The Importance of Family, Race, and Gender for Multiracial Adolescent Well-being

Abstract

Using data from the National Longitudinal Study of Adolescent Health (Add Health), this study investigates patterns of well-being among multiracial adolescents. Specifically, this article addresses three questions. First, using various categorizations for multiracial background, are there measurable differences in emotional and social well-being among White, minority, and multiracial adolescents? Second, do multiracial adolescents with a White mother tend to fare differently than those with a minority mother? Third, does variation in family-based social capital—including parental involvement, parent-child relationship quality, and family structure—contribute to observed well-being differences among multiracial and monoracial adolescents? Results suggest that multiracial adolescents experience more negative social and emotional well-being outcomes when their mother is a minority. This finding persists even when controlling for sources of family-based social capital.

As of 2000, the U.S. Census began allowing individuals to identify themselves as belonging to more than one racial category. This reflects an important demographic development. Not only is the multiracial population growing at a significant rate, but also, socially, multiracial identity is becoming more acceptable, and its acknowledgment by the Census Bureau is just one piece of evidence supporting this trend. A considerable amount of research on multiracial individuals and households has been conducted over the last 20 years, including investigations of the self-identification process of multiracial children (D. R. Harris & Sim, 2002), the demographic characteristics of the multiracial household (Chew, Eggebeen, & Uhlenbeen, 1989), and variation in the salience of multiracial identity across settings (D. R. Harris & Sim, 2002; Jaret & Reitzes, 1999). Some studies suggest that multiracial adolescents tend to experience relatively poorer outcomes than their monoracial non-Hispanic White (hereafter referred to as White) counterparts but relatively better outcomes than their monoracial minority counterparts (D. R. Harris & Thomas, 2002). The underlying processes responsible for these group differences in well-being, however, are not well understood.

This article attempts to provide a better understanding of the determinants of well-being among multiracial adolescents using data from a nationally representative sample of adolescents in the United States, the National Longitudinal Study of Adolescent Health (Add Health). The analysis pays particular attention to whether and how parent-child relationships and the intersection of gender and racial identity of parents contribute to patterns of social and emotional well-being among multiracial adolescents. The analysis addresses three specific questions. First, are there measurable differences in emotional and social well-being among White, minority, and multiracial adolescents? Following Campbell and Eggerling-Boeck (2006), four specific measures of social and emotional well-being are considered: depression, serious consideration of suicide, feeling of social acceptance, and connection to school. Second, do multiracial adolescents with a White mother tend to fare differently than those with a minority mother? Third, does variation in family-based social capital—including parental involvement, parent-child relationship quality, and family structure—contribute to observed well-being differences among multiracial and monoracial adolescents?

Why Multiracials?

Aside from the growing number of individuals with recognized multiracial heritage in the United States, there are a number of reasons why this population is important to study and may be susceptible to particular risk. For example, in our highly racially stratified society, multiracials occupy a distinct category that is not congruent with other single-race populations. This differential categorization makes it harder for multiracials to assimilate with either of their single-race heritages (Campbell & Eggerling-Boeck, 2006; Park, 1928). This disconnect may leave multiracials lacking the social capital of their monoracial counterparts and heighten the risk of poor well-being outcomes (Coleman, 1988; Park, 1928). Furthermore, despite the occurrence of biracial unions throughout U.S. history, these types of unions have been generally uncommon and stigmatized. We know multiracial families are on the rise: In the 1970s, 1 in 100 children was born to parents of different races, and in recent years this figure has increased to 1 in 19 children. Yet social stigma still exists for biracial couples, and a primary source of unease comes from relatives who are uncomfortable with the unions producing children (O’Donoghue, 2004; St. Jean, 1998). These findings suggest the potential for rejection of multiracial families by extended family and community members; this rejection could lead to social isolation and distress (Campbell & Eggerling-Boeck, 2006).

What Do We Mean by “Multiracial”?

The term “multiracial” is very broad and, like any racial category, is socially constructed, and its meaning can change over time and across contexts (D. R. Harris & Sim, 2002; Rockquemore & Laszloffy, 2005). Although most people in the United States have had some type of racial, ethnic, or cultural mixing at some point in their family history, relatively few recognize themselves as multiracial, biracial, or multiethnic. Before the 1980s, the majority of researchers and the general public alike understood race in terms of the “one-drop rule,” such that multiracials were identified exclusively by their minority heritage (Roth, 2005). It was not until the late 1980s that individuals and researchers reconceptualized the understanding of race and began to include a category for those of multiple race heritages. Although this was a move toward more accurately describing individuals’ racial characteristics, it also led to the stigmatization of those of multiple race heritages who claimed a single-race identity. More recently, sensitivity to this issue has led researchers to recognize the legitimacy of the variety of ways in which multiracial individuals choose to identify themselves (Rockquemore & Laszloffy, 2005).

In this study, the understanding of multiracial identity is based on self-reported information from children and parents. The term “multiracial” is used to refer to youth who have any multiple race heritages. The study focuses particularly, however, on multiracials with both a minority racial background and a White background. This is done because the majority of African Americans, Asians, Hispanics, and Native Americans who interracially married have White partners (Bratter & Eschbach, 2006). Because the experience of a multiracial identity may vary considerably across specific race groups (e.g., Jaret & Reitzes, 1999) and for children, depending on the racial and gender location of parents (Roth 2005), these factors are taken into account when considering outcomes associated with multiracial identity.

Multiracials and Well-being

Historically, scholars depicted the well-being outcomes for multiracial populations as being particularly bleak. In the 1920s, Park (1928) concluded that multiracials experience particular disadvantage, as they lie on the margin of two cultures and lack connection to either. This position became known as the “Marginal Man Hypothesis.” Since then, further research on the hypothesis that Park proposed almost a century ago paints a more complex picture of well-being among multiracials. For example, Cheng and Lively (2009) found that, although multiracials do have relatively poorer psychological well-being outcomes than their monoracial peers, they also have somewhat more positive well-being outcomes related to social interaction with peers. Specifically, multiracials tend to report a similar or greater number of friends to discuss problems with as do monoracials but report the quality of these relationships to be lower than do monoracials (Cheng & Lively, 2009). For general health and risk-taking behaviors, multiracial adolescents fare worse than monoracial adolescents. For example, using data from the National Longitudinal Study of Adolescent Health, Udry, Li, and Hendrickson-Smith (2003) found that individuals who identify with more than one race are at higher risk than those who identify with one race with respect to general health, school experience, smoking, and drinking. This study suggests that overall disadvantage could exist for the multiracial population.

Other research has found no well-being differences at the aggregate level for multiracials but has found differences between specific multiracial groups once they are examined separately. For example, Campbell and Eggerling-Boeck (2006) examined social and emotional well-being and found no overall difference for multiracials compared to monoracials as aggregate groups; however, they did find race-specific disadvantages. Specifically, they found that Native American-White adolescents have significantly worse outcomes compared to their monoracial White counterparts (Campbell & Eggerling-Boeck, 2006). D. R. Harris and Thomas (2002) found that for school achievement, multiracials often fall between their minority and White counterparts, depending on the specific outcome assessed. Black-White adolescents do fall between their monoracial counterparts for vocabulary test scores but resemble Whites for grade-point average (GPA). Meanwhile, Asian-White adolescents have vocabulary scores similar to those of monoracial Whites, yet have worse GPAs than either of their monoracial counterparts. Binning, Unzuta, Huo, and Molina (2009) complicated the story further by showing that psychological well-being for multiracials differs depending on the self-label that a multiracial associates with. Specifically, if multiracials identify with a multiple racial group label, they have lower stress levels and lower levels of alienation from peers than do multiracials who self-label as monoracial (Binning et al., 2009).

Why Might Well-being Differences Exist?

There are a number of reasons to expect well-being of youth in multiracial families to differ from that of youths in monoracial families. For example, systematic differences exist across race groups in social background and economic characteristics. Couples in an interracial union have a higher chance of union dissolution than do those in a racially homogenous union (Bratter & King, 2008). Additionally, although parents in multiracial families tend to have comparable or higher levels of education than do parents in monoracial families, they tend to have a lower family income than monoracial Whites but a higher family income than monoracial minorities (Cheng & Powell, 2007; Chew et al., 1989).

Prior research on well-being outcomes for multiracials has also supported the claim that accounting for the race and gender combination of multiracial parents is important. For example, gender differences exist in marital instability among biracial couples. Compared to monoracial White couples, White female/Black male and White female/Asian male marriages are more prone to divorce (Bratter & King, 2008). Additionally, Chew and colleagues (1989) analyzed the general characteristics of biracial households and found that the socioeconomic status of multiracial households tends to fall somewhere between that of the average two-parent minority households and two-parent White households. Important differences emerge, however, once the gender and race of each parent are taken into account. For example, among Black-White multiracials, White mother-Black father households have a lower family income than households with a White father-Black mother. Furthermore, with the exception of Asian father-White mother households, all biracial combinations have an average family income that is less than homogenous White families. These divorce and income disparities suggest potential for other measurable differences among multiracial families on the basis of gender location and racial category of parents within the households.

Race, Well-being, and Family-Based Social Capital

Social capital offers a useful framework for understanding how families contribute to youth well-being. Coleman (1988) proposed the idea of social capital as a resource that stems from social interactions where people can draw on social ties for help and support. Scholars often make a distinction between social capital generated through relationships within the family (intrafamily social capital) versus through connections family members have to outside social networks (extrafamily social capital). Measures of intrafamily-based social capital measures are geared toward capturing the ways in which adolescents are able to use ties with parents to access human and cultural capital resources. Research suggests that adolescents with parents who understand them, have positive educational expectations for them, and have high levels of family cohesion will have stronger bonds to their parents and will have more access to parental resources and thus improved well-being (Crosnoe, 2004; Yabiku, Axinn, & Thornton, 1999). Measures of extrafamily-based social capital are instead geared toward capturing the ways in which adolescents are able to use network ties to access community resources. Research here indicates that adolescents in two-parent households with parents who have high community involvement and strong network ties will have more access to community resourses and thus improved well-being outcomes (Coleman, 1988; Portes, 2000). Some evidence suggests that family-based social capital may be important to consider when trying to understand the association between race and adolescent well-being. For example, Furstenberg and Hughes (1995) found that at-risk, disadvantaged minority youth are able to use higher levels of both intra- and extrafamily-based social capital to diffuse the negative effects of poverty and teen pregnancy. If minority youth can use family-based social capital to diffuse disadvantage, multiracial youth might also be able to use this social capital. Yet, little prior research has explicitly considered this possibility. Social capital may be protective of emotional well-being among multiracial youth by increasing self-esteem that might be otherwise tarnished by racism and protective of social well-being through the creation of social ties that would otherwise be lacking from extended family abandonment. This study extends these previous findings to the multiracial population and investigates whether family-based social capital explains differences in well-being that were addressed in previous studies.

There is some evidence, however, that social capital available to multiracial children may vary depending on gender and race of parents. For example, a number of qualitative studies suggest that White mothers of half-Black children face the highest level of stigma of any biracial coupling (O’Donoghue, 2004; Winddance Twine, 1996), which may reduce children’s access to extrafamilial social networks and also affect the quality of parent-child bonds. This is especially salient for White mothers when they are dealing with their family of origin (Childs, 2005). Qualitative research documents the turmoil associated with their interracial unions, going as far as being disowned by family members (Byrd & Garwick, 2006). In addition to this extra stigma, White mothers have a difficult time relating to the effect of race and racism on their children (Reddy, 1994; Winddance Twine, 1996). This type of disconnect can be a disadvantage to the children, as they may have less maternal support if they do in fact experience racial discrimination.

This article builds on prior efforts to understand patterns of well-being among multiracial adolescents and the role of family-based social capital in generating well-being differentials across race groups. This research focuses on outcomes for multiracials during the period of adolescence for a variety of reasons. First, experience in adolescence has enduring effects throughout the life course. For example, early socialization in adolescence can affect such outcomes as age at marriage, adult psychological well-being, and long-term life satisfaction (Clausen, 1991). Furthermore, adolescence is a time when racial identity becomes particularly salient (Wind-dance Twine, 1996). In fact, with the onset of dating, multiracial individuals—sometimes for the first time—start to feel distinct from their peers and realize their own racial heritage (Winddance Twine, 1996). This suggests that even if multiracial identity is something that is fluid over time and place, studying well-being outcomes in adolescence is of particular importance.

In the study, three specific questions are addressed. First, the study explores whether measurable differences in emotional and social well-being exist among White, minority, and multiracial adolescents using two unique classifications of race. Second, this research investigates whether multiracial youths with a White mother tend to fare differently than those with a minority mother. Finally, this study uses multiple measures of family-based social capital (including parental involvement, parent-child relationship quality, and family structure) to better understand mechanisms underlying race differences in well-being.

A number of specific hypotheses are tested in the current research. First, based on Park’s (1928) “marginal man” hypothesis, adolescents in the multiracial category are expected to show the poorest well-being outcomes. Specifically, their outcomes should be worse than either their White or minority counterparts, as they will not be able to access social capital from either of their racial heritage groups and are more likely to be disconnected from their extended family (Childs, 2005). Second, it is hypothesized that multiracial adolescents with White mothers will have measurably worse well-being than monoracial White adolescents or other multiracials. As discussed above, previous research on White mothers in multiracial families suggests overwhelming disadvantage. For Black-White biracial families in particular, not only does the research state that White mother families will be more likely to come from households with lower family income, but that they will also experience higher levels of stigma, have more issues with their family of origin, and that their lack of previous experiences with racism will not allow them to help their child navigate racism. Finally, it is expected that the race-based well-being differentials described above will diminish once controls for family-based social capital are added. Family-based social capital has proven particularly useful for understanding the association between race and adolescent well-being. Specifically, family-based social capital has been shown to explain how adolescents overcome disadvantages associated with race. As suggested by previous research, it is expected that this will operate through self-esteem and social ties.

Method

Data and Sample

The National Longitudinal Study of Adolescent Health is a federally funded, nationally representative school-based sample of adolescents in grades 7 – 12 collected using stratified random sampling techniques that has gathered four waves of data starting in 1995 (K.M. Harris, 2011). The original goal of colleting these data was to better understand adolescents’ health and risk behaviors, including personal traits, families, friendships, romantic relationships, peer groups, schools, neighborhoods, and communities (K. M. Harris, 2011). Data from this sample were collected at home and at school from the adolescents and through parent surveys. This study utilizes the restricted access data from Wave I of the in-home and parent surveys. The in-home adolescent surveys yielded 20,745 respondents and the parent surveys yielded 17,700 respondents.

These data are particularly useful for studying the multiracial population because Add Health not only has a large sample size but it also allows the researcher to use both the adolescent’s self-reported race and parental self-reported race to classify multiracials. For the in-home adolescent survey and the parent survey, respondents were asked to report their race and told that they might give more than one answer. They might select White, Black or African American, American Indian or Native American, Asian or Pacific Islander, or other. Having both of these reports means that adolescent race can be established by adolescent self-report or by parent self-report. From adolescent self-report alone, 1,038 respondents identified themselves as multiracial (803 non-Hispanic multiracial respondents).

In this study, the analyses are restricted to a sample of non-Hispanic adolescents with at least one biological non-Hispanic parent. Specifically, individuals with Hispanic ethnicity for parents and respondents are omitted because in this survey this ethnicity was asked as a separate question from race, making it impossible to distinguish a multiracial Hispanic-White. This therefore makes it unclear if adolescents who were choosing Hispanic identity and a single race were considering themselves multiracial (Campbell & Eggerling-Boeck, 2006). Previous research suggests that when surveys are designed in this way it is not possible to know if responses are reflecting mixed racial heritage, national origin, or just the a pressure to answer both sets of questions (Campbell & Eggerling-Boeck, 2006; Rodriguez, 2000). Furthermore, D. R. Harris and Sim (2002) maintain that when using Add Health data, Hispanic multiracial identity cannot be measured and that it can only be accurately measured by a survey that combines a Hispanic race and ethnicity question. Finally, because Add Health only provides valid parent racial data for the adolescents who had at least one biological parent respondent, nonbiological parent respondents are omitted from the sample.

Because of concerns discussed in Binning et al. (2009), it needs to be acknowledged that the way in which multiracial identity has been operationalized could be flawed. Specifically, just because individuals check multiple race boxes does not mean that their psychological interpretation of their race matches their racial heritage selection (Binning et al., 2009). Additionally, multiple studies have focused on identification based on racial heritage instead of self-identification. Campbell and Eggerling-Boeck (2006), however, demonstrate that social and emotional well-being outcomes among multiracial adolescents do not significantly differ when using a racial classification based on self-identification versus parent racial heritage.

Race

Adolescent race is the primary analytic measure within this study. Specifically, shown in Table 1, three distinct race measures were constructed to address the research questions posed above. The first expanded categorization was separated into nine White-minority categories. These categories are based on adolescent self-reports and include monoracial White, monoracial Black, monoracial Native American, monoracial Asian, monoracial other, multiracial other, Black-White, Native American-White, and Asian-White. The category of multiracial other includes individuals with more than two racial heritages, meaning this category encompasses all other multiracials that did not simply have a White/minority background.

Table 1.

Race Variable Descriptions

Race Categorization	Description	Source
Detailed	Race broken into nine categories. Based on adolescent self reports. Monoracial White Monoracial Black Monoracial Native American Monoracial Asian Monoracial other Multiracial Black-White Multiracial Native American-White Multiracial Asian-White Other multiracials	in-home survey
Condensed	Race broken into three categories. Based on the same adolescent self-reports as detailed race variable. Monoracial White Monoracial minority Multiracial	in-home survey
Mother-specific racial status	Race broken into four categories. Based on a combination of adolescent self-report from the in-home and parent surveys. An inference for parent nonrespondent’s race for self-declared multiracial adolescents was made using adolescent self-reported race. Monoracial White Monoracial minority Multiracial: White mother Multiracial: Minority mother	in-home survey and parent survey

With that in mind, the second categorization of race takes on a more condensed form. These categories, also based on adolescent self-reports, include monoracial White, monoracial minority, and multiracial.

The final categorization of race uses a combination of adolescent self-reports and parent respondent self-reports. Specifically, parent race and adolescent race are matched, and adolescents are placed in one of four categories: mono-racial White; monoracial minority; multiracial, White mother; and multiracial, minority mother. If adolescents identified as multiracial and had a mother who identified as White, they were placed in the multiracial, White mother category. If adolescents identified as multiracial and had a mother who identified as minority, they were placed in the multiracial, minority mother category. Additionally, adolescents who identified as White-minority multiracial and had a minority father were inferred to have a White mother. All other monoracial adolescents were placed into either monoracial White or mono-racial minority categories based on their own self-reported race.

Social and Emotional Well-being

As discussed above, the outcome measures are divided into two categories, social and emotional well-being. These measures were previously used by Campbell and Eggerling-Boeck (2006) with the purpose of studying racial variation in well-being outcomes. Emotional well-being was measured by suicide ideation and a modified version of the CES-D depression scale. Suicide ideation was measured by adolescents’ response to the question of whether they had seriously considered suicide in the last 12 months. A dichotomous variable was created where a value of 1 means that the respondent had seriously considered suicide. Depression was measured using a 19-item “self-rating” scale intended to assess depressive symptoms that range from 0 to 3, with a high score indicating a high level of depression. The items for this scale were summed and logged (after 1 was added).

Social well-being was measured by perceptions of social acceptance and whether the adolescent feels socially integrated into the school. Specifically, social acceptance was measured by respondents’ agreement or disagreement with the statement “I feel socially accepted.” A dichotomous variable was created where a value of 1 means that the respondents strongly agree or agree with the statement. School engagement was measured based on the mean of the responses to two questions: The first asked how close the respondent feels to people at school, and the second asked if the respondent feels like a part of his or her school. The final average is coded on a 5-point scale, with a higher score indicating higher level of school engagement.

Family-Based Social Capital

Measures of social capital reflect both intra- and extrafamily-based social capital. Intrafamily-based social capital was measured by close family relationships, family understanding of the adolescent, family cohesion, parental educational expectations, and family structure. Close family relationships were measured by an item in which respondents were asked, “How much do you feel that your family pays attention to you?” Family understanding of the adolescent was measured by an item in which respondents were asked, “How much do you feel that people in your family understand you?” Family cohesion was measured by an item in which respondents were asked, “How much do you feel that you and your family have fun together?” For each of these measures a dichotomous variable was created where a value of 1 was assigned to the responses “quite a bit” or “very much.” Parental educational expectations were measured by parents’ response to the question, “How disappointed would you be if your child did not graduate from college?” A dichotomous variable was created where the value of 1 was assigned to parent respondents who would be very disappointed if their child did not graduate college. Family structure was measured by parent respondents’ report of parents living in the household. The variable was coded 1 for one having two biological parents in the household and 2 for one biological parent.

Extrafamily social capital was measured by parental community involvement and closure of parental networks. Parental community involvement was measured by the average of two questions asking whether parents would tell a neighbor if the neighbor’s child was getting in trouble, and whether a neighbor would tell them if their own child was getting in trouble. A dichotomous variable was created where the value of 1 (strong community ties) was assigned to a parent response of a combination of “probably would,” “would,” or “might” to both of the questions. Closure of parental networks was measured by parents’ response to the question, “How many parents of your child’s friends have you talked to in the last four weeks?” From this question an ordered categorical variable was created with four categories, 1 if parent knows no friends (weak), 2 if parent knows 1 – 2 friends (moderate), 3 if parent knows 3 – 4 friends (strong), 4 if parent knows 5 or more friends (very strong).

Controls

A basic set of controls was also included in each model: adolescent’s sex, mother’s level of education, and family income. The measure of adolescent’s sex was taken from the in-home child survey, and the measures for mother’s level of education and family income were taken from the parent survey. Controlling for sex is important, as previous research points to some sex-specific differences for various outcomes relating to multiracial children (Cooney & Radina, 2000). Controlling for mother’s level of education and family income provides a way of distinguishing race from class background (Elder, Eccles, Ardelt, & Lord, 1995).

Analysis

Methodologically, logistic regression models were estimated for binary outcomes (suicide ideation and social acceptance) and OLS regression models were used for continuous outcomes (depression and school engagement). Each of these regression models allows for simultaneous estimation of all of the coefficients included in the models. Models were estimated using the survey commands in Stata, which are specifically designed to account for the statistical requirements of complex large-scale survey data such as Add Health.

Results

Descriptive Statistics

Descriptive statistics, shown in Table 2, suggest a socioeconomic disadvantage for monoracial Native Americans and monoracial Black adolescents in this sample. For example, monoracial Blacks have the lowest family income, and monoracial Native Americans have the lowest levels of mother’s completed education. This table also suggests socioeconomic advantage for multiracial and monoracial Asians in the sample. For example, multiracial Asian-Whites have the highest levels of family income and monoracial Asians have the highest levels of mother’s education. The family-based social capital measure results are more complicated, with particular racial groups faring the best for some outcomes while simultaneously faring the worst on others. For example, multiracial Black-White adolescents have the highest reported levels of having a close relationship to their family. Multiracial Black-White adolescents are also, however, the most likely to be in a single-parent household and also to have the lowest reported rates of parental community involvement.

Table 2.

Descriptive Statistics, Detailed Race ( N = 9,222)

Variable	Total			White			Black			Native American			Asian			Other			Black-White			Native-White			Asian-White			Other Multiracial			Range
	%	M	SD	%	M	SD	%	M	SD	%	M	SD	%	M	SD	%	M	SD	%	M	SD	%	M	SD	%	M	SD	%	M	SD
Dependent variables
Emotional well-being
Depression		10.3	6.8		9.95	6.76		11.65	7.53		13.28	7.46		12.1	6.51		9.89	5.13		10.59	7.41		11.86	8.00		12.35	8.16		11.49	6.61	(0 – 50)
Has seriously considered suicide	0.13			0.13			0.11			0.24			0.11			0.07			0.13			0.17			0.21			0.22			(0 – 1)
Social well-being
Feels socially accepted	0.86			0.86			0.88			0.87			0.83			0.77			0.89			0.76			0.93			0.76			(0 – 1)
Engagement in school		2.20	0.84		2.19	0.85		2.24	0.88		2.11	0.96		2.13	0.71		2.38	0.74		2.24	0.85		2.25	0.90		2.19	0.87		2.32	1.00	(1 – 5)
Independent variables
Intrafamily social capital
Close family relationship	0.72			0.74			0.69			0.76			0.64			0.57			0.79			0.56			0.65			0.64			(0 – 1)
Family understanding	0.55			0.55			0.59			0.69			0.50			0.46			0.60			0.44			0.52			0.59			(0 – 1)
High parent educational expectations	0.40			0.36			0.53			0.51			0.80			0.55			0.32			0.41			0.31			0.42			(0 – 1)
Fun with family	0.63			0.64			0.66			0.62			0.54			0.62			0.56			0.58			0.61			0.55			(0 – 1)
Single parent household	0.39			0.34			0.66			0.57			0.17			0.54			0.85			0.37			0.40			0.40			(0 – 1)
Extrafamily social capital
Parent community involvement	0.74			0.72			0.83			0.76			0.68			0.64			0.69			0.83			0.80			0.81			(0 – 1)
Closure of parental networks
Moderate	0.40			0.39			0.43			0.40			0.47			0.19			0.35			0.4			0.44			0.41			(0 – 1)
Strong	0.23			0.25			0.19			0.16			0.15			0.14			0.25			0.25			0.35			0.18			(0 – 1)
Very strong	0.16			0.18			0.08			0.12			0.07			0.11			0.09			0.19			0.06			0.14			(0 – 1)
Control variables
Male	0.51			0.51			0.48			0.63			0.17			0.42			0.51			0.58			0.36			0.39			(0 – 1)
Mother’s education
Less than HS	0.11			0.09			0.18			0.26			0.21			0.25			0.15			0.14			0.16			0.22			(0 – 1)
HS or some college	0.65			0.66			0.18			0.26			0.32			0.56			0.56			0.74			0.42			0.55			(0 – 1)
College or more	0.24			0.25			0.17			0.11			0.47			0.19			0.29			0.12			0.42			0.23			(0 – 1)
Log of family income		3.59	0.78		3.7	0.74		3.05	0.89		2.79	0.7		3.73	0.76		3.13	0.85		3.3	0.63		3.67	0.70		3.83	0.67		3.35	0.76	(0 – 6.9)
n		9,222			6,153			2,088			41			462			39			90			146			67			136

Notes: Data are adjusted for the complex sampling design of Add Health. Source: National Longitudinal Study of Adolescent Health (Add Health).

Table 2 also illustrates preliminary evidence of variation in well-being across racial groups. Specifically, when adjusting for design effects, results suggest that for depression, monoracial others have the best outcomes (mean = 9.89), whereas Native Americans have the poorest outcomes (mean = 13.28). For suicide ideation, multiracial Native Americans have the worst outcomes, with approximately 24% considering suicide, and monoracial others have the most positive outcomes, with only 7% considering suicide. As for social acceptance, multiracial Black-White adolescents have the most positive outcomes, with approximately 89% feeling socially accepted, whereas multiracial Native American-Whites have the worst outcomes, with only approximately 76% feeling socially accepted. Finally, for school engagement, scaled from 1 to 5, Native Americans have the lowest levels with a mean score of 2.11, and monoracial others have the most positive outcomes with a mean score of 2.38.

Are There Overall Race Differences in Well-being?

In the analyses of whether measurable differences exist in the emotional and social well-being among White, minority, and multiracial adolescents when controlling for sex, family income, and mothers’ education, both the detailed and condensed racial categorizations were analyzed. As shown in Tables 3 and 4, some significant differences across detailed race groups in levels of well-being were found.

Table 3.

Multiracial Identification and Emotional Well-Being ( N = 9,222)

	Depression OLS Coefficient			Suicide Logit Coefficient
	Model 1	Model 2	Model 3	Model 1	Model 2	Model 3
Expanded race: Model 1a
Monoracial Black	1.11** (0.32)			−0.29** (0.10)
Monoracial Native American	2.72 (2.05)			0.73* (0.38)
Monoracial Asian	2.25** (0.82)			−0.19 (0.23)
Monoracial other	−0.76 (1.59)			−0.73 (0.76)
Multiracial other	0.98 (0.96)			0.52 (0.31)
Black-White	0.40 (1.51)			−0.01 (0.48)
Native American-White	1.83* (0.75)			0.36 (0.26)
Asian-White	2.32 (1.56)			0.50 (0.45)
Compressed Race: Model 2
Minority		1.26** (0.33)			−0.26** (0.10)
Multiracial		1.41* (0.60)1			0.36* (0.18)1
Mother-specific racial status: Model 3
Minority			1.26** (0.33)			−0.26** (0.10)
Multiracial: White mother			1.41 (0.74)			0.11 (0.26)
Multiracial: Minority mother			1.42 (0.81)			0.70** (0.26)2
Control variables
Male	−1.70** (0.22)	−1.68** (0.22)	−1.68** (0.22)	−41** (0.09)	−.40** (0.09)	−.40** (0.09)
Mother’s education (omitted: less than HS)
HS or some college	−0.89** (0.33)	−0.92** (0.33)	−0.92** (0.33)	−0.07 (0.59)	−0.07 (0.12)	−0.07 (0.12)
College or more	−1.87** (0.36)	−1.88** (0.36)	−1.88	−0.14 (0.15)	−.14 (0.15)	−.15 (0.15)
Log of family income	−0.59** (0.13)	−0.56** (0.13)	−56	−0.02 (0.05)	−0.02 (0.05)	−0.02 (0.05)
Results from adjusted Wald test
Model 1 vs. Model 2 (df = 6)		8.35, p < .05			7.15, p < .01
Model 2 vs. Model 3 (df = 6)			5.53, p < .05			5.74, p < .01

Notes: Standard errors are shown in parentheses. Data are adjusted for the complex sampling design of Add Health. Source: National Longitudinal Study of Adolescent Health (Add Health).

^aOmitted category is monoracial White.

¹Coefficient for “Multiracial” differs from that for “Minority” at the p < .05 level.

²Coefficient for “Multiracial: Minority mother” differs from that for monoracial “Minority” at the p < .01 level.

^*Coefficient differs from 0 at the p < .05 level.

^**Coefficient differs from 0 at the p < .01 level.

Table 4.

Multiracial Identification and Social Well-Being ( N = 9,222)

	School Engagement OLS Coefficient			Social Acceptance Logit Coefficient
	Model 1	Model 2	Model 3	Model 1	Model 2	Model 3
Expanded race: Model 1a
Monoracial Black	.01 (.03)			.34 ** (.13)
Monoracial Native American	−.14 (.20)			.16 (.39)
Monoracial Asian	−.06 (.05)			−.19 (.23)
Monoracial other	.15 (.24)			−.44 (.59)
Multiracial other	.10 (.14)			−.52 (.30)
Black-White	.03 (.14)			.34 (.43)
Native American-White	.04 (.11)			−.66* (.28)
Asian-White	.01 (.16)			.91 (.61)
Compressed race: Model 2
Minority		.00 (.03)			0.22 (.12)
Multiracial		.05 (.08)			−.34 (.19)
Mother-specific racial status: Model 3
Minority			.00 (.03)			.22 (.06)
Multiracial: White mother			−.02 (.09)			−.118 (.23)
Multiracial: minority mother			.16 (.10)			−.56* (.25)
Control variables
Male	−.01 (.03)	−.01 (.03)	−.01 (.03)	.34 ** (.08)	.34 ** (.09)	.34 ** (.09)
Mother’s education (omitted: less than HS)
HS or some college	−10* (.05)	−.10* (.04)	−.10* (.04)	.09 (.12)	.11 (.12)	.11 (.12)
College or more	−.01 ** (.05)	−.18 ** (.05)	−18 ** (.05)	.06 (.13)	.08 (.13)	.08 (.13)
Log of family income	−.03 (.02)	−.03* (.02)	−.03* (.02)	.11* (.05)	.09* (.05)	.09* (.05)
Results from adjusted Wald test
Model 1 vs. Model 2 (df = 6)		0.23			4.63, p < .05
Model 2 vs. Model 3 (df = 6)			1.08			3.51, p < .01

Notes: Standard errors are shown in parentheses. Data are adjusted for the complex sampling design of Add Health.

^aOmitted category is monoracial White.

^*Coefficient differs from 0 at the p < .05 level.

^**Coefficient differs from 0 at the p < .01 level.

With respect to depression, as shown in the first column of Table 3, multiracial Native American-White respondents have significantly worse outcomes compared to their monoracial White counterparts, meaning that being Native American-White versus being monoracial White results in a 1.83-point increase on the modified CES-D depression scale, holding constant sex, family income, and mother’s level of education. A Wald test, however, does not reveal that their outcomes significantly differ from monoracial Native Americans. When the compressed categorization of race for depression was used, Table 3 shows that minorities and multiracials both have worse outcomes than their monoracial White counterparts—meaning being minority versus being monoracial White results in a 1.26-point increase on the modified CES-D depression scale, holding constant sex, family income, and mother’s level of education. Also, being multiracial versus being monoracial White results in a 1.41-point increase on the modified CES-D depression scale, holding constant sex, family income, and mother’s level of education. A further Wald test shows that the multiracial coefficient differs from the minority coefficient at a p < .05 significance level.

With respect to suicide ideation, the second column of Table 3 reveals that no significant values were found for multiracial adolescents using the detailed categorization of race. Table 3 also reveals, however, that, when the compressed categorization of race for suicide ideation is used, differences by racial category for multiracial adolescents do exist. Specifically, monoracial minorities are less likely than monoracial Whites to consider suicide, and multiracials are more likely than monoracial Whites to consider suicide. Thus, being monoracial minority versus being monoracial White results in a .26 decrease in the log odds of considering suicide, holding constant sex, family income, and mother’s level of education. Also, being multiracial versus being monoracial White results in a .36 increase in the log odds of considering suicide, holding constant sex, family income, and mother’s level of education.

With respect to school engagement, the first column of Table 4 reveals that no statistically significant values were found for multiracial adolescents using the detailed categorization, compressed, or mother-specific categorization of race for school engagement.

As to the examination of differences in social acceptance, depicted in the second column of Table 4, no significant values were found using the compressed categorization of race. Also in Table 4, however, when the expanded categorization of race for social acceptanceis used, differences by racial category are noted. Specifically, multiracial Native American-White respondents have significantly worse outcomes (coefficient of −.66) compared to their monoracial White counterparts, meaning that being multiracial Native American-White versus being monoracial White results in a .66 decrease in the log odds of feeling socially accepted, holding constant sex, family income, and mother’s level of education. A Wald test, however, does not reveal that their outcomes significantly differ from monoracial Native Americans.

In summary, the above results provide support for the conclusion that there are measurable differences pertaining to emotional and social well-being between White, minority, and multiracial adolescents. Specifically, Tables 3 and 4 show that for social and emotional well-being, multiracial Native American-Whites have statistically significant and more negative outcomes than their monoracial White counterparts. Furthermore, for emotional well-being, the condensed multiracial category shows significant and more negative effects than its monoracial White counterpart.

Does Gender of the Minority Status Parent Matter?

Tables 3 and 4 illustrate the results of the regression analyses conducted to test the importance of gender and race of the parent for multiracial adolescents on adolescents’ reports of emotional and social well-being. With respect to emotional well-being, multiracial adolescents with a minority mother fare worse than both their monoracial minority and White counterparts. This is particularly relevant for suicide; multiracial adolescents with a minority mother are significantly more likely to consider suicide than monoracial Whites, meaning that being multiracial and having a minority mother versus being monoracial White and having a White mother results in a .70 increase in the log odds of considering suicide, holding constant sex, family income, and mother’s level of education. Furthermore, a Wald test reveals that this coefficient is significantly different from the coefficient for monoracial minority (coefficient of −.26) at a p < .05 significance level. The analysis also revealed, however, no significant emotional well-being differences for multiracial adolescents with a White mother.

The regression analysis of social well-being, summarized in Table 4, tells a similar story for multiracials with a minority mother. Specifically, for social acceptance, multiracials with a minority mother are less likely to feel socially accepted (coefficient of −.56) than their monoracial White counterparts, meaning that being multiracial and having a minority mother versus being monoracial White and having a White mother results in a .56 decrease in the log odds of feeling socially accepted, holding constant sex, family income, and mother’s level of education. Wald tests, however, do not provide results to suggest that this coefficient is significantly different from any other category. The analysis also revealed, however, no significant social well-being differences for multiracial adolescents with a White mother.

In sum, these results provide support for the claim that gender and minority status of parents is important in analyzing both social and emotional well-being outcomes for multiracial adolescents. Furthermore, it appears that there is a specific emotional disadvantage for multiracial adolescents who have a minority mother versus their monoracial counterparts.

Does Family-Based Social Capital Explain the Difference?

Finally, variations in family-based social capital, including parental involvement, parent-child relationship quality, and family structure, were examined, as these contribute to observed well-being differences among multiracial and monoracial adolescents. Comparing the results from Tables 5 and 6 to those from Tables 3 and 4 provides direct tests of this question.

Table 5.

Multiracial Identification and Emotional Well-Being Including Measures of Family-Based Social Capital ( N = 9,222)

	Depression OLS coefficient			Suicide Logit Coefficient
	Model 1	Model 2	Model 3	Model 1	Model 2	Model 3
Expanded race
Monoracial Black	1.10** (.32)			−0.28* (0.12)
Monoracial Native American	3.21 (1.83)			1.02* (0.44)
Monoracial Asian	1.85* (0.85)			−.31 (0.25)
Monoracial other	−1.35 (1.50)			−0.85 (0.72)
Multiracial other	0.76 (0.85)			0.49 (0.31)
Black-White	0.18 (1.30)			0.03 (0.47)
Native American-White	1.05 (0.73)			0.13 (0.27)
Asian-White	1.84 (1.35)			0.35 (0.54)
Compressed race
Minority		1.20** (0.32)			−0.26* (0.11)
Multiracial		0.92 (0.55)1			0.24 (0.18)1
Mother-specific racial status
Minority			1.20** (0.32)			−0.26* (0.11)
Multiracial: White mother			0.88 (0.68)			−0.02 (.26)
Multiracial: Minority mother			0.98 (0.77)			0.60* (0.27)2
Family measures
Intrafamily social capital
Family relationship	−0.03** (0.00)	−0.03** (0.00)	−0.03** (0.00)	−0.01** (0.00)	−0.01** (0.00)	−0.01** (0.00)
Family understanding	−0.02** (0.00)	−0.02** (0.00)	−0.02** (0.00)	−0.01** (0.00)	−0.01** (0.00)	−0.01** (0.00)
Parent educational expectations	−0.01** (0.00)	−0.01** (0.00)	0.01** (0.00)	−0.00 (0.00)	−0.00 (0.00)	−0.00 (0.00)
Family cohesion	−0.02** (0.00)	−0.02** (0.00)	0.02** (0.00)	−0.00** (0.00)	−0.00** (0.00)	−0.00** (0.00)
Family structure	0.75** (0.00)	0.72** (0.27)	0.72** (0.27)	0.04 (0.10)	0.03 (0.10)	0.04 (0.10)
Extrafamily social capital
Parent community involvement	0.00 (0.00)	0.00 (0.00)	−0.03** (0.00)	0.00 (0.00)	0.00 (0.00)	0.00 (0.00)
Closure of parental networks
Moderate	−0.01 (0.24)	0.01 (0.24)	0.01 (0.24)	0.00 (0.12)	0.01 (0.12)	0.00 (0.12)
Strong	−0.62* (0.27)	−0.61** (0.27)	−0.61* (0.27)	0.01 (0.13)	0.02 (0.13)	0.01 (0.13)
Very strong	−0.68** (0.28)	−0.67** (0.28)	−0.67* (0.28)	−0.08 (0.15)	−0.08 (0.15)	−0.08 (0.15)
Control variables
Male	−1.43	−1.42	−1.42	−0.35** (0.08)	−0.35** (0.08)	−.34** (.08)
Mother’s education (omitted: less than HS)
HS or some college	−0.70	−0.72	−0.72	−0.06 (0.14)	0.06 (0.14)	−.06 (.14)
College or more	−1.38	−1.38	−1.38	−0.08 (0.17)	−0.08 (0.16)	−.08 (.16)
Log of family income	−0.27	−0.26	−0.26	0.03 (0.06)	0.03 (0.06)	.03 (.06)
Results from adjusted Wald test
Model 1 vs. Model 2 (df = 6)		7.19, p < .01			4.06, p < .05
Model 2 vs. Model 3 (df = 6)			4.76, p < .01			3.50, p < .05

Notes: Standard errors are shown in parentheses. Data are adjusted for the complex sampling design of Add Health. Source: National Longitudinal Study of Adolescent Health (Add Health).

^aOmitted category is monoracial White.

¹Coefficient for “Multiracial” differs from that for “Minority” at the p < .05 level.

²Coefficient for “Multiracial: Minority mother” differs from that for monoracial “Minority” at the p < .01 level.

^*Coefficient differs from 0 at the p < .05 level.

^**Coefficient differs from 0 at the p < .01 level.

Table 6.

Multiracial Identification and Social Well-Being Including Measures of Family-Based Social Capital ( N = 9,222)

	School Engagement OLS Coefficient			Social Acceptance Logit Coefficient
	Model 1	Model 2	Model 3	Model 1	Model 2	Model 3
Expanded racea
Monoracial Black	.02 (.03)			.29* (.13)
Monoracial Native American	−.10 (.19)			−.06 (.29)
Monoracial Asian	−.12* (.05)			−.12 (.24)
Monoracial other	.07 (.22)			−.31 (.63)
Multiracial other	.08 (.12)			−.55 (.31)
Black-White	−.00 (.15)			.31 (.47)
Native American-White	−.01 (.11)			−.51 (.33)
Asian-White	−.05 (.15)			1.07 (.60)
Compressed race
Minority		−.00 (.03)			.19 (.12)
Multiracial		.01 (.07)1			−.28 (.21)1
Mother-specific racial status
Minority			−.00 (.03)			.19 (.12)
Multiracial: White mother			−.06 (.08)			−.08 (.25)
Multiracial: Minority mother			.13 (.09)			−.56* (.28)2
Family measures
Intrafamily social capital
Family relationship	−.00** (.00)	−.28** (.04)	−.00** (.00)	.60** (.08)	.01** (.00)	.01** (.00)
Family understanding	−.00** (.00)	−.00** (.04)	−.00** (.00)	.97** (.10)	.01** (.00)	.01** (.00)
Parent educational expectations	−.00** (.00)	−.00** (.00)	−.00** (.00)	.24** (.08)	.00** (.00)	.00** (.00)
Family cohesion	−.00** (.00)	−.00** (.00)	−.00** (.00)	.25** (.10)	.00** (.00)	.00** (.00)
Family structure	.03 (.03)	.04 (.03)	.04 (.03)	−.01 (.10)	.01 (.10)	.01 (.10)
Extrafamily social capital
Parent community involvement	−.00* (.00)	−.01 (.02)	−.00* (.00)	.14 (.09)	.00 (.00)	.00 (.00)
Closure of parental networks
Moderate	−.08** (.03)	−.08** (.03)	.08** (.03)	.13 (.11)	.13 (.11)	.13 (.11)
Strong	−.15** (.03)	−.15** (.03)	−.15** (.03)	.43** (.11)	.45** (.11)	.45** (.11)
Very strong	−.28** (.04)	−.28** (.04)	−.28** (.04)	.17 (.16)	.18 (.16)	.18 (.16)
Control variables
Male	−.01 (.03)	−.01 (.03)	.01 (.03)	.07 (.13)	.27** (.09)	.27** (.09)
Mother’s education (omitted: less than HS)
HS or some college	−.10* (.04)	−.06 (.04)	−.06 (.04)	.07 (.13)	.08 (.13)	.08 (.14)
College or more	−.18** (.05)	−.09 (.07)	−.09 (.05)	−.07 (.15)	−.06 (.15)	−.06 (.15)
Log of family income	−.03 (.06)	−.01 (.02)	−.01 (.02)	.06 (.06)	.06 (.06)	0.06
Results from adjusted Wald test
Model 1 vs. Model 2 (df = 6)		0.02			2.78
Model 2 vs. Model 3 (df = 6)			1.05			2.74, p < .05

Notes: Standard errors are shown in parentheses. Data are adjusted for the complex sampling design of Add Health. Source: National Longitudinal Study of Adolescent Health (Add Health).

^aOmitted category is monoracial White.

¹Coefficient for “Multiracial” differs from that for “Minority” at the p < .05 level.

²Coefficient for “Multiracial: Minority mother” differs from that for monoracial “Minority” at the p < .01 level.

^*Coefficient differs from 0 at the p < .05 level.

^**Coefficient differs from 0 at the p < .01 level.

The regression results suggest that family-based social capital contributes to some of the observed well-being differences between multiracials and monoracial Whites. For example, poorer depression and social acceptance outcomes, which are seen in Tables 3 and 4 for multiracial Native American-Whites, are no longer significant after the introduction of family-based social capital measures in Tables 5 and 6. Also, for the condensed categorization of race, previously negative outcomes associated with multiracials versus monoracial Whites seen in Tables 3 and 4 no longer exist after the introduction of family-based social capital measures seen in Tables 5 and 6. Nevertheless, differences in well-being between monoracial Whites and monoracial minorities do not appear to be explained by family-based social capital. Family-based social capital was also used to test the previous established differences in emotional and social well-being associated with the gender of the minority-status parent. The previously significant results in Tables 3 and 4 remain significant in Tables 5 and 6, however, and the magnitudes only experience a slight decrease, meaning family-based social capital, unlike in the previous two examples, does not explain much of the variation associated with the gender of the minority-status parent.

In sum, the previously significant well-being effects for multiracials versus monoracial Whites disappear after the introduction of measures of family-based social capital. This is true for both the condensed and detailed categorizations of race. Family-based social capital is not useful, however, for explaining well-being variation for multiracials when the mother-specific categorization of race is used. Specifically, the social and emotional well-being effects remain significant for multiracials with a minority mother, even after the introduction of measures of family-based social capital.

Discussion

This study attempts to answer three research questions. First, is there variation in well-being for multiracial adolescents compared to their monoracial counterparts? Second, do multiracial adolescents with a White Mother tend to fare differently than those with a minority mother? Third, can these effects be explained by measures of family-based social capital?

To begin, the results suggest that there are, in fact, significant well-being differences for multiracials compared to their monoracial counterparts. For social and emotional well-being, multiracial Native American-Whites have more negative outcomes than their monoracial White counterparts. These results are not surprising, as previous literature suggests that both suicide and depression are relatively high among the monoracial Native American population when compared to monoracial Whites (Dinges & Duong-Tran, 1992). Additionally, for emotional well-being, the condensed multiracial category shows significant and more negative effects than its monoracial White counterpart. The results also suggest that the gender and race combination of parents is important. Specifically, for social and emotional well-being, there is specific emotional disadvantage for multiracial adolescents who have a minority mother versus those who have a White mother. And finally, family-based social capital is in fact important. The findings show that some previously significant well-being differences are no longer significant with the introduction of measures of family-based social capital.

The results found in this analysis can also be understood in three main themes. First, having significant well-being effects in the detailed categorization of race shows that disaggregating racial categories is useful and important. Additionally, having well-being results that were only significant in the condensed racial categorization shows that studying multiracials as a group is also useful. Specifically, having three separate categorizations of race allowed trends to be compared across the multiracial category as a whole and then specific outcomes for each individual group to be examined. These comparisons provide further support to the claims of Jaret and Reitzes (1999) that individuals from different multiple race heritages experience being multiracial in different ways.

Second, the significance of gender and race of the parent show that being multiracial and having a minority mother appears to be disadvantageous. Multiple studies have pointed to the importance of gender and racial location of parents being important predictors of racial identification for multiracial children (Bratter & King, 2008; Roth, 2005; Xie & Goyette, 1997). This study can extend this finding to assert that parent racial and gender location is also an important predictor of well-being outcomes for multiracial children.

Third, and finally, the disappearance of effects for the detailed and condensed racial categories after the introduction of measures of family-based social capital shows that the family and their relationships with children and the community are important when understanding outcomes, particularly for multiracial adolescents. On the basis of work by Furstenberg and Hughes (1995), we know that minority youth can use higher levels of family-based social capital to diffuse negative life outcomes. Findings from this study extend these results to suggest that multiracials also can use family-based social capital to diffuse negative well-being effects. Additionally, these findings make sense in terms of exosystem models established by Bronfenbrenner (1986) in the description of the ecology of the family. Bronfenbrenner suggests that community and social networks of parents have an indirect effect on psychological outcomes for children by way of their families. Specifically, parents’ workplace, social networks, and community influences can have an effect on the well-being outcomes of children. This theoretical perspective could explain why controlling for family-based social capital makes some of the negative social and emotional well-being outcomes for multiracial children disappear.

Overall, this article suggests that when examining race in our modern society it is important to use multiple measures of racial categorization to understand the growing multiracial population. The analysis supports this claim, particularly through conflicting results within the suicide and social acceptance measures by racial categorization. Whereas grouping multiracials together as a single category does not produce any significant outcomes for social acceptance, separating the group into detailed categories shows significant effects for the Native American-White population. This finding would have been overlooked without a more detailed categorization of race. Just the opposite happens for the well-being measure of suicide; specifically, no values are statistically significant for the detailed classification of race, but when we examine multiracials as a whole, we see they are significantly more likely to consider suicide. In this case, using the condensed classification as a way to understand what is causing this general poorer outcome is important.

Furthermore, this article also illustrates that racial location of the minority parent is an important factor in studying multiracial adolescents. In fact, the analysis shows that multiracials with a White father and minority mother have the worst outcomes compared to their monoracial counterparts. This is particularly interesting as the previous qualitative literature that focused on gender and racial composition systematically suggested more negative outcomes for adolescents with White mothers in multiracial families. This could be happening for a variety of reasons. Differences in parenting or mother-child relationships could be the explanation for this finding. Research on the monoracial population has shown that differences in parenting practices do exist across racial groups, and these parenting practices affect child well-being (Amato & Fowler, 2002). Future research in the area of parenting practices within multiracial families could provide great insight into the intricacies of the effect of gender and race combination of parents in multiracial families.

Finally, we have claimed that when studying multiracial adolescents’ outcomes, it is important to account for family-based social capital. Specifically, for both the detailed and condensed classification of race, after the introduction of measures for family-based social capital, the previously significant well-being effects for multiracials versus monoracial Whites disappear. This is particularly important, as it suggests that the effects were just a product of in-home variation of multiracials that this analysis was able to capture by using this measure of social capital. Specifically, this analysis suggests that poorer outcomes for multiracial adolescents operate through various mechanisms of family-based social capital.

Relating back to the stated hypotheses, the data suggest that the world for multiracial adolescents is not as bleak as proposed by Park (1928). The multiracial adolescent does seem to have more negative outcomes for social and emotional well-being, but these outcomes can be attributed to particular racial groups (in the case of suicide, negative outcomes for the general multiracial category are being driven by Asian-White and Native American-White multiracials), and these outcomes are no longer significant after controlling for family-based social capital.

Contrary to one of the stated hypotheses, it is not multiracial children with White mothers that have the most negative outcomes. Although previous qualitative literature has suggested that a disadvantage exists for White mothers themselves in multiracial families, it does not appear that the same disadvantage exists for their children. In this analysis it was shown that it was multiracial children with minority mothers who have significantly negative outcomes compared to their monoracial White counterparts. Further, this finding was still present after controlling for family-based social capital. This suggests that there is something occurring in the homes of these multiracial children with minority mothers that is beyond the proposed measures of intra-and extrafamily-based social capital, and this experience is translating into disadvantage for these children.

Moreover, in congruence with the stated hypothesis, we have found that various aspects of the family are particularly useful when we attempt to understand the multiracial population. The results show, as supported by the literature, that the family and more specifically strong intra- and extrafamily ties can diffuse the racial disadvantage that would previously be associated with being multiracial.

In closing, the results of this study provide a multifaceted look at well-being outcomes for multiracial adolescents. The findings from this analysis show that the increasing multiracial population is a group that is distinct and important to study. These results specifically advance our knowledge of the multiracial population by showing that social and emotional outcome differences are present and that they can be explained by controlling for intra-and extrafamily-based social capital. Also, this analysis shows that the gender and racial composition of the family is important for understanding multiracial adolescent well-being. Moreover, it shows that something is occurring in multiracial households with a minority mother that is causing negative well-being outcomes for adolescents.

It is important to note, however, that the outcome measures used in this analysis are only crude classifications of adolescent well-being. These measures do not get at the exact nature of how and why an adolescent may feel less socially accepted or more depressed. Future research that is able to use more sophisticated measures to disentangle the processes of negative well-being occurring in these households would be very useful. It should also be mentioned that this study did not take into account the effects of nativity status. Prior research has found multiracial identification by parents as being affected by both gender and nativity status of parents (Qian, 2004). This finding suggests that a future study might also find well-being outcome differences for multiracials on the basis of both gender and nativity status. Additionally, a future study that is able to disaggregate between race and ethnicity and is able to include the Hispanic population would provide great additional insight to the multiracial population and broaden our understanding of that significant and growing ethnic group. Finally, a future study with a larger sample size that could disaggregate the detailed race categorization by race and gender of parents could provide more precise results as to what specific minority mother racial group is causing different outcomes. Prior identity research on the multiracial Asian population has found that race and gender location of parents may not be equally as important for this group (Khanna, 2004; Xie & Goyette, 1997). These studies suggest that future research might also find that race and gender are not the factors contributing to negative well-being outcomes for multiracial Asians.