risk was the year a participant received a high school diploma, a GED or high school equivalency certificate, or exited school without a high school diploma or GED. Participants continued to be at risk until they exited their family home or were censored at their last interview date; those who exited the parental home prior to their onset of risk were dropped from these analyses (n = 239; analytic sample n = 3,364). From those remaining, 2,927 exited the family home during the duration of observation and 437 did not exit 14 the family home before their last interview date (note, in 5 cases, participants exited the family home to live with roommates; these were excluded from the analyses resulting in a final analytic sample of 3,359). We conducted preliminary Kaplan-Meier survival estimates to examine the timing until home leaving (to any living arrangement) for the full sample and by gender. We then conducted multiple-decrement Kaplan-Meier survival estimates to examine time until home leaving separately for living alone and with a romantic partner for the full sample and separately by gender. Discrete-time multinomial logistic regression models predicted exiting the parental home to live alone or with a romantic partner relative to remaining in the family home from individual, household, and residential characteristics. Additionally, we analyzed these models by gender.
Missing data was estimated for all independent and control variables using multiple imputation (MI command in Stata) that creates multiple datasets containing imputed values on missing information. Twenty-five different datasets were created with the multiple imputation using chained questions technique (MICE); each variable was treated as the dependent variable and all variables regressed onto it (Johnson & Young, 2011). Further, while each dataset included the dependent variable in the imputation process, the imputed dependent variables were excluded from the final analyses (von Hippel, 2007). All analyses were adjusted using the survey setting in
Stata to reflect the sampling procedure of the NLSY97 (Cleves et al., 2010; Center for Human
Resource Research, 2014).
Results
Descriptive Results
Sample Characteristics. Weighted descriptive statistics for the full sample and by gender are presented in Table 1. Chi-square and t-tests also tested for significant differences by gender. In the first year following school completion about 20% of the sample have exited the parental home, half to live alone and half to live with a romantic partner. Overall, at the time of home leaving, the sample is 22 years old, and most are working only and not enrolled in any school. 15 Significant gender differences exist on many characteristics including type of home leaving, work and schooling, household size, presence of biological children, family structure of parental home, and residential location. Notably, significantly more young women have left home at this time
(25%) compared to young men (18%), and more young women exit to live with a partner (15%) compared to young men (9%). In addition, a much larger share of young women have a biological child (32% compared to 8% of young men), and have experienced a pregnancy (12% compared to 7%). A larger share of young women in the sample also come from a single parent family household, and a larger share of young men come from a household containing two biological parents. Overall, among this sample of young adults who do not attend college, this table provides descriptive evidence that a larger share of young women come from less advantaged family backgrounds, and that early family formation motivates home leaving more so young women than for young men.
Kaplan-Meier survival estimates. Figures 1 and 2 provide survival estimates for first home leaving events overall and separately by gender. Figure 1 shows that within 3 1/2 years of finishing high school about 50% of the sample exits the family home. Figure 2 shows that young adult women exit sooner overall compared to men. For example, about 50% of young women have exited the parental home within 3 years, whereas it takes about 4 years for 50% of young men to exit the parental home. This gender difference persists throughout the period of observation. Tables 2 and 3 provide survival estimates for first home leaving events by type of exit (to live alone or to live with a romantic partner) and also survival estimates for first home leaving event by type separately for young women and young men. Table 2 shows that within 5 years of finishing school, 37% of young adults had exited to live along compared to 32% who exited to live with a partner. When examining these exits by gender, we show in Table 3 that within 5 years of finishing school 40% of women have exited to live alone compared to 35% of men, and 43% of young women had exited to live with a partner compared to 23% of young men,
Thus, we observe gender differences in the timing of exits to both living arrangement types, and 16 that gender differences are especially large for exits to live with a partner. This finding is consistent with existing literature documenting early family formation as a common pathway to adulthood, especially among less educated young women (Osgood et al., 2005).
Multivariate Results
Table 4 presents discrete-time logistic regression results that estimate home leaving for the full sample, and examines exits to living alone and with a romantic partner. When comparing exits to live alone versus remaining in the family home, we find significant differences by work, parental education and region of residence. Compared to those who combine work and schooling, those who only work have a lower relative risk (RR=.55) of exiting to live alone versus remaining at home, and parent’s education is positively associated with exiting to live along (RR=1.04).
Compared to those living in the eastern region of the U.S., those living in all other regions have a higher relative risk of exiting to live alone (RR=1.62) versus remaining living in the family home.
We also find significant differences when examining exits to live with a romantic partner versus remaining in the family home. Females have a significantly higher relative risk of exiting to live with a romantic partner than males (RR=1.73), which is consistent with our life table estimates, and young adults who work only have a lower relative risk of exiting to live with a romantic partner (RR=.76), compared to those who combine work and schooling. Finally, compared to living in the east, living in the north central region of the nation increases the relative risk of exiting to live with a romantic partner (RR=1.44). The final model in Table 4 presents the competing risks of exiting to live with a partner versus exiting to live alone. We found that young adult women have significantly higher relative risks of living with a partner, as do those who are only working. Higher parental education lowered the relative risk of living with a partner. These findings confirm the significant gender differences in the timing of first exits, and the type of living arrangement upon first exit from the family home.
We also find that some of the control variables are associated with the type and timing of home exits for young adults. For example, having one’s own biological child speeds the transition to 17 living alone, as does the presence of a stepmother in the parental home. In contrast, larger parental household size, Hispanic ethnicity and having dropped out of high school significantly reduce the relative risk of exiting to live alone. Control variables significantly positively associated with exiting to live with a romantic partner include pregnancy and the presence of one’s own biological child, and those significantly negatively associated with exits to live with a romantic partner include being either Hispanic or non-Hispanic Black, and coming from a single parent family home. Age significantly reduces the relative risk of exiting to live with a romantic partner, but this effect decreases with age.
Discrete-time multinomial logistic regression models by gender. Both theory (Arnette, 2006) and empirical studies (Garasky et al., 2001) motivate separate analyses by gender. For young men, we find that higher parental education is associated with a higher relative risk (RRR=1.07) of remaining in the parental home regardless of the competing risk, whereas exclusively working significantly reduces the relative risk of exiting to live alone, regardless of the competing risk.
Only one of our independent variables, residence in the North Central region of the U.S., significantly increases the relative risk of exiting to live with a romantic partner (RR=1.49), and none significantly decrease the relative risk. When considering the control variables we find that relative risk of exiting to live alone is significantly decreased among those from a large family household, who experience a pregnancy, and who are either non-Hispanic Black or Hispanic; and is significantly increased by the presence of a stepmother in the family household. Young adult men with biological children in the household and who experience a pregnancy have significantly higher relative risks of exiting to live with a romantic partner, and the opposite is true for non-
Hispanic Black men.
For young adult women, higher parental education is not associated with either the type or timing of home leaving, whereas working only increases the relative risk of remaining in the parental home, regardless of the competing risk. Residence in the North Central region increases the relative risk of exiting to both living arrangements (RR=1.56 and 1.41), but is not associated 18 with the competing risk of exiting to live either alone or with a partner. When considering the control variables we find that pregnancy increases the relative risk of exiting to live with a romantic partner, whereas having a biological child increases the relative risk of exiting to live alone. In general, being anything other than non-Hispanic White decreases the relative risk of any home leaving, with the exception of non-Hispanic Blacks who have a significantly higher relative risk of exiting to live alone. The presence of a stepmother more than doubles the relative risk of exiting to live alone, whereas a single parent family home decreases the relative risk of exiting to live with a romantic partner. For young women, age significantly reduces the relative risk of exiting to live with a romantic partner, but this effect decreases with age.
Discussion
This study examined the home leaving process as a way to better understand the
transition to adulthood among young adults in the U.S. We used a nationally-representative
sample of young adults who do not pursue post-secondary education—the forgotten half. We
focus on this understudied group in part to fill a gap in our research knowledge about transitions
to independence among young adults, but also as a way to evaluate the relevance Emerging
Adulthood Theory for the transition to adulthood among less educated young adults. We also
emphasized gender differences in home leaving due to the recognized sex differences in the early
adult life course, as has been done in numerous prior studies (Billari, Philipov & Baizan, 2001; 19
Billari & Leifbroer, 2010; Holdsworth, 2000; Iacovou, 2010; Stone, Berrington & Falkingham,
2011).
Living away from the family home is a key marker of the transition to adulthood (Sassler et al., 2008), and existing research finds that achieving this independent status now takes longer for young adults (Yelowitz, 2007). One explanation for the lengthening of this process is the increase time required to build human capital and explore one’s identity—both thought to be critical developmental tasks that best prepare today’s young adults for adulthood (Arnette, 2006). This study questions whether or not this extended period of human development is relevant for the lives of less educated young adults, and provides empirical evidence that perhaps it is not, at least when considering exiting the family home as an indicator of independence during young adulthood.
We find significant gender differences in the timing and type of living arrangement upon first exit from the parental home. Young adult women in our sample exited from the parental home sooner than young adult men, and these differences were especially pronounced when considering exits to live with a partner. Early family formation appears to be a common part of the life course for young adult women who do not go on to schooling after finishing high school. We also found that the work efforts of young adults are associated with exits from the family home, although the effects are different for young men and women. For men, work efforts are associated with being in a partnership, or living at home, and for young women work efforts are associated with remaining in the family home. An unexpected finding was the positive association between parent’s education and the risk of exiting to live alone, but only young men. The multivariate analyses also find that pregnancy significantly increases the risk of exiting to live with a partner for young men and women, but the presence of a biological child is associated to exits to live with a partner for young men. For young women, the presence of a biological child prompts exits to live alone. 20
One key criticism of emerging adulthood theory is that it may not be relevant for all, but mostly for more privileged groups that attend and complete college, often with support from parents
(Bynner, 2005; Côté, 2014; Côté & Bynner, 2008; Hendy & Kloep, 2007). Critics argue that a majority of young adults do not experience an extended period of young adulthood that can be used to pursue post-secondary education and explore one’s identity, preferences and life goals before assuming full adult roles and responsibilities, in large part because of their lack of economic resources (e.g., Silva 2013). Empirical findings from studies of the transition to adulthood among
American youth provide some evidence for this concern (Osgood, Ruth, Eccles, Jacobs & Barber,
2005).
Opponents of emerging adulthood theory often favor the life course perspective as an explanation for changes in the lives of today’s young adults (Bynner, 2005; Côté & Bynner, 2008;
Stone et al., 2011). Life course theory (Elder, 1998) is thought to better balance the effects of individual choice and structural factors when studying human development and outcomes.
Numerous studies using the life course perspective have examined how structural factors, including macroeconomic shocks, shape human lives by providing opportunities or constraints that direct young adults along limited pathways (Conger & Elder, 1994; Elder, 1999; Elder, Johnson &
Crosnoe, 2003). Perhaps most importantly, the main tenants of life course theory— socio-historical context, timing of lives, linked lives, and human agency—are thought to be relevant for young adults from all backgrounds.
We believe our findings also provide evidence that questions the relevance of Emerging Adult
Theory for our sample of less educated young adults. Work and early family formation are prominent features of the life course for these young adults, and are associated with transitions out of the parental home during the late teens and early 20’s. Very few of these young adults are engaged in 21 any schooling. These findings provide some evidence that these young adults are more quickly launched into adult roles relatively soon after completing high school. Whether or not these transitions are permanent is yet to be seen. The duration of these first living arrangements and likelihood of returning to the family home are beyond the scope of this study, but these are areas where future research efforts should be directed.
References 22
Aquilino, W.S. (2006). Family relationships and support systems in emerging adulthood. In J. J. Arnett
& J. L. Tanner (Eds.), Emerging adults in America: Coming of age in the 21st century (pp.193-
218). Washington, DC: American Psychological Association.
Arnett, J. J.(2000). Emerging adulthood: A theory of development from the late teens through the
twenties. American Psychologist, 55, 469-480.
Arnett, J. J. (2006). Emerging adulthood: Understanding the new way of coming of age. In J. J. Arnett
& J. L. Tanner (Eds.), Emerging adults in America: Coming of age in the 21st century (pp. 3–
19). Washington, DC: American Psychological Association.
Arnett, J. J. (2007). Emerging adulthood, a 21st century theory: A rejoinder to Hendry and Kloep.
Child Development Perspectives, 1, 80-82.
Arnett, J. J. (2016). Emerging adulthood and social class rejoinder to Furstenberg, Silva, and du Bois-
Reymond. Emerging Adulthood, 4, 244-247.
Arnett, J. J., & Tanner, J. L. (2011). Themes and variations in emerging adulthood across social
classes. In J. J. Arnett, M. Kloep, L.B. Hendry & J. L. Tanner (Eds.), Debating emerging
adulthood: Stage or process? New York, NY: Oxford University Press.
Avery, R., Goldschneider, F., & Speare, Jr., A. (1992). Feathered nest/gilded cage: Parental income
and leaving home in the transition to adulthood. Demography, 29, 375-388.
Bynner, J. (2005). Rethinking the youth phase of the life-course: The case for emerging adulthood?
Journal of Youth Studies, 8, 367-384.
Carr, P. J., & Kefalas, M. J. (2011). Straight from the heartland: Coming of age in Ellis, Iowa. In M. C.
Waters, P. J. Carr, & M. J. Kefalas (Eds.), Coming of age in America: The transition to
adulthood in the twenty-first century. Berkeley: University of California Press. 23
Center for Human Resource Research (2014). NLSY Handbook. The Ohio State University, Columbus,
OH.
Clark, W. A. (2013). The aftermath of the general financial crisis for the ownership society: What
happened to low-income homeowners in the US? International Journal of Housing Policy, 13,
227-246.
Clark, W.A., & Mulder, C.H. (2000). Leaving home and entering the housing market. Environment
and Planning A, 32, 1657-1671.
Cleves, M., Gutierrez, R. G., Gould, W. & Marchenko, Y. V. (2010). An Introduction to Survival
Analysis using Stata (3rd ed.). College Station, TX: Stata Press.
Côté, J., & Bynner, J. M. (2008). Changes in the transition to adulthood in the UK and Canada: The
role of structure and agency in emerging adulthood. Journal of Youth Studies, 11, 251-268.
Côté, J. E. (2014). The dangerous myth of emerging adulthood: An evidence-based critique of a flawed
developmental theory. Applied Developmental Science, 18, 177-188.
Danziger, S., & Ratner, D. (2010). Labor market outcomes and the transition to adulthood. The Future
of Children, 20, 133-158.
De Marco, A., & Berzin, S. C. (2008). The influence of family economic status on home-leaving
patterns during emerging adulthood. Families in Society: The Journal of Contemporary Social
Services, 89, 208-218.
Downey, D. B. (1995). When bigger is not better: Family size, parental resources, and children's
educational performance. American Sociological Review, 60, 746-761.
Furstenberg, F. (2016). Social class and development in early adulthood: Some unsettled issues.
Emerging Adulthood, 4, 236-238. 24
Garasky, S., Haurin, J. R., & Haurin, D. R. (2001). Group living decisions as youths transition to
adulthood. Journal of Population Economics, 14, 329-349.
Glick, J. E., Ruf, S. D., White, M. J., & Goldscheider, F. (2006). Educational engagement and early
family formation: differences by ethnicity and generation. Social Forces, 84, 1391-1415.
Goldscheider, F. K., & DaVanzo, J. (1989). Pathways to independent living in early adulthood:
Marriage, semiautonomy, and premarital residential independence. Demography, 26, 597-614.
Goldscheider, F. K., & Goldscheider, C. (1998). The effects of childhood family structure on leaving
and returning home. Journal of Marriage and the Family, 60, 745–756.
Halperin, S. (1998). The Forgotten Half revisited. American youth and young families, 1988-2008.
American Youth Policy Forum: Washington, D.C.
Hendry, L. B., & Kloep, M. (2007). Conceptualizing emerging adulthood: Inspecting the emperor’s
new clothes? Child Development Perspectives, 1, 74–79.
Hamilton, S. F., & Hamilton, M. A. (2006). School, work, and emerging adulthood. In J. J. Arnett & J.
L. Tanner (Eds.), Emerging adults in America: Coming of age in the 21st century (pp. 257–
277). Washington, DC: American Psychological Association.
Iacovou, M. (2010). Leaving home: Independence, togetherness and income. Advances in Life Course
Research, 15, 147-160.
Jang, B. J., & Snyder, A. R. (2015). Moving and union formation in the transition to adulthood in the
United States. Advances in Life Course Research, 23, 44-55.
Johnson, D. R. & Young, R. (2011). Toward best practices in analyzing datasets with missing data:
Comparisons. Journal of Marriage and Family 73, 926 – 945.
Karen, D. (2002). Changes in access to higher education in the United States: 1980–1992. Sociology of
Education, 75, 191–210. 25
Kins, E., & Beyers, W. (2010). Failure to launch, failure to achieve criteria for adulthood? Journal of
Adolescent Research, 25, 743-777.
Lichter, D. T., & Johnson, K. M. (2007). The changing spatial concentration of America's rural poor
population. Rural Sociology, 72, 331-358.
Lichter, D. T., Parisi, D., & Taquino, M. C. (2012). The geography of exclusion: Race, segregation,
and concentrated poverty. Social Problems, 59, 364-388.
Manning, W. D., Brown, S. L, & Payne, K. K. (2014). Two decades of stability and change in age at
first union formation. Journal of Marriage and Family, 76, 247-260.
Manning, W. D., & Lamb, K. A. (2003). Adolescent well‐being in cohabiting, married, and single‐
parent families. Journal of Marriage and Family, 65, 876-893.
Mitchell, L. L., & Syed, M. (2015). Does college matter for emerging adulthood? Comparing
developmental trajectories of educational groups. Journal of youth and adolescence, 44, 2012-
2027.
Mulder, C. H., & Clark, W. A. (2002). Leaving home for college and gaining independence.
Environment and Planning A, 34, 981-999.
Osgood, D. W., Ruth, G., Eccles, J. S., Jacobs, J. E., & Barber, B. L. (2005). Six paths to adulthood. In
R. A. Settersten, F. F. Furstenberg Jr, & R. G. Rumbaut (Eds.), On the frontier of adulthood:
Theory, research, and public policy (pp. 320–355). Chicago, IL: University of Chicago Press.
Qian, Z. (2012). During the Great Recession, more young adults lived with parents. US2010 Project,
Providence, RI.
Sandefur, G. D., Eggerling-Boeck, J., & Park, H. (2005). Off to a good start? Postsecondary education
and early adult life. In R. A.Settersten, Jr., F. F.Furstenberg, Jr., & R. G.Rumbaut (Eds.), On the 26
frontier of adulthood: Theory, research, and public policy (pp. 292–319). Chicago: University of
Chicago Press.
Settersten, R. A., & Ray, B. (2010). What’s going on with young people today? The long and twisting
path to adulthood. The Future of Children, 20, 19–41.
Silva, J. M. (2012). Constructing Adulthood in an Age of Uncertainty. American Sociological Review,
77, 505-522.
Silva, J. M. (2013). Coming up short: working-class adulthood in an age of uncertainty. Oxford:
University Press.
Sironi, M., & Furstenberg, F.F. (2012). Trends in the economic independence of young adults in the
United States: 1973-2007. Population and Development Review, 38, 609-630.
Stone, J., Berrington, A., & Falkingham, J. (2014). Gender, turning points, and boomerangs: Returning
home in young adulthood in Great Britain. Demography, 51, 257-276.
Schwartz, S. J., Côté, J. E., & Arnett, J. J. (2005). Identity and agency in emerging adulthood two
developmental routes in the individualization process. Youth & Society, 37, 201-229.
Tickamyer, A. R., & Duncan, C. M. (1990). Poverty and opportunity structure in rural America.
Annual Review of Sociology, 16, 67-86.
Von Hippel, P. T. (2007). Regression with missing ys: An improved strategy for analyzing multiply
imputed data. Sociological Methodology, 37, 83-117.
White, L. (1994). Coresidence and leaving home: Young adults and their parents. Annual Review of
Sociology, 20, 81–102.
Yelowitz, A. S. (2007). Young adults leaving the nest: The role of the cost of living. In S.Danziger &
C.Rouse (Eds.), The price of independence (pp. 170–207). New York: Russell Sage.
27
Table 1. Weighted descriptive statistics (at the time of home-leaving) Full sample Men Women (n =3,359) (n=2,040) (n=1,319) Mean (SD) Mean (SD) Mean (SD) Home-leaving Living with family 0.80 0.821 0.75 Living alone 0.10 0.09 0.10 Living with partner 0.11 0.091 0.15 Age 22.00 (2.97) 22.221 (2.97) 21.54 (2.93) Race/ethnicity Non-Black, non-Hispanic 0.58 0.57 0.59 Black 0.23 0.24 0.22 Hispanic 0.20 0.19 0.20 Work and school Work only 0.91 0.91 0.89 School only <0.01 <0.01 <0.01 Both 0.02 0.021 0.03 Neither 0.07 0.06 0.08 High school drop-out 0.42 0.43 0.40 Female 0.33 - - Household size 3.79 (1.84) 3.721 (1.79) 3.92 (1.92) Biological children in HH 0.16 0.081 0.32 Pregnancy 0.08 0.071 0.12 Family structure Two bio parents 0.23 0.261 0.16 Single parent 0.23 0.221 0.26 Stepmother 0.01 0.01 0.01 Stepfather 0.08 0.08 0.08 Parental education 11.62 (2.73) 11.63 (2.75) 11.61 (2.68) Residence Urban 0.75 0.731 0.77 County unemployment rates 5.26 (2.32) 5.31 (2.33) 5.15 (2.30) Census region Northeast 0.19 0.19 0.18 North central 0.21 0.21 0.20 South 0.40 0.40 0.40 West 0.20 0.19 0.22 Note. 1Indicates significant gender difference. SD = Standard Deviation
Running head: HOME-LEAVING AND LIVING ARRANGMENTS
Figure 1. Kaplan-Meier survival estimate for first homeleaving
Survival curve for first homeleaving 1.00 0.75 0.50 0.25 0.00 0 5 10 15 analysis time
Figure 2. Kaplan-Meier survival estimates for first homeleaving by gender 29
Survival curves for first homeleaving by gender 1.00 0.75 0.50 0.25 0.00 0 5 10 15 analysis time
Men Women
30
Table 2. Life table survival estimates of first homeleaving by exit type Percentage exiting home1 Year Live alone Live with partner 1998-1999 99% 99% 1999-2000 95% 96% 2000-2001 86% 88% 2001-2002 76% 80% 2002-2003 63% 68% 2003-2004 48% 55% 2004-2005 35% 42% 2005-2006 25% 32% 2006-2007 18% 25% 2007-2008 13% 19% 2008-2009 10% 14% 2009-2010 7% 9% 2010-2011 4% 6% 2011-2012 3% 3% 2012-20132 - - 2013-2014 0% 0% Person-years/months 5,096 6,384 N 1,395 1,523 Note. 1441 individuals do not exit the family home. 2The NLSY97 does not collect data from 2012 and homeleaving estimates from 2012-2013 cannot be estimated. 31
Table 3. Life table survival estimates of first homeleaving by exit type and gender Percentage exiting home1 Live alone Live with partner Year Men Women Men Women 1998-1999 99% 99% 99% 99% 1999-2000 95% 93% 97% 95% 2000-2001 88% 82% 92% 84% 2001-2002 78% 73% 87% 72% 2002-2003 65% 60% 77% 57% 2003-2004 49% 46% 63% 44% 2004-2005 36% 31% 51% 31% 2005-2006 27% 23% 41% 22% 2006-2007 19% 16% 31% 17% 2007-2008 14% 11% 24% 13% 2008-2009 11% 9% 17% 9% 2009-2010 8% 6% 12% 6% 2010-2011 4% 4% 8% 3% 2011-2012 3% 3% 4% 2% 2012-20132 - - - - 2013-2014 0% 0% 0% 0% Person-years/months 3,259 1,837 4,042 2,342 n 872 523 837 686 Note. 1441 individuals do not exit the family home (331 men and 110 women). 2The NLSY97 does not collect data from 2012 and homeleaving estimates from 2012-2013 cannot be estimated.
32
Table 4. Discrete-multinomial logistic regression model (n =3,359) Live alone Live with partner vs. In Live with partner vs. vs. In family home (ref) family home (ref) Live alone (ref) RRR SE RRR SE RRR SE Age 0.99 0.15 0.75* 0.10 0.75 0.13 Age2 0.99 0.01 1.01* 0.01 1.01* 0.01 Race/ethnicity Black 0.97 0.08 0.47*** 0.04 0.48*** 0.06 Hispanic 0.64*** 0.07 0.74*** 0.07 1.16 0.16 Work and school Work only 0.55*** 0.05 0.76** 0.07 1.37** 0.15 School only 1.06 0.22 1.22 0.27 1.15 0.29 High school drop-out 0.85* 0.06 0.91 0.06 1.08 0.10 Female 0.98 0.07 1.73*** 0.12 1.76*** 0.15 Household size 0.92*** 0.02 1.01 0.02 1.09** 0.03 Biological children in HH 1.96*** 0.21 1.36** 0.14 0.68** 0.08 Pregnancy 0.95 0.12 3.03*** 0.29 3.18*** 0.45 Family structure Single parent 0.93 0.08 0.85* 0.07 0.91 0.09 Stepmother 1.97*** 0.41 1.17 0.28 0.59 0.16 Stepfather 1.24 0.15 0.99 0.12 0.80 0.13 Parental education 1.04** 0.02 0.99 0.01 0.95** 0.02 Residence Urban 0.96 0.08 0.89 0.08 0.92 0.11 County unemployment 1.00 0.02 1.01 0.01 1.01 0.02 rates Census region North central 1.62*** 0.17 1.44*** 0.15 0.89 0.14 South 1.30** 0.13 1.19 0.11 0.92 0.12 West 1.47*** 0.17 1.11 0.12 0.75 0.12 Person-years 15,128 Note. RRR = Relative Risks Ratio. SE = Standard Error. * p < .05, ** p < .01, *** p < .001
33
Table 5. Discrete-multinomial logistic regression model by gender Men Women Live alone vs. Live with partner Live with Live alone vs. In Live with partner Live with In family home vs. In family partner vs. live family home vs. In family partner vs. liv (ref) home (ref) alone (ref) (ref) home (ref) alone (ref) RRR SE RRR SE RRR SE RRR SE RRR SE RRR SE Age 0.96 0.18 1.09 0.21 1.14 0.29 1.07 0.29 0.89* 0.13 0.55 0.18 Age2 0.99 0.01 0.99 0.01 0.99 0.01 0.99 0.01 1.01* 0.01 1.02* 0.01 Race/ethnicity Black 0.79* 0.08 0.67*** 0.07 0.84 0.12 1.34* 0.17 0.30*** 0.04 0.22*** 0.04 Hispanic 0.67*** 0.08 0.82 0.10 1.23 0.21 0.60** 0.11 0.66*** 0.09 1.11 0.23 Work and school Work only 0.56*** 0.06 0.80 0.10 1.42* 0.23 0.52*** 0.08 0.72* 0.10 1.40 0.26 School only 0.68 0.22 0.84 0.34 1.22 0.60 1.39 0.41 1.33 0.38 0.96 0.35 High school drop-out 0.87 0.08 0.95 0.08 1.09 0.13 0.81 0.10 0.84 0.09 1.04 0.17 Household size 0.92** 0.03 0.99 0.03 1.08 0.04 0.93 0.04 1.01 0.03 1.09 0.05 Biological children in HH 0.94 0.23 2.31*** 0.38 2.46** 0.70 2.34*** 0.32 1.24 0.16 0.53*** 0.09 Pregnancy 0.57* 0.13 3.55*** 0.46 6.27*** 1.60 1.27 0.22 2.41*** 0.33 1.90*** 0.37 Family structure Single parent 0.99 0.10 0.90 0.10 0.91 0.13 0.83 0.12 0.78* 0.09 0.93 0.16 Stepmother 1.98** 0.49 1.17 0.38 0.59 0.23 2.14* 0.79 1.17 0.42 0.55 0.24 Stepfather 1.32 0.20 1.16 0.19 0.87 0.18 1.10 0.21 0.85 0.14 0.78 0.17 Parental education 1.07** 0.02 0.99 0.02 0.93** 0.02 1.01 0.03 0.99 0.02 0.99 0.03 Residence Urban 0.99 0.11 0.85 0.09 0.85 0.13 0.91 0.13 0.95 0.12 1.05 0.19 County unemployment rates 1.01 0.02 0.99 0.02 0.99 0.02 1.01 0.03 1.01 0.02 0.99 0.03 Census region North central 1.62*** 0.22 1.49** 0.20 0.92 0.17 1.56* 0.28 1.41* 0.23 0.90 0.20 South 1.41** 0.18 1.10 0.14 0.78 0.13 1.12 0.19 1.27 0.19 1.13 0.23 West 1.64*** 0.23 1.05 0.15 0.64* 0.12 1.23 0.24 1.14 0.17 0.93 0.22 Person-years 10,013 5,115 N 2,040 1,319 Note. Ref = reference. RRR = Relative Risks Ratio. SE = Standard Error * p < .05, ** p < .01, *** p < .001 34