Introduction

A critical time of self and identity development, adolescence is a period in which children must figure out their place among the social groups and categories that exist in society. Ethnic or cultural background represents an important category within the U.S., particularly for youth from ethnic minority families because their minority status heightens the salience of their ethnic membership. One commonly used conceptualization of ethnic identity, proposed by Phinney (1992), describes the construct as comprised of two dimensions that theoretically stem from social identity and Eriksonian perspectives (Erikson 1968; Tajfel and Turner 1986). Ethnic affirmation or belonging refers to one’s attachment or belonging to one’s ethnic or cultural background, and ethnic identity achievement entails one’s exploration of what it means to be a member of one’s ethnic group (Phinney 1992). Although a number of studies have observed that higher levels of exploration and belonging are associated with better psychological well-being, self-esteem, and academic motivation and achievement (e.g., Fuligni et al. 2005; Rowley et al. 1998; Wong et al. 2003), less is known about how ethnic identity relates to adolescents’ relationships with their families. In particular, there is little information on how ethnic identity may be related to general indicators of family functioning (e.g., cohesion) as compared to more culturally familistic processes (e.g., family obligation). In addition, given that the family is one of most important influences in development (Dmitrieva et al. 2004; Parke 2004), a key question that has yet to be fully addressed is how family processes and ethnic identity evolve over time.

In this article, we delineated different aspects of family relationships and examined whether ethnic exploration and belonging are more strongly linked to family respect, obligation, and daily family assistance, which are key examples of culturally relevant family variables, as compared to dyadic parent–child cohesion and generic levels of family leisure activities. We then examined whether associations between ethnic identity and family processes varied by gender or ethnic background, and whether ethnic identity serves to mediate broad ethnic group differences that are typically found in terms of family obligation and assistance. Finally, we examined longitudinal associations to determine whether family processes predict stronger ethnic identity over time, or whether ethnic identity increases later family connectedness.

Ethnic Identity and Family Processes

The scant research examining ethnic identity and family processes is surprising given that, presumably, the construct of ethnic identity is derived from social relationships and socialization influences in which the family typically is at the forefront (Farver et al. 2002; Hamm 2001; Umaña-Taylor and Fine 2004). Indeed, the family often represents the first and foremost setting for ethnic socialization, that is, the process by which children learn about their ethnic background (Demo and Hughes 1990; Stevenson 1995). Above and beyond ethnic socialization research, much of the existing work specifically linking ethnic identity and the family has been limited in its focus on dyadic family relationships such as parent–child cohesion. Adolescents with a strong and positive sense of ethnic identity tend to report more cohesive relationships with parents (Okagaki and Moore 2000; Wilson and Constantine 1999), perhaps because parental efforts to instill cultural pride or a positive ethnic identity are more effective if the parent–child relationship is close and conducive to such socialization experiences.

Yet, links between ethnic identity and family processes may be more specific than ethnic identity simply being associated with better and more positive relationships. These links may even go beyond the basic ethnic socialization messages that are transmitted through the parent–child relationship (e.g., cultural pride or preparation for bias; Hughes 2003). Given that ethnic identity involves exploration and feelings of belonging toward one’s ethnic or cultural background (Phinney 1992), perhaps ethnic identity is most strongly associated with aspects of the family that are relevant to one’s cultural background, for instance, family obligation or respect. Aspects of social identity theory (Tajfel 1972) also can be used to support associations between ethnic identity and familistic constructs, particularly for youth whose cultural backgrounds do not emphasize respect or obligation as strongly (e.g., European Americans).

Links Between Ethnic Identity and Family: Cultural Perspectives

Ethnic identity development involves establishing what it means to be member of one’s ethnic group, including how this meaning translates into one’s broader value system, cultural orientation, and everyday life (Sue et al. 1998). Members of ethnic groups often share a cultural heritage, traditions, values, and group affiliation (Kwan and Sodowsky 1997). Members of one’s own family are typically members of one’s ethnic group, which suggests that families share a common ethnic identification and familiarity with cultural attitudes and traditions. Hence, although prior work has linked ethnic identity with general indicators of family relationships (e.g., Okagaki and Moore 2000), ethnic identity may be even more strongly linked to family processes that reflect a sense of cultural values or shared ethnic experience.

The obligations of children to respect and assist their family are particularly important cultural values among those with Latin American and Asian backgrounds. Based on principles of familism, considered core features in Latin American cultures, one is expected to place the family before one’s personal needs and to show loyalty and respect for parents and other family members (Zinn 1982). Similar themes are found in the Asian construct of filial piety (Lieber et al. 2004). Akin to family respect, family obligation is a culturally relevant value that refers to the expectation that children support and assist their family in the interest of family maintenance. For instance, children might be expected to pick up groceries, cook meals, or care for other family members (Fuligni et al. 1999; Phinney et al. 2000). Such obligation extends into the future, such as when parents can no longer care for themselves or when young adults can contribute portions of income to the family (Fuligni and Pedersen 2002).

Ethnic identity involves both exploring and committing to the values and cultural orientation of one’s ethnic group (Phinney 1992); hence, Latin American and Asian youth with high levels of ethnic exploration and belonging should display positive feelings toward their ethnic group’s values, traditions, and beliefs. For example, an individual from an Asian American background who is exploring what it means to be Asian, or who feels positively about being Asian American, should endorse some of the values and beliefs associated with the Asian culture (e.g., family respect, obligation). Such internalization of cultural values could transpire as a natural result of identity formation (Sue et al. 1998) or more specifically through cultural socialization processes. That is, in learning more about one’s ethnic background, one could be simultaneously learning more about one’s cultural values (Hughes et al. 2006). The general process of cultural socialization could thus translate into not only a stronger sense of ethnic identity but also a solid commitment to the values and belief systems that are commonly held by one’s group. Given their cultural relevance, links between ethnic identity and familistic values thus may be more closely tied than links between ethnic identity and generic indicators of family life (e.g., dyadic parent–child cohesion).

Links Between Ethnic Identity and Family: Social Identity Perspectives

Although the associations between ethnic exploration and belonging and culturally relevant values of family respect and obligation might be naturally expected from adolescents with Latin American and Asian backgrounds, would associations exist among American adolescents from European backgrounds, for whom family respect and obligation typically are not as salient (Fuligni et al. 1999)? On the one hand, the lower emphasis placed upon family respect and obligation in European American families could mean that these processes are less strongly associated with ethnic identity as compared to adolescents with Latin American and Asian ancestry. On the other hand, basic tenets of social and group identity dynamics suggest that, regardless of ethnicity or other background variables (e.g., gender), conceptual links between ethnic identity and familistic processes should be expected.

Drawing on principles of social identity theory (Tajfel 1972; Tajfel and Turner 1986), ethnic or group identity should relate to within-group variables that reflect group connectedness and affiliation. For example, a sense of group identity, as shown through affiliation, has been linked with a greater willingness to voluntarily provide support and assistance to that group, even if the group does not define itself in terms of such values and traditions (Tyler 1999). Hence, even if such behaviors are not explicitly valued or expected by the group, feeling a sense of group connectedness tends to motivate individuals to act in respectful and helpful ways toward group members. Similarly, social identity theoretically involves a “moral” dimension which incorporates group obligation, commitment, and solidarity (Isajiw 1990, as cited by Kwan and Sodowsky 1997). Based on these ideas, adolescents from European families who report strong levels of ethnic exploration and belonging should be willing to endorse positive attitudes toward their group and to assist and maintain the integrity of their group (e.g., their family), even when constructs of respect or obligation are not outwardly emphasized in their cultural traditions. These patterns of association should also transcend other demographic variables such as gender.

Ethnic Identity as a Mediator of Ethnic Group Differences in Family Processes

Despite theoretical expectations that support similar patterns of association between ethnic identity and family respect and assistance across diverse ethnic groups, differences in the average levels of these constructs can be expected to vary across ethnic background. Numerous studies have found that adolescents from Latin American and Asian backgrounds report higher levels of family respect and obligation, and more acts of assistance compared to those from European backgrounds (Fuligni et al. 1999; Ho 1995; Phinney et al. 2000; Phinney et al. 2005). As one explanation for these trends, perhaps ethnic identification serves as an important mediator of ethnic group differences.

Adolescents from ethnic minority families, including those from Latin American and Asian backgrounds, typically report higher ethnic exploration and belonging as compared to their counterparts from European backgrounds (Phinney 1990). If ethnic exploration and belonging are indeed linked to familistic traditions, then higher levels of ethnic identity among Latin American and Asian adolescents may be one reason why these individuals tend to report higher levels of respect, obligation, and assistance (Fuligni and Flook 2005). Hence, based on prior work that has documented ethnic differences in both ethnic identity and familistic processes, we might expect that higher levels of ethnic identification are one reason why Latin American and Asian youth tend to report higher levels of family obligation and respect than European Americans. Gaines et al. (1997) provided support for this possibility when they found that higher ethnic and cultural identity partially accounted for the tendency for adults from ethnic minority backgrounds to endorse higher collectivism and familism than those from European backgrounds.

In examining ethnic differences in family processes, and mediation by ethnic identity, it is important to consider potential confounds such as generation and socioeconomic status. For instance, recent immigrants may be particularly likely to engage in family obligation behaviors because their foreign-born parents place more emphasis on them, or because their family’s newcomer status creates a very real demand for children to step in and assist (Suárez-Orozco and Suárez-Orozco 2001). Youth from families with low socioeconomic status may have to assist their family out of necessity as, for instance, parents may be working multiple jobs and require additional help at home (Elder and Conger 2000). Although important to consider, it should be noted that demographic variation in income and parent education have not always accounted for ethnic group effects (Fuligni et al. 1999; Fuligni and Pedersen 2002). Nonetheless, we control for generation and socioeconomic status in all multivariate analyses presented here.

Longitudinal Associations Between Ethnic Identity and Family Processes

In order to better understand how ethnic identity and family processes are related, it is important to consider directionality of associations. Drawing on the ethnic socialization and social identity literatures, it appears that directionality could go both ways. Given that ethnic identity develops out of the family and, presumably, through close contact with family members (Demo and Hughes 1990; Hughes 2003), greater family respect and obligation, including daily interactions that reflect such attitudes, could contribute to more opportunities for socialization experiences which then serve to strengthen ethnic identity over time. Alternatively, drawing on Tajfel’s (1972) social identity perspectives, a strong sense of ethnic identification could lead individuals to subsequently develop positive attitudes toward members of their ethnic group, including their family, and to therefore exhibit stronger attitudes toward family respect, obligation, and engage in more activities to help in the maintenance and solidarity of their group.

The Current Study

We examined four key questions concerning ethnic identity and family processes among adolescents from Latin American, Asian, and European backgrounds. First, are ethnic exploration and belonging more strongly associated with family respect, obligation, and daily assistance as compared to more general family processes? We next examined whether associations between ethnic identity and family processes varied by ethnic background or gender. Based on fundamental principles of social identity theory that should apply to all adolescents, we expected that associations would not vary by ethnicity or across males and females. Third, we examined whether ethnic identity mediates ethnic differences in average levels of respect and obligation. We expected that youth from Latin American and Asian backgrounds would report higher levels of familistic processes than European Americans, and that ethnic identity would account for a substantial portion of this effect. Fourth, we explored how ethnic identity and family processes are longitudinally related over time, given competing hypotheses that point to two possible directions of association. In summary, by examining direct, mediational, and longitudinal associations between ethnic exploration and belonging and a variety of family processes, we hoped to gain a deeper understanding of how ethnic identity relates to specific aspects of family relationships during a developmental period when both identity and the family are significant for children’s development and well-being.

Method

Participants

Ninth grade students were recruited from three high schools in the Los Angeles metropolitan area. Although schools with large numbers of Latin American, Asian, and European adolescents were specifically targeted, they collectively represent a sample varying in socioeconomic status and academic achievement, as described by the state (California Department of Education 2006). The first school consisted of predominantly Latin American and Asian American students from families with lower-middle to middle class educational, occupational, and financial backgrounds. The second school consisted of mostly Latin American and European American students from families with lower-middle to middle class backgrounds. The third school included mostly Asian and European American students from middle to upper-middle class families. No single ethnicity dominated any of these schools; rather, the two largest ethnic groups each comprised 30–50% of each school’s total population. All ninth graders in two of the three schools were invited to participate. Due to the unfeasibility of recruiting the large numbers of ninth graders in the third school, a random selection of approximately half of the ninth graders were invited to participate. Of those invited, 65% participated resulting in a total of 783 students with a wide range of ethnic, socioeconomic, and immigrant backgrounds. In all three schools, recruitment and participation took place during the spring semester.

Participants were presented with a large list of ethnic labels (e.g., Mexican, Korean, Italian) and instructed to choose the label(s) that they use to describe themselves. They were asked to choose as many as apply. Based on selections, adolescents were grouped according to whether their chosen ethnic labels represented broad categories of Latin, Asian, or European American ancestry. Adolescents from ethnic minority groups comprising too small a number for meaningful comparisons (e.g., Middle Eastern, African American, Multiethnic) were omitted (approximately 13% of the full sample), resulting in a final N of 679 (50% male, 50% female). Of this final sample, approximately 41% (n = 279) had Latin American backgrounds, 38% (n = 258) had Asian backgrounds, and 21% (n = 142) had European (non-Hispanic) backgrounds. Within the Latin American and Asian ethnic categories, 86% of Latin Americans had Mexican ancestry and 67% of Asian Americans had Chinese ancestry. Other ethnic subgroups within these broader categories included Brazilian, Guatemalan, Honduran, Nicaraguan, Salvadoran, Asian Indian, Filipino, Japanese, Korean, Pacific Islander, Taiwanese, Thai, and Vietnamese. The majority of European American adolescents chose White as their primary label (86%). See Table 1 for a breakdown of our final sample by gender, ethnic background, and school.

Table 1 Sample breakdown by demographic variables
Full size table

Significant ethnic group differences in generational status were found (F (2, 669) = 189.68, p < .001, partial η² = .36). Most of the participants from Latin American and Asian backgrounds were from immigrant families, being predominantly of the first (born outside of the U.S.) and second (U.S. born, but at least one parent was foreign-born) generations. Adolescents from European backgrounds were predominantly of the third generation or greater (youth and both parents born in the U.S.). Significant group differences in parents’ education were also found such that parents of European families exhibited highest levels of education, followed by parents of Asian families, and, lastly, parents of Latin American families (Mothers: F (2, 575) = 25.09, p < .001, partial η² = .08; Fathers: F (2, 575) = 20.83, p < .001, partial η² = .07). All adolescents lived at home and reported similar living arrangements with one exception. Compared to Asian and European youth, a lower proportion of Latin American youth lived with their father. Average age in the ninth grade was 14.87 years (SD = .40), with a range of 13–16.

Approximately 78% of the ninth grade sample participated in the tenth grade (n = 532). Data collection occurred during the spring semester, approximately 12–14 months after ninth grade data collection. Informed consent procedures in the tenth grade were similar to consent procedures used in the ninth grade.

Procedure

Students who provided their own assent to participate and who returned parental consent forms completed self-report questionnaires in small group settings during class time. Study materials were available in English, Spanish, and Chinese. Eight participants completed measures in Spanish (n = 4) and Chinese (n = 4). Questionnaires took about 30 min to complete and assessed demographic information as well as information on adolescents’ ethnic identity, family respect, obligation, and parent–child cohesion.

Upon completion, students were given a 14-day supply of daily diary checklists and told to complete one each night before going to bed. Daily family interactions were assessed via these checklists, each of which took about 5–10 min to complete. Adolescents reported on their family activities each day as well as how much time they spent assisting and spending leisure time with the family. Participants sealed each day’s responses in a manila envelope and stamped the seal with a hand-held electronic time stamper provided by the researchers. The stamper imprinted the current date and time and was programmed to be unalterable. Research assistants entered schools at the end of the two-week period to collect completed checklists. Adolescents received $30 for participation and were told that they would receive two movie passes if inspection of the data indicated that they completed diaries correctly and on-time (e.g., completed on consecutive days with correct date stamped on seal). The time stamper method of monitoring daily diary completion and cash and movie pass incentives resulted in a high rate of compliance. Approximately 95% of the diaries were completed and, of these, 86% were completed on time, on either the correct night or before noon the following day.

All questionnaire and daily diary measures have been successfully used in prior research on adolescents from diverse ethnic backgrounds (e.g., Fuligni et al. 2002).

Measures

Socioeconomic Status

An index of socioeconomic status was created using adolescents’ reports of their parents’ education and occupation. Youth reported on the education level of their mothers and fathers on a six-point scale using the following response options: Elementary/junior high school, Some high school, Graduated from high school, Some college, Graduated from college, Graduated from law, medical, or graduate school. Adolescents also listed their parents’ current job, if any. Responses were coded on a five-point scale using the following codes (sample occupations in parentheses): Unskilled (housecleaning, parking attendant), Semi-skilled (cashier, waitress), Skilled (electrician, bank teller), Semi-professional (nurse, small business owner), Professional (engineer, physician). Socioeconomic status was computed using the mean of the standardized values of both parents’ educational and occupational statuses.

Ethnic Identity

Two subscales of the Multigroup Ethnic Identity Measure (Phinney 1992) assessed ethnic identity. The Affirmation and Belonging subscale, consisting of five items, assesses ethnic pride and feelings of belonging and attachment to one’s ethnic group. Sample items read, “I am happy that I am a member of the group I belong to,” and, “I have a lot of pride in my ethnic group.” The Ethnic Identity Achievement subscale consists of five items and measures exploration of and commitment to one’s ethnic group. Sample items read, “I have spent time trying to find out more about my ethnic group, such as its history, traditions, and customs,” and, “I have a clear sense of my ethnic background and what it means for me.” All items are scored on a five-point scale ranging from Strongly Disagree to Strongly Agree with higher scores reflecting higher Belonging and higher Exploration. Internal consistencies were similarly high across ethnic groups (αs = .68–.78, Belonging; αs = .85–.88, Exploration).

Family Respect

A seven-item measure that has demonstrated good reliability in adolescents from diverse ethnic groups in prior research (Fuligni et al. 1999) was used to assess beliefs about the importance of respecting and following the wishes of the family. Using a five-point scale ranging from Not At All Important to Very Important adolescents responded to items such as, “Treat your parents with great respect,” “Follow your parents’ advice about choosing friends,” and “Respect your older brothers and sisters.” Higher scores indicate higher respect. Internal consistencies across ethnic groups were good (αs = .74–.80).

Family Obligation

Also found to be highly reliable in prior work (Fuligni et al. 1999, 2002), a 12-item scale was used to assess attitudes toward family obligation and the provision of family assistance. Adolescents were asked to determine the importance of each item on a five-point scale ranging from Not At All Important to Very Important. Higher scores reflect higher family obligation. Sample items include, “Help take care of brothers and sisters,” “Run errands that the family needs done,” and, “Help out around the house.” Items concerning attitudes toward future support and obligation were also presented. Samples include, “Help your parents financially in the future,” “Spend time with your parents even after you no longer live with them,” and, “Help take care of your brothers and sisters in the future.” The internal consistencies of current (αs = .73–.84) and future support (αs = .79–.84) across all ethnic groups were good. Both scales were significantly correlated (r = .61, p < .001) and thus combined to reflect an overall index of family obligation.

Parent–Child Cohesion

The Cohesion subscale of the Family Adaptation and Cohesion Evaluation Scales II (Olson et al. 1979) was used to assess parent–child cohesion separately for mother and father. The subscale consists of ten items scored on a five-point scale ranging from Almost Never to Almost Always. Sample items read, “My mother [father] and I are supportive of each other during difficult times,” “My mother [father] and I feel very close to each other,” and, “My mother [father] and I avoid each other at home.” Higher scores reflect higher parent-child cohesion. Internal consistencies obtained across ethnic groups were good (αs = .85–.87, Mother; αs = .86–.88, Father).

Daily Family Assistance

For each of the 14 daily assessments, adolescents indicated whether they engaged in any of the following activities reflecting family assistance or support: Helped to clean your apartment or house, Took care of your brothers or sisters, Ran an errand for your parents or family, Helped your brothers or sisters with their schoolwork, Helped your parents with official business (for example translating letters, completing government forms), Helped to cook a meal for your family, Helped your parents at their work, Anything else to help or assist your family. Events were drawn from prior focus group studies conducted with adolescents and were intentionally designed to correspond with familial obligations (Fuligni et al. 2002). On any given day, the possible number of assistance acts ranged from zero to eight. Daily acts of assistance were aggregated across the 14-day period to reflect an average indicator of daily assistance; that is, on any given day, how many acts of assistance do adolescents engage in? To assess the amount of daily time spent assisting the family, adolescents were asked to estimate the total time spent engaging in these activities, if they answered Yes to any of the activities above. These estimates also were aggregated across the 14-day period to reflect an average index of time spent assisting the family.

Daily Family Leisure

Using daily checklists, adolescents indicated whether they engaged in any of the following: Ate a meal with family, Spent leisure time with family, Spent time with aunts, uncles, cousins, or grandparents. These items also were drawn from prior work with adolescent focus groups (Fuligni et al. 2002). On any given day, the possible number of leisure activities ranged from zero to three. Daily responses were averaged across the fourteen days to reflect an overall indicator of daily leisure interactions. To assess daily leisure time, youth were asked to estimate the total amount of time spent engaging in these activities each day, if they answered Yes to any of the above. Time estimates also were aggregated across the fourteen-day period to reflect an average index of time spent interacting with the family on a leisurely basis.

Results

Rates of Attrition and Missing Data Analyses

Between ninth and tenth grades, adolescents from Latin American and European backgrounds exhibited similar rates of attrition, 27% and 30%, respectively, whereas those from Asian backgrounds exhibited 11% attrition. Within-ethnicity independent samples t-tests determined whether those who participated only in the ninth grade and those who participated in both the ninth and tenth grades differed in ninth grade study variables. For youth from Latin American backgrounds, those who participated only in the ninth grade engaged in more assistance activities with their families (t (267) = 2.87, p < .01) and spent more time assisting their families (t (251) = 3.24, p < .01) compared to those who remained in the study for both waves. Compared to Asian Americans who participated in both waves, those participated only in the ninth grade reported lower levels of ethnic exploration (t (253) = 2.19, p < .05) and lower attitudes toward family obligation (t (256) = 2.79, p < .01). The magnitudes of these differences were moderate (Cohen’s d range = .39–.53). No differences in ninth grade variables were found between European Americans who participated only in the ninth grade and those who participated in both the ninth and tenth grades (ts (132–142) = .05–1.05, ns).

Approximately 0–9% of cases was missing across each of our main study variables. Slightly higher rates of missing data were found in the daily diary components of the study as compared to responses in the initial questionnaire, suggesting that most of our missing data were missing by design (e.g., as a function of fatigue or the relatively intense nature of the daily diaries rather than as a function of a more systematic effect that would call into question our results) (see Acock 2005). To handle missing data, we utilized pairwise deletions in correlation analyses, and listwise deletions in all other analyses.

Hypotheses and Analytic Strategy

We started with preliminary analyses of ethnic and gender differences in key study variables. To examine basic links between ethnic exploration and belonging and multiple indicators of family relationships, we compared the magnitudes of the bivariate correlations between ethnic identity and the different measures of family relationships and activities. We expected that ethnic identity would be more strongly associated with family assistance, obligation, and daily assistance behaviors than with dyadic levels of parent–child closeness and daily family leisure behaviors. Second, we examined whether these bivariate associations differed across adolescents from Latin American, Asian, and European backgrounds, or across males and females, using tests of equal regression slopes. We did not expect groups to differ, based on our theoretical orientation stemming primarily from social identity theory.

Statistical mediation analyses were used to address the third key question of the study: does ethnic exploration and belonging explain ethnic group differences in the tendency for adolescents to value and provide family respect, support, and assistance? In order to properly estimate the role of ethnic identity in explaining these group differences, we conducted multiple regression analyses in which we first estimated ethnic differences in family respect, support, and daily assistance while controlling for generation and socioeconomic status. Our analyses, therefore, examined the extent to which ethnic identity accounted for any observed ethnic differences in family respect, obligation, and assistance above and beyond effects of adolescents’ generational status and their parents’ education and occupational status.

Finally, we conducted longitudinal analyses to address a fourth question: to what extent does ethnic identity drive family processes, or do family processes drive ethnic identity? Although a definitive answer cannot be obtained with our non-experimental data, we conducted cross-lagged regressions across a one-year period of time to determine whether earlier measures of ethnic identity were more strongly predictive of later measures of family respect, obligation, and assistance, or vice versa, after controlling for generation and socioeconomic status.

Preliminary Analyses

Ethnic and gender differences in key study variables were examined using 3 (ethnic background: Latin American, Asian, European) × 2 (gender: boys, girls) analyses of variance (ANOVAs). As shown in Table 2, a number of ethnic group differences were found. With the exception of family leisure activities (F (2, 648) = 1.17, ns), main effects of ethnic background were significant (F range = 7.26–25.95, p < .001). Bonferonni posthoc tests revealed that Latin American and Asian adolescents reported higher levels of ethnic exploration and belonging compared to European Americans. Adolescents from ethnic minority backgrounds also reported greater family respect and obligation and more daily assistance time compared to their European counterparts. Further, Latin American adolescents reported more leisure activities with their families compared to both Asian and European Americans. In terms of parent–child cohesion, adolescents from European and Latin American backgrounds reported greater cohesion with their mother compared to their Asian American counterparts, and adolescents from European backgrounds reported greater cohesion with their father compared to other adolescents.

Table 2 Means (SDs) of study variables across adolescents from Latin, Asian, and European American backgrounds
Full size table

Main effects of gender were found only for parent–child cohesion such that females compared to males reported greater cohesion with their mother (F (1, 658) = 9.14, p < .01). Compared to females, males reported greater cohesion with their father (F (1, 622) = 12.06, p < .001). No interactions between ethnic background and gender were found.

To examine whether generation and socioeconomic status accounted for the ethnic differences found, we ran additional analyses with generation (dummy coded with the third generation serving as the reference) and socioeconomic status (parent education and occupation status as proxies) as covariates. Results from ANCOVAs indicated that generation and socioeconomic status did not account for ethnic group differences found. With the addition of covariates, effect sizes decreased an average of .02 and, with one exception, all previously found ethnic differences remained significant. The one exception was family respect; after controlling for generation and socioeconomic status, ethnic differences were no longer significant.

Bivariate Associations between Ethnic Identity and Family Processes

As shown in Table 3, both subscales of ethnic identity were significantly correlated with family variables, with the exception of leisure time. Specifically, adolescents with a stronger sense of ethnic belonging and exploration reported more cohesive relationships with their mother and father, greater family respect and obligation, and engaged in more daily family interactions. A series of planned statistical comparisons examined whether ethnic identity was more strongly associated with family respect and assistance than with dyadic cohesion and family leisure variables. We utilized Hotelling’s t test, a standard procedure used to determine the statistical difference between two dependent correlations in the same sample (Hotelling 1940, as noted in Chen and Popovich 2002). Results revealed that both ethnic identity subscales were significantly more strongly correlated with family respect and obligation (ts = 4.40–6.70, p < .001) than with parent–child cohesion. In general, both identity subscales also were more strongly linked with daily family assistance than with daily leisure acts, and with daily assistance time than with leisure time (ts = 1.72–2.45, p < .05). One marginal difference was found between belonging and assistance versus leisure time (t = 1.24, p = .11). Since both identity subscales were similarly related to family variables and highly correlated with each other (r = .60, p < .001), they were combined in further analyses as a more parsimonious index of ethnic identity.

Table 3 Correlations among primary study variables and socioeconomic status
Full size table

Ethnic and Gender Differences in Associations between Ethnic Identity and Family Processes

To address our second question, potential moderating effects of ethnic background and gender were examined by conducting tests of equal slopes within analyses of covariance predicting parent–child cohesion, family respect, obligation, and the four indicators of daily family interactions. Ethnic group membership and gender were entered as independent variables and ethnic identity was entered as a covariate. The equality of slopes was tested by estimating the interaction effects between ethnic identity and ethnic background and gender, respectively. No moderating effects were found (F scores of two- and three-way interactions ranged from .01 to 2.20, ns), suggesting that the association between ethnic identity and family processes did not vary across ethnic background or gender.

Ethnic Identity as a Mediator of Ethnic Differences in Family Processes

Following guidelines by Baron and Kenny (1986), we utilized multiple regressions in order to address our third question, namely, whether adolescents’ ethnic identity mediated ethnic group differences in family respect, obligation, and daily family assistance activities. Results from Table 2 confirm that the initial variable, ethnic background, is significantly associated with the mediator, ethnic identity. Results from Table 3 confirm that the mediator is significantly associated with family processes. In order to estimate ethnic group differences in family processes, two dummy-coded variables were created for Latin American and Asian group membership and were entered at Step 1 into a series of regressions. Adolescents’ generational status and socioeconomic status (SES) also were entered at Step 1. The mediating effect of ethnic identity was examined at Step 2. Notably, preliminary regressions further confirm conditions for mediation such that ethnicity is associated with ethnic identity above and beyond the effect of generational status and SES (bs = .25–.45, p < .05). After controlling for generational and SES, ethnic differences remained in terms of family obligation and daily family assistance. Hence, we tested whether ethnic identity significantly mediated ethnic differences in family obligation, daily family assistance activities, and daily time spent assisting the family.

As shown in Table 4, adolescents from Latin American and Asian American backgrounds exhibited higher levels of family obligation than adolescents from European backgrounds. Compared to adolescents from European backgrounds, adolescents from Latin American backgrounds also engaged in more acts of daily assistance and spent more time assisting the family. Again, there were no ethnic differences in family respect above and beyond adolescents’ generation and socioeconomic status. Thus, differences in family respect were not eligible for mediational analyses and are not presented in the table.

Table 4 Regression with demographic variables and ethnic identity predicting family processes
Full size table

Mediational effects of ethnic identity were examined for family obligation and daily family assistance. When ethnic identity is accounted for at Step 2, most of the group differences that were previously significant became nonsignificant. Proportions of ethnic group differences that could be explained by ethnic identity were calculated by dividing indirect effects of ethnicity by total effects. Approximately 35% of the initial ethnic group difference in family obligation that was found between Asian and European youth was accounted for by ethnic identity. In terms of differences between Latin American and European youth, ethnic identity accounted for 61% of the group difference in family obligation, 29% of the group difference in family assistance activities, and 18% of the difference in time spent assisting the family. Sobel tests of the significance of mediating effects confirmed that ethnic identity was a statistically significant mediator of all the initially observed ethnic group differences (Z range = 2.31–3.82, p < .05).

Longitudinal Associations Between Ethnic Identity and Family Processes

We conducted a series of cross-lag regressions in order to address our fourth research question and to shed further light on how ethnic identity and family processes are longitudinally related. Ninth grade ethnic identity was used to predict tenth grade family variables, after controlling for ninth grade family variables. Similarly, each ninth grade family variable was used to predict tenth grade ethnic identity, after controlling for ninth grade ethnic identity. We focused only on family respect, obligation, and daily assistance variables. Generation and socioeconomic status were again included as controls. Standardized regression coefficients depicted in Fig. 1 indicate that ethnic identity in the ninth grade predicted tenth grade family respect and obligation. In contrast, ninth grade attitudes toward family respect and obligation did not predict tenth grade ethnic identity. As shown in Fig. 2, ninth grade ethnic identity did not significantly predict daily assistance in tenth grade; however, ninth grade assistance activities were marginally significant in predicting later ethnic identity.

Fig. 1
figure 1

Cross-lag associations between ethnic identity, family respect, and family obligation (N range = 481–484). Note: Values reflect standardized regression coefficients after controlling for generation and socioeconomic status. Bold lines indicate significant associations. ***p < .001

Full size image
Fig. 2
figure 2

Cross-lag associations between ethnic identity, daily assistance acts, and daily time spent assisting the family (N range = 426–474). Note: Values reflect standardized regression coefficients after controlling for generation and socioeconomic status. Bold lines indicate significant associations. † p < .10, ***p < .001

Full size image

Discussion

Adolescence heralds a time in which children establish who they are and where they fit in terms of the larger society. Ethnic identity is a highly salient social category and has been linked to a number of social and psychological outcomes such as self-esteem and academic motivation (Fuligni et al. 2005; Rowley et al. 1998). Yet, little research has focused explicitly on how ethnic identity and different aspects of family relationships are related. More specifically, although a limited amount of research has linked ethnic identity to general indicators of family functioning (e.g., parent–child cohesion) (Wilson and Constantine 1999), less is known about how ethnic identity may relate to culturally relevant family processes such as obligation or respect. We addressed these unanswered questions in the field and utilized both survey and daily diary methodology in order to gain a better understanding of how ethnic identity may be directly, mediationally, and longitudinally associated with adolescents’ family life.

Our overall results suggest that ethnic identity influences adolescents’ family processes on a number of levels. Adolescents who reported greater exploration of and belonging to their ethnic and cultural background reported greater family connectedness across a variety of indicators, but ethnic identity was most closely tied to aspects of the family that convey a sense of respect and support. Specifically, adolescents’ identification with and connection to their ethnic group showed the strongest associations with family respect and obligation, which are processes that are embedded in the cultural backgrounds of youth with Latin American and Asian origins (e.g., Lieber et al. 2004; Zinn 1982). Similar associations were observed at the level of adolescents’ daily interactions with the family. Compared to simply spending leisure time with the family, ethnic identity was particularly strongly associated with adolescents’ frequency and time spent assisting the family on a daily basis.

Although links between ethnic identity and family support and obligation may be readily expected in cultural groups that emphasize familism or filial piety, what about adolescents from European backgrounds, for whom constructs such as family respect and obligation may not be as culturally relevant? Does ethnic identity reflect a fundamental dynamic that is associated with family respect, obligation, and assistance regardless of ethnic background? Our results indicate that associations between ethnic identity and the family did not vary by gender or ethnic background, and that similar patterns may indeed be expected even among European Americans whose familistic values are typically less salient (Fuligni et al. 1999). Notably, although the vast majority of European American adolescents (86%) chose White as their primary ethnic label, it is possible that some of these youth identified with European heritage cultures that share familistic and collectivistic values (e.g., Italian, Portuguese, Ukranian). Hence, perhaps cultural values play a partial role in explaining associations between ethnic identity and family processes, at least for some European youth.

Nonetheless, above and beyond the values of a particular group, the similarity of the associations across ethnic background and gender provides broad support for a social identity perspective which argues that individuals with a strong sense of group identity tend to act in the best interest of that group (Tyler 1999). Prior theoretical and empirical work has indeed shown that a sense of group connectedness leads to a greater willingness to respect, assist, and feel committed to the members of one’s social group (Tajfel and Turner 1986; Tyler 1999). Our results build on this literature and suggest that, regardless of ethnic background, attitudes and behaviors that are supportive of the solidarity and maintenance of one’s ethnic group can extend to family attitudes regarding respect, obligation, and actual acts of family assistance. Ethnic identification can thus increase one’s social connectedness and provide opportunities for one to have a meaningful and purposeful role in relationships. It is this group connection and feeling like an important and valued member of one’s family or social group that may lead to the positive outcomes that are often associated with ethnic identity (e.g., self-esteem, as shown by Rowley et al. 1998).

Although the actual magnitude of the associations between ethnic identity and family respect and obligation were similar across Latin American, Asian, and European youth, average levels of ethnic identity and family respect, obligation, and support tended to be higher among adolescents from ethnic minority backgrounds. After controlling for generation and socioeconomic status, ethnic group differences in family respect no longer existed. However, ethnic differences remained in family obligation and daily family assistance. Youth with Latin American and Asian backgrounds reported higher obligation compared to their European American counterparts. In addition, Latin American youth reported more daily assistance and spent more time assisting their families compared to European youth. However, once ethnic identity was taken into account, ethnic differences decreased. Ethnic identity accounted for a significant proportion of the overall effect of ethnicity (18–61%), which further speaks to the impact of ethnic identity on the culturally relevant variables of family obligation and assistance. Our findings are in line with prior work using adult samples in which ethnic identity mediated ethnic differences in collectivism and familism (Gaines et al. 1997); yet, we extended this literature by focusing on adolescents and by examining variables that represent more concrete and specific aspects of familism, as well as daily indicators of adolescents’ family life.

One notable pattern found was that group differences in daily assistance behaviors were not as striking between Asians and Europeans as between Latin Americans and Europeans. One explanation is that, for practical reasons, daily manifestations of family assistance and obligation (e.g., assisting with household chores and taking care of siblings while parents work overtime) are related to families’ social and financial resources. Since economic discrepancies are typically greater between Latin American and European families than between Asian and European families, greater differences in family assistance might be expected across Latin American and European families. Although we included an index of socioeconomic status as a control, our proxy was rather distal and may not have directly reflected the economic stresses and demands that families truly face. Indeed, traditional measures of socioeconomic resources that rest primarily on income or education may not provide accurate estimates of the resources and strains that ethnic minority and immigrant families experience (Fuligni and Yoshikawa 2003). In future research, a more precise analysis of social and economic factors could allow for a better understanding of how different indices of economic strain can potentially explain ethnic group differences and cultural processes.

In terms of longitudinal associations, cross-lag analyses revealed that ethnic identity in the ninth grade significantly predicted family respect and obligation in the tenth grade. These findings are again supportive of a social identity framework in that adolescents’ social identification longitudinally predicted commitment and assistance toward members of one’s ethnic group (Tajfel and Turner 1986; Tyler 1999). Interestingly, longitudinal links, albeit modest in size, were significant despite the stability in variables across the ninth and tenth grades. In terms of a reverse directionality that could be reflective of a socialization framework, less support was found. That is, although it is conceptually possible that ethnic identity strengthens over time due to ethnic socialization experiences that arise from family interactions, our data lend only marginal support for this perspective.

Our overall results should be considered in light of limitations. First, our longitudinal data spanned only a one-year period. Future research should examine a longer developmental lag in order to truly disentangle the directionality of effects. Further, the basis for both ethnic identity and family processes could be influenced by extraneous variables, for instance, the socialization influence of peers or the cultural diversity of adolescents’ communities. It may be that, for youth in culturally segregated communities in which their particular ethnic group is in the majority, there are greater opportunities for ethnic socialization from others outside of the family or a greater demand for the enculturation of ethnic values (e.g., familism). Explicit information regarding adolescents’ ethnic socialization history is a similarly important variable to consider (Demo and Hughes 1990; Hughes 2003). Another variable that may be important to consider in future research is collectivism, which may be closely tied to familistic constructs. Perhaps ethnic identity serves to also mediate adolescents’ ethnic group differences in collectivism, as found by Gaines et al. (1997) in their sample of adults. It is also possible that collectivism plays a role in predicting both ethnic identity and culturally relevant family processes in adolescent development.

Potential variation in additional dimensions of ethnic identity also should be examined in future work. We focused on exploration and belonging, which largely reflect social identity and Eriksonian frameworks. Recent research has implicated other dimensions of ethnic identity as important in development, for instance, ethnic centrality or the degree to which one feels that one’s ethnicity is a core aspect of one’s self (Rowley et al. 1998). Perhaps ethnic identity is even more strongly associated with the family if one reports high levels of ethnic centrality. Further, much of the current research on ethnic identity focuses on its positive aspects (e.g., belonging, pride), but it may be informative to consider potentially negative dimensions of the construct. For instance, does having an excessive amount of ethnic pride, perhaps to the exclusion of other ethnic groups, carry negative developmental effects? In a similar vein, our results suggest that a positive sense of ethnic identity relates to more positive connections with one’s family members, as members of one’s ethnic group. Although the reverse also may be true (e.g., that low levels of ethnic identity would relate to a dis-identification with one’s family), more work is needed to thoroughly understand the potential pitfalls of having a negative or low ethnic identity.

In conclusion, results from this study provide evidence that ethnic identity, as conceived through exploration and belonging, is particularly strongly related to family obligation, respect, and daily assistance. Appearing to transcend ethnicity, such associations between ethnic identity and family processes have important implications in light of the potential strengths, resources, and social networks that are available to adolescents. For instance, having a strong sense of ethnic identity may allow for a greater connection to the family (e.g., by predicting family respect and frequency of daily interactions), which can then impart a unique sense of support from which to draw in times of need. The greater sense of family obligation that arises out of ethnic identity could provide a sense of purpose or meaning in life which, in turn, potentially serves as a positive strength. Given the fundamental role that ethnic identity plays in adolescents’ family lives and social connectedness, it would be fruitful for future work to continue to understand the complex ways in which ethnic identity pervades the family context.