Adolescent Family Context and Adult Identity Formation

Abstract

This study examines the links between adolescent family context and coming to see oneself as an adult. Using data from the National Longitudinal Study of Adolescent Health, we investigate how adolescent family structure, resources, and processes together influence adult identity and whether they do so similarly for men and women. We find that youth in single- or step-parent families, but not in two parent adoptive families, are more likely to identify as adults compared to those in two biological parent families. These relationships, however, are mediated by both family resources and processes. We also find that one of these processes, parental control, is especially influential for youth in single-father and “other” family structures, and that parent-adolescent relationship quality and living in “other” structure families are more consequential for young women than men.

INTRODUCTION

Research on the transition from adolescence to adulthood has grown in recent years as scholars grapple with the extension of this period of the life course and its precursors and consequences, both macro-historically and with respect to individual variations in trajectories (Furstenberg et al, 2004). Today the transition to adulthood takes place over an extended period of time and is marked by greater variability in the timing and order in which young people move into traditional adult roles such as spouse, parent, and worker (Shanahan, 2000; Furstenberg et al, 2004) Although most of this research has focused on adult role transitions, a small but growing literature is attending to the subjective transition to adulthood, including the formation of an adult age identity (Shanahan et al., 2005; Benson & Furstenberg, 2007; Johnson et al., 2007; Macmillan, 2007; Arnett, 2004).

This early research suggests that adult identity formation is closely linked to family of origin (Shanahan et al. 2005; Benson & Furstenberg, 2007; Johnson et al., 2007), but scholars have not yet identified the mechanisms through which it operates. Our understanding to date is limited to a rudimentary observation: young people in their late teens and twenties who grew up in families without two biological parents are more likely to see themselves as adults than those of the same age who grew up in families with two biological parents (Benson & Furstenberg, 2007; Johnson et al., 2007).

The effects of family structure are often understood in terms of economic resources, but developmental scholars also point to the importance of understanding the nature and type of relationships within families (Settersten, forthcoming; Cooney & Mortimer, 1999; Musick & Bumpass, 1999). During adolescence and young adulthood, youth engage in an on-going negotiation with their parents to develop autonomy and independence (O’Connor et al., 1996). While research suggests that parents serve as a key force that can either push youth toward or pull youth away from maturity and adult development (Eccles et al., 1993; Settersten, forthcoming; Galambos et al., 2003), existing research has not specified how family context influences the types of messages young people receive from their families.

In this paper, we conceptualize family context to include structure, resources, and processes. Drawing upon Amato’s (1993) general model of “resources and stressors,” we contend that family resources and processes will have direct effects on identifying as an adult and will provide key mediating pathways through which family structure influences self-perceived adulthood. To examine these processes, we use longitudinal data from the National Longitudinal Study of Adolescent Health. First, we investigate whether living in different family structures during adolescence shapes subjective age in young adulthood. Second, we examine whether family resources, including family income and parental education, and family processes, including conflict, monitoring, hierarchy and shared responsibility, influence subjective age and mediate the effects of family structure on self-perceived adulthood. Finally, we investigate the ways in which gender moderates the relationship between family context and subjective age. The next section describes our theoretical approach and reviews extant literature on family context that informs an understanding of how young people come to see themselves as adults.

Contextual Perspective on Adult Identity Formation

According to identity theory, social identities are constructed through interaction with significant others and are largely influenced by the social structures or contexts within which people are embedded (Stryker & Serpe, 1994; Erikson, 1968). During childhood and adolescence, the family is a particularly important context of socialization (Elder, 1968). Developmental scholars argue that initial identity content is based on feedback youth receive from their parents, although youth will incorporate new information and contexts into their identities as they grow older (Bronfenbrenner 1979; Eccles et al., 1993; Erikson, 1968). Since families are such an important context for identity formation, we expect that the cues and signals adolescents receive from their families, especially their parents, about where they fall along the developmental continuum from adolescence to full-fledged adulthood will shape youth’s subjective age.

Family Structure

Research consistently shows that young people growing up in families without two biological parents are more likely to consider themselves adults than those reared in two biological parent families (Benson & Furstenberg, 2007; Johnson et al., 2007). Prior studies have oversimplified family structure, however, grouping all young people who did not grow up with two biological parents together. These various structures are distinct from one another in important ways, including on parent-adolescent relationships and economic and social support, and some may be more similar to families with two biological parents than they are to family structures with which they are grouped (Musick & Bumpass, 1999; Sweeney, 2007; Cavanagh, 2008).

Family Resources

Youth from families with lower socioeconomic status, as indexed by parental education levels and household income, tend to feel more like an adult than their more advantaged peers (Johnson et al., 2007; Benson & Furstenberg, 2007). Less able to rely on their families for economic and social capital, adolescents from lower socioeconomic backgrounds generally expect to move through major transition markers at earlier ages than their more advantaged counterparts (Furstenberg et al., 2004; Neugarten et al., 1965). In childhood and adolescence, they often take on greater household and financial responsibilities, and have more adult-like interactions with their parents (Burton, 2007). Thus, we expect that youth from less advantaged families will be more likely to perceive themselves as adults than those from more advantaged families.

In addition, because household income and parental education differs across family structures, we expect that resources will account for some of the family structure differences in self-perceived adulthood (Casper & Bianchi, 2002). For example, studies consistently show that single-mother families are at the bottom of the income distribution while two-parent families are at the top (Casper & Bianchi, 2002).

Family Processes

Multiple pathways exist through which families may influence adult identity formation, including levels of social control and monitoring, warmth and closeness, responsibility, and hierarchical family relations (Musick & Bumpass, 1999). We conceptualize these family processes not as competing pathways but rather as mechanisms that may operate simultaneously on the development of an adult identity.

Monitoring and Control

Moderate levels of parental monitoring and control are important to successful development, but excessive levels (e.g., being too overprotective, having too many rules, and not allowing youth to make independent decisions) can undermine independent identity formation (Elder, 1968; Zimmer-Gembeck, 2001). At the same time, very low levels of parental monitoring may accelerate adult identity formation by allowing adolescents to engage in and experiment with age inappropriate behaviors (Hogan & Kitagawa, 1985; Newcomer & Udry, 1987). Accordingly, we hypothesize that strong monitoring and control during adolescence will be negatively associated with young adults’ self-perceived adulthood, and very low levels of monitoring will be positively associated with it. Based on studies showing that single-parent families provide less monitoring than two biological parent families, we also expect that the effects of family structure operate partially through differences in monitoring and control (Hogan & Kitagawa, 1985; Newcomer & Udry, 1987).

Closeness and Conflict

Although developmental research suggests that warm relationships are important for healthy psycho-social development, research on how conflict influences autonomy development is mixed. It is argued that some parent-adolescent conflict during adolescence is normative and important because it facilitates parental detachment and self-reliance (Steinberg, 1989; Zimmer-Gembeck, 2001). There is not consensus, however, about whether detachment, particularly premature detachment caused by high-conflict relationships, leads to positive outcomes. Some argue that detachment is a sign of psychological maturity and should be associated with positive outcomes (Soenens et al., 2007). On the other hand, it is argued that detachment, especially early detachment, may have negative consequences for youth and can lead to more distress and greater susceptibility to peer pressure (Chen & Dornbusch, 1998). Because high conflict relationships can lead to both increased self-reliance and susceptibility to peer pressure and age inappropriate behaviors, we hypothesize adolescents who experience high conflict with their parents will be more likely to self-identify as adults.

Closeness and conflict should also partially mediate the effects of family structure on self-perceived adulthood, and the effects of parent-adolescent conflict on adult identity formation may differ by gender. Studies consistently show that non-intact family types, particularly divorced single-parent and step-parent families, have higher levels of conflict and detachment between parents and children than two parent, intact families (Hanson, McLanahan, & Thomson, 1996). In addition, women may be more sensitive than men to conflict within families. For example, Cooney and Mortimer (1999) found that family conflict accelerated home leaving for young women but not for young men.

Hierarchy Within the Family

Adult identity formation may also develop out of the status position adolescents hold within the household vis-á-vis their parental figure. Nock (1988) and Weiss (1979) suggest that the level of hierarchy within the household is associated with self-sufficiency. In households where the hierarchy between parent and adolescent is less defined and structured, generational boundaries tend to be blurred; reciprocal dependency between parent and adolescent replaces subordinate-superordinate relations (Nock 1988). Although not tested empirically, Nock (1988) contends that a peer-like, parent-adolescent relationship accelerates self-sufficiency because adolescents are typically given more independence and called on to contribute more to the household than those growing up in more hierarchal families. Based on this argument, we hypothesize that youth who grow up in families with peer-like parent-adolescent relationships will be more likely to self identify as adults than those in families with more hierarchical relationships.

We also expect authority relations within families to mediate the effects of family structure on self-perceived adulthood. Role ambiguity within step-parent and cohabiting families creates issues concerning the legitimacy and authority of parental figures (Buchanan, Maccoby & Dornbusch, 1996; Peterson & Hann, 1999).

Shared Responsibility

Taking on responsibility within the household is an important signal of adulthood and a developmental task through which adolescents gain a sense of competence and mastery (Cooney & Mortimer, 1999). By granting young people more responsibilities, parents signal to their children that they trust them to complete tasks typically designated to adults. Studies show that young adults who had more responsibilities within the household are more likely to define themselves as adults than those who had fewer responsibilities (Benson & Furstenberg, 2007). Accordingly, we expect that greater household responsibilities in adolescence will predict self-perceived adulthood.

In addition, we expect household responsibility to partially explain differences by family structure, with potential variations by gender. Children and adolescents in single-parent and step-parent families do more housework than those in two-parent households (Cooney & Mortimer, 1999), and there is evidence that household responsibility explains some family structure differences in adult transition timing. For example, Cooney and Mortimer (1999) found that household responsibilities accelerates early home leaving, and accounts for some of the difference between single- and two-parent families in the timing of home leaving. The authors also note that women tend to take on more responsibilities within the household, and that these responsibilities are associated with women’s but not men’s early home leaving patterns. The effect of household responsibilities on adult identity formation may likewise be stronger for young women.

METHODS

Data

This research employs data from the National Longitudinal Study of Adolescent Health (Add Health). Add Health is a nationally representative study of U.S. adolescents in grades 7–12 from 134 middle and high schools in 80 communities. Students were selected using a stratified sampling technique. Schools were selected from a complete list of high schools (Quality Education Database) based on their region, urbanicity, school type (public vs. private, racial composition, and size). Each of the selected high schools was matched to a feeder school (typically a middle school), with the probability of the feeder school being selected proportional to its contribution to the high school’s student body. Data collection began with an in-school questionnaire, administered to all students present in selected schools in 1994–95. A sub-sample of students (and one parent or parent-like figure) was then selected for in-depth interviews at home (n=20,745). These Wave I respondents were re-interviewed in 1996 (Wave II) and again in 2001–2002 (Wave III). A total of 15,197 original respondents were re-interviewed in Wave III.

The analytic sample in this research includes respondents who were interviewed in both Wave I and III and who were not already living with a spouse or partner at the Wave I (adolescent) interview (n=13,673). Family structure, resources, and processes were measured in Wave I when respondents were 12 to 17 years old. Adult identity was measured in Wave III when respondents were 18 to 26. We expect the effects of adolescent family context on subjective age to begin in adolescence, although we cannot assess them at that time because subjective age was measured for the first time in Wave III. Instead, we assess them several years later at a time in which some young people have begun to view themselves as adults and others have not (Arnett, 2004)ⁱ.

Measures

Self-perceived adult identity is based on the following survey question asked in Wave III: “How often do you think of yourself as an adult?” (0=never; 1= seldom; 2=sometimes; 3=most of the time; 4=all of the time). We measure self-perceived adult identity as a dichotomous variable coded “1” if a respondent reported feeling like an adult “all of the time” otherwise “0”. Descriptive statistics for adult identity and other study measures appear in Table 1. Just under half of young people reported feeling like an adult all of the time. The validity of this measure is demonstrated through relationships to theoretically relevant developmental characteristics, such as independence and maturity, and to adult role behaviors such as financial independence, leaving school, starting full-time work, and family formation (Johnson et al., 2007).

Table 1.

Description and Summary Statistics (weighted) of Study Measures

			Entire Sample (n=13,673)		Men (n=6502)		Women (n=7171)
Variable	Range	α	Mean/Prop	s.d	Mean/Prop	s.d	Mean/Prop	s.d
Self-perceived adult identity	0,1		40%		39%		42%	*
Family Structure
Two biological parent family	0,1		57%		58%		57%
“Step-parent” family	0,1		16%		16%		16%
”Adoptive” two parent family	0,1		1%		1%		1%
Single mother family	0,1		19%		18%		19%
Single father family	0,1		3%		3%		3%
All other	0,1		4%		4%		4%
Missing family income	0,1		0.20	0.01	0.20	0.01	0.21	0.01
Family income (log)	0–6.91		3.58	0.03	3.57	0.03	3.59	0.03
Parent education	1–8		5.31	0.08	5.32	0.09	5.31	0.09
Number of siblings	0–4		1.37	0.03	1.38	0.03	1.37	0.03
Parent-adolescent conflict	0,1		0.40	0.01	0.36	0.01	0.43	0.01***
Parent-adolescent closeness	0–4	0.83	3.20	0.02	3.28	0.02	3.12	0.02***
Parental control	0–7	0.63	1.86	0.05	1.85	0.06	1.89	0.05
Parental presence	0–4	0.36	2.77	0.02	2.79	0.02	2.76	0.02
Peer-like communication	0,1		0.65	0.01	0.58	0.01	0.72	0.01***
Household responsibility	0–4		2.04	0.02	1.93	0.02	2.14	0.02***
Male	0,1		51%
Race/ethnicity
Non-Hispanic White	0,1		67%		67%		68%
Black	0,1		16%		16%		16%
Hispanic	0,1		12%		12%		11%
Asian	0,1		3%		3%		3%
Other	0,1		2%		2%		1%
Age	18–26		21.76	0.12	21.85	0.12	21.67	0.12

Note: Asterisks note results of t-tests of mean differences by gender

^***p<.001;

^**p<.01;

^*p<.05.

All independent variables are measured at Wave 1.

Source: National Longitudinal Study of Adolescent Health. n=13,673

Family structure is measured in this analysis as a six-category variable: two biological parents, “step-families” (including all families with a biological parent and the parent’s spouse or cohabiting partnerⁱⁱ), “adoptive” two-parent familiesⁱⁱⁱ (including all other two parent families types such as two adoptive or foster parents), single biological mother families, single biological father families, and “all other” families (including a diverse array of single, non-biological parent family types).

Our measures of family resources include family income and parent education. The natural log of family income is a continuous measure based on parents’ self-reports. Parent educational attainment is measured by the highest of mothers’ and fathers’ educational attainments (1=completed 8^th grade or less; 8=graduate or professional training). When only one parent’s educational attainment was available, it serves as the final value, and when neither was available, we substitute students’ reports. We control respondents’ number of siblings, top-coded at 4, which arguably influences the amount of resources available to the adolescent within the family.

We consider several dimensions of family process. Parent-adolescent conflict is coded “1” if a respondent reported having an argument with either parent in the past four weeks, otherwise “0”. No measure capturing a higher degree of conflict was available. Parent-adolescent closeness is the mean response to five items (0=low to 4=high) that asked adolescents about the closeness, satisfaction, warmth, caring, and communication in relationships with their parents (alpha=0.83). We first created separate scales for closeness to residential mother and to residential father (using information for a non-residential parent only when information for a residential parent was missing). When measures of closeness to mother and father were both available, we select the least close because we anticipate that the presence of a poor relationship might affect family dynamics even if the adolescent also had a close relationship with another parent (Cooney & Mortimer, 1999).

We measure parental monitoring with indicators of parental control and presence. Parental control is an additive index ranging from 0 to 7 (alpha=0.63), and it is based on seven items that asked adolescents if their parents allowed them to make their own decisions regarding 1) weekend curfew, 2) friends, 3) clothes, 4) amount of television watching, 5) TV show selection, 6) weeknight bedtime, and 7) what food to eat. Each item is reverse coded so that higher values indicated greater parental control. Parental presence (see Sweeney, 2007; Cavanagh, forthcoming) is an additive index ranging from 0 to 4, and it is based on whether at least one parent was present in the home most or all of the time the adolescent went to school in the morning, came home from school in the afternoon, ate evening meals (5–7 dinners a week), and went to bed at night (alpha=0.36).

Social position within the family is captured in this analysis by examining peer-like communication between parents and adolescents. Peer-like communication is a dichotomous measure based on whether adolescents report talking with at least one of their parents about intimate issues they would typically talk about with their peers, including personal problems, people they are dating, and/or parties they go to. We contend that in families with more hierarchical parent-adolescent relationships adolescents would be less likely to freely discuss such intimate details of their lives.

Finally, we use participation in household chores as a proxy for shared responsibility within the household. Household responsibility is based on the number of times in the past week adolescents did “work around the house, such as cleaning, cooking, laundry, or caring for a pet” (0=not at all to 4=5 or more times).

We also include several demographic controls in our analysis, including racial/ethnic group, age, and sex. Race/ethnicity is measured by self-report and distinguished by 5 major groups: Hispanic American, non-Hispanic white, non-Hispanic black, Asian American, and Native American or other race. Age is measured in years at Wave III, and sex is dichotomous variable coded “1” for male and “0” for female. We also considered birth order, but as it was not related to adult identity, we exclude it here.

Due to the complex sampling design of Add Health, we weight all analyses and use the Huber/White estimator of variance to calculate corrected standard errors (see Chantala & Tabor, 1999). We use listwise deletion to handle missing data on all variables with the exception of family income, which had a non-trivial percent of cases missing. For data missing on family income, we use mean substitution, and we include a dummy variable for missing family income in our models.

RESULTS

Differences in self-perceived adulthood and family characteristics across the family structure types are shown in Table 2. As expected, respondents who were living with their two biological parents in Wave I are less likely to feel like adults all of the time in Wave III compared to those from all family types, with the exception of “adoptive” two parent families. Respondents who lived in the residual family structure category (“all other” family structures) at Wave I were most likely to self-identify as adults.

Table 2.

Means on Focal Study Measures by Family Structure

	Two biological parent family	“Step- parent” family	“Adoptive” two-parent family	Single mother family	Single father family	All other family structures
Self-perceived adult identity	.36	.45a	. 39	.47a	.46a	.52abcd
Family Resources
Family income	3.77	3.59a	3.78b	3.07abc	3.50acd	3.25abcde
Parental education	5.58	5.41a	5.71	4.71abc	4.99abc	4.20abcde
Family Processes
Parent-adolescent conflict	.39	.44a	.41	.40	.42	.28abcde
Parent-adolescent closeness	3.19	2.99a	3.22b	3.41abc	3.17bd	3.21bd
Parental control	1.92	1.89	2.00	1.77a	1.39abcd	1.78e
Parental presence	2.95	2.70a	3.00b	2.42abc	2.03abcd	2.82de
Peer-like communication	.62	.71a	.76a	.67ac	.71a	.66c
Household responsibilities	2.04	2.06	2.10	2.02	2.03	1.88abcd

Note: Means significantly different (p<.05) from

^atwo biological parent family,

^b“step-parent” family,

^c“adoptive” two-parent family,

^dsingle mother family, and

^esingle father family.

Family resources and processes differ by family structure in ways consistent with the findings of prior studies. In terms of resources, two biological parent families and “adoptive” two parent families are the most advantaged. Adolescents growing up in single mother and “all other” family structures experienced the lowest incomes and parent education levels. In regard to parent-adolescent relationship quality, step-parent families had the highest levels of conflict and the lowest levels of closeness. Parental control and presence tended to be highest in two biological parent families and “adoptive” two-parent families. Adolescents in single parent families, especially single father families, reported lower levels of parental control and presence. Peer-like communication occurred less frequently in two biological parent families than in any other family structure. Few differences were observed, however, in the housework involvement of adolescents across family structures.

Our first model predicting adult identity controls only respondents’ demographic characteristics. Consistent with the bivariate pattern, young people who had lived in “step-parent,” single mother, single father, and “all other” families are more likely to feel like adults all of the time compared to those from two biological parent families. Those in single mother families have 36 percent greater odds, those in single father families have 43 percent greater odds, those in step-parent families have 44 percent greater odds, and those in “all other” family structures have 51 percent greater odds of perceiving themselves as adults than those in two biological parent families. Differences amongst the non-two biological parent family types are not statistically significant.

Model 1 also documents differences in self-perceived adulthood by gender and age. Males are 13 percent less likely to feel like adults all of the time compared to females. And with each year older, the net odds of feeling like an adult increased 14 percent. It is important to note that although the tendency to identify as an adult increases with age, at no age does a majority report feeling like an adult all the time. Approximately 28% of 18–19 year olds feel adult all the time; about 46% of 25–26 year olds do so.

In Model 2 we introduce measures of family resources to evaluate their impact on adult identity and whether they mediate the effects of family structure. Greater family resources, as captured by parental educational attainment and family income, reduce the likelihood of feeling like an adult. They also attenuate the effects of adolescent family structure. Differences in family resources by family structure fully explain the effects of living with a single father and in an “all other” family structure in adolescence and substantially mediate the effect of living with a single mother. Family resources also somewhat attenuate the effects of living in a step-parent family during adolescence^iv.

We introduce the measures of family process in Model 3. Higher levels of family closeness and parental control reduce the likelihood of feeling like an adult. Peer-like communication with parents, parent-adolescent conflict, and more frequent housework, in contrast, facilitate feeling like an adult. These measures partially mediate the effects of having lived in a step-parent family on adult identity^v, but family structure differences remain statistically significant. Notably, the effect of gender is completely explained by family processes. In a series of models (not shown) we examine the family process measures singly and in combination and find that parental control, parent-adolescent conflict, and household responsibilities account for this reduction. As adolescents, girls shouldered more frequent household responsibilities, had less close relationships with their parents, and experienced more conflict with their parents (see Table 1). These differences in family experiences between adolescent boys and girls are clearly tied to identity development.

In Model 4 we consider both family resources and processes together. The results confirm the findings of Models 2 and 3. Family resources explain the effect of living with a single parent^vi and in an “all other” family structure on self-perceived adulthood. Family processes and resources together partially mediate the effects of living in a step-parent family (about 9%).

In order to evaluate our hypotheses about gender differences in the effects of family structure and process on self-perceived adulthood, we examine interaction terms between gender and family structure and between gender and each family resource and process measure. And in order to evaluate whether the family resources and processes operate similarly across family structure, we also examine interactions between these factors. These models indicate similarity in the effects of family structure and process on the formation of an adult identity. Only four differences achieve statistical significance, and the final model in Table 3 includes these four interaction terms together. First, the effect of living in an “all other” family structure during adolescence, compared to a two biological parent family, is larger for females than males. Stratified models by gender indicate that living in an “all other” family structure facilitates self-identifying as an adult for both sexes, but that the effect is larger for young women than for young men. Second, the facilitating effect of higher levels of parent-adolescent conflict on adult identity is again stronger for females than males. Stratified models by gender indicate that parent-adolescent conflict is not a significant predictor of self-perceived adulthood for young men, but it is for young women. Finally, we find that growing up in single- father households or “adoptive” two parent households is more closely associated with self-perceived adulthood when parental control is low.

Table 3.

Logistic Regression Models of Self-Perceived Adulthood on Family Structure, Resources, and Processes (N=13,673)

	Model 1			Model 2			Model 3			Model 4			Model 5
	B	SE	eβ	B	SE	eβ	B	SE B	eβ	B	SE	eβ	B	SE B	eβ
Family Structure
Two biological parent family			--			--			--			--			--
“Step-parent” family	0.37	0.06	1.44***	0.34	0.06	1.41***	0.34	0.06	1.40***	0.30	0.06	1.35***	0.30	0.06	1.35***
”Adoptive” two-parent family	0.14	0.24	1.15	0.14	0.24	1.15	0.11	0.24	1.11	0.11	0.24	1.11	0.68	0.32	1.97*
Single mother family	0.30	0.07	1.36***	0.14	0.07	1.15*	0.33	0.07	1.38***	0.13	0.07	1.14	0.13	0.07	1.14
Single father family	0.36	0.16	1.43*	0.24	0.17	1.27	0.35	0.16	1.42*	0.19	0.17	1.21	0.49	0.22	1.63*
All other family types	0.41	0.14	1.51***	0.19	0.15	1.21	0.43	0.14	1.54***	0.20	0.15	1.22	0.65	0.19	1.92**
Family Resources
Family income (log)			--	-0.11	0.03	0.90***			--	−0.12	0.03	0.89**	−0.12	0.03	0.89***
Parent education			--	-0.11	0.01	0.90***			--	−0.12	0.01	0.89***	−0.12	0.01	0.89***
Family Processes
Parent-adolescent conflict			--			--	0.14	0.05	1.15**	0.14	0.05	1.15**	0.26	0.07	1.30***
Parent-adolescent closeness			--			--	−0.08	0.04	0.92*	−0.07	0.04	0.93	−0.07	0.04	0.93
Parental control			--			--	−0.05	0.02	0.95*	−0.06	0.02	0.94**	−0.05	0.02	0.95*
Parental presence			--			--	0.05	0.03	1.05	0.02	0.03	1.02	0.01	0.03	1.01
Peer-like communication			--			--	0.21	0.05	1.23***	0.24	0.05	1.27***	0.25	0.05	1.28***
Household responsibilities			--			--	0.09	0.03	1.09**	0.11	0.03	1.11***	0.11	0.03	1.11***
Male	−0.14	0.05	0.87**	−0.14	0.05	0.87***	−0.07	0.05	0.94	−0.06	0.05	0.94	0.08	0.07	1.08
Age	0.13	0.02	1.14***	0.13	0.02	1.13***	0.11	0.02	1.11***	0.10	0.02	1.10***	0.10	0.02	1.10***
Interactions
Male * All other family types													−0.86	0.24	0.42***
Male * Parent-adolescent conflict													−0.25	0.10	0.78*
Single father family * Parental control													−0.22	0.09	0.81*
”Adoptive” family * Parental control													−0.31	0.12	0.73*

Note: Controls are race/ethnicity, number of siblings, and missing income (omitted from the table). eβ=exponentiated β.

^***p<.001;

^**p<.01;

^*p<.05

DISCUSSION

Families play an important role in their children’s transition to adulthood. This research contributes to our understanding of the subjective side of the adult transition by providing insights about how adolescent family context influences identification as an adult in the years following. Consistent with identity theory (Stryker & Serpe, 1994; Erikson, 1968), age-related identities are developed within social contexts, and the family is one of the most fundamental contexts for adolescents engaged in the autonomy-building that leads to adulthood. Our results show that family structure, resources, and a range of family processes all serve to constrain and promote the formation of adult identity.

Young people who lived in non-two biological parent families, except those in “adoptive” two parent families, were more likely to perceive themselves as adults than those from two biological parent families. These differences can be largely explained by the different economic resources and family processes observed across family structures. Family resources completely accounted for the differences in self-perceived adulthood between single-father and “all other” families and two-biological parent families, and they also largely reduced the effects of living in a single mother household. Family resources and processes both explained some but not all of the difference between two biological parent families and step-parent families. That they did not fully explain the difference may be due in part to our inability to fully capture family processes, an issue we discuss further below. These results suggest that accelerated subjective adulthood within single-parent and “all other” family types is driven by the lack of economic resources while in step-parent families it is also at least partly due to family processes, such as relationship quality, responsibilities, and monitoring.

These findings are largely consistent with literature on adolescent development and family structure more generally. Although research links child outcomes in single-parent families with family processes, studies also point to the discrepancies in economic resources between single- and two-parent families to account for these differences (McLanahan & Sandefur, 1994; Demo & Acock, 1996). Adoptive families provide similar levels of resources and social support to their children as two biological parent families (Schwartz & Finely, 2006; Lansford et al., 2001), which may account for the lack of difference between these two family types. Studies show that the effects of step-families on adolescent development tend to be due to a combination of family process factors, such as conflict, low monitoring and role ambiguity (Amato, 1993; Demo & Acock, 1996; Bray & Berger, 1993). We suspect that higher parent-adolescent conflict and greater role ambiguity in step-families accounts for the observed differences and that our measures of family process simply did not capture these experiences fully.

Our results also confirm that greater family resources are associated with subjective age in young adulthood (Johnson et al., 2007; Benson & Furstenberg, 2007), and offer new evidence showing that family processes also facilitate or hinder age-related identity development. Consistent with our hypotheses, parent-adolescent conflict and warmth were associated with young people’s adult identity. While high-conflict families promote self-identifying as an adult, families marked by a high level of closeness limit it. As developmental research suggests, the level of warmth and conflict within the parent-adolescent relationship may signify the type of attachment adolescents have with parents (Zimmer-Gembeck, 2001). Adolescents in conflict-ridden families may prematurely detach from their parental figures in attempt to separate themselves from a negative family context. As a result of this detachment, these youth develop self-reliance, although it may be through potentially negative pathways such as involvement in age-inappropriate behaviors.

Consistent with studies showing that girls are more sensitive to family conflict (Cooney & Mortimer, 1999), we also find that adolescent-parent conflict facilitates women’s but not men’s self identification as an adult. Future research is needed to identify the specific pathways through which early detachment may lead to pre-mature age identities, especially for young women^vii. In addition, we need to more fully understand how gender differences in adolescent development shape later adult identification. Our results show that adolescent men and women experience different family contexts, particularly in regard to family responsibilities and parent relationship quality, and these differences fully account for differences in subjective adult identity in the young adult years.

The results also provide some support for our hypothesis that high levels of parental monitoring and control inhibits adult identity formation. We found that parental control but not parental presence was associated with self-perceived adulthood. Through tight monitoring of adolescent activities, parents not only limit adolescents’ autonomy but also signal that they do not (yet) trust them to make their own choices. Close parental control may also delay onset of adult identity through limiting adolescents’ exposure to pre-mature and age inappropriate activities.

The findings also provide support for our hypothesis regarding hierarchical relationships within the family and family responsibilities. Adolescents who have a more peer-like relationship with their parents were more likely to feel like an adult in the years following than adolescents growing up in more hierarchal families. It appears that the blurring of generational boundaries does have implications for how young people perceive themselves (Nock, 1988; Weiss, 1979). By treating a young person as a peer rather than as a subordinate, parents signal that they see their adolescent children as adults like themselves. Consistent with previous research (Benson & Furstenberg, 2007), we also found that higher levels of household responsibilities promoted the development of an adult identity. Household labor provides an opportunity for young people to develop competence and mastery (Elder, 1968), but for youth who take on an unusually high level of responsibilities, this labor can force them to grow up more quickly than their same aged peers (Burton, 2007).

These family processes largely operated in similar ways across family structures. The only significant differences we found were that the negative effect of parental control on self-perceived adulthood was stronger in single-father and “all other” families. Single-father households typically offer less monitoring and control compared to two biological parent households, and this lack of monitoring, in turn, may facilitate seeing oneself as adult. When control and monitoring is present in single father families, it makes a big difference. Levels of control and monitoring are not lower in “other” family structures (see Table 2), however, indicating that other explanations need to be examined.

The Add Health data provide a unique opportunity to take a closer look at adolescent family context and its influence on adult identity formation, but it is important to acknowledge that these data are limited in several ways that restrict our conclusions and ultimately call for additional research. Our examination of family context is limited in that it is a snapshot representation. For our purposes, adolescence is the best time for such a snapshot, but no such measure can capture the complex family histories we know children and adolescents experience. Having identified primary distinctions in family structure and the type of family processes that predict adult identity among young people, however, this study provides an important foundation upon which more dynamic examinations of family structure can build.

Further limitations stem from several measures of key concepts that were not as strong as we would have liked. We tapped into shared responsibility in the family rather narrowly through housework contributions and were not able to measure other types of responsibilities including carework, participation in household decision making and the like. Our measure of hierarchical parent-adolescent relations was also indirect, focusing on intimate peer-like communication, and our measure of conflict with parents was overly simplified, unable to distinguish highly conflictual relationships. In each case we were able to document associations between these family processes and the formation of an adult identity, but it remains to be seen whether more complete measures would demonstrate stronger effects and account for more of the variation by family structure. Each of these measures also represents the adolescents’ perceptions of family processes, which may differ from those of other family members or outside observers. Finally, we capture perceived adult identity through a single item. While several studies have successfully used this approach (Johnson et al., 2007; Benson & Furstenberg, 2007), future research could benefit from a more contextualized measure of adult identity.

Despite these limitations, this research contributes to our understanding of family context and the subjective side of the transition to adulthood. Scholars stress the increasing importance of family background in the adult transition (Furstenberg et al., 2004; Settersten, forthcoming) but have not fully identified why and how family context matters. The results of this study provide new information about the ways in which families provide youth with powerful cues and signals about their own development, and opportunities to exercise an emerging sense of self as an adult. In addition, our results suggest that adolescent family socioeconomic context has important and far-reaching implications for young adult development. Today, families are increasingly called upon to provide financial support well into the young adult years (Furstenberg et al. 2004), and our study suggests that youth from families with lower socioeconomic resources, that may not provide the same type of financial safety net, often enter adulthood far earlier than their more advantaged peers. Scholars need to examine how this type of precocious development may impact health and well-being in the later adult years.

In addition, these results highlight that identity formation is a longitudinal developmental process. Future research needs to attend to the dynamics of age identity as it plays out over the years of adolescence into adulthood, including more specific attention to the pace of subjective aging and cyclical movement in feeling older and younger (Settersten, forthcoming). Future research on young adult identity development also needs to take into account how multiple developmental contexts during adolescence, such as family context, peer relationships, pubertal development and psychosocial functioning, shape the way youth come to define themselves as adults.