Introduction

Self and identity have long been considered multidimensional and dynamic phenomena that can adapt or vary as a function of interpersonal contexts (James 1892). The adaptable nature of ethnic identity has been similarly touted by both early and contemporary scholars (e.g., Du Bois’s 1903 dual consciousness framework), yet the construct continues to be operationalized in general and relatively stable ways (Phinney 2003). The goal of the current study was to utilize a relational approach to determine whether average levels of ethnic identity vary as a function of relational context, and whether the amount of variation in ethnic identity reported across multiple relationships is associated with personal and social adjustment. These issues were examined within young adults from Latin, Asian, Filipino, and European American backgrounds. We focused on the developmental period of young adulthood, during which both identity formation and social relationships are in the forefront of individuals’ lives.

Defining the Construct of Relational Ethnic Identity

Although the field has yet to form a consensus on a single definition of ethnic identity (Sue et al. 1998), many approaches utilize Eriksonian and social identity theories to operationalize the construct. Based on these two approaches, Phinney’s (1992) commonly used measure of ethnic identity focuses on two dimensions, namely, ethnic exploration and ethnic affirmation or belonging. Drawing on Eriksonian (1968) perspectives, ethnic identity involves elements of exploration with the ultimate goal of achieving a fully developed sense of self. Ethnic exploration involves an active search into what it means to be a member of one’s ethnic group, including an examination of one’s values, traditions, and history. The rationale behind ethnic affirmation or belonging stems from social identity approaches, which define group identity as being embedded within the emotional value and significance attributed to one’s group (Tajfel 1981). Ethnic belonging thus reflects an affectively based sense of connectedness with one’s ethnic group. Consistent with these approaches, we defined ethnic identity as a multidimensional construct consisting of ethnic exploration and belonging.

The idea that context can alter individuals’ sense of identity, personality, or even one’s communication styles has been consistently established across diverse areas of research (e.g., Deaux and Ethier 1998; Lawson and Sachdev 2000; Swann et al. 2002). For instance, research on multiple selves has found that individuals present themselves and behave differently across different social contexts (Oyserman and Markus 1993). Similarly, the concept of relational self-worth suggests that individuals’ self-evaluations depend on the specific relationship in which they are interacting (e.g., feeling positively with close friends, but depressed and unhappy with parents) (Harter et al. 1998). Borrowing from these multiple fields of research, we defined relational ethnic identity as individuals’ ethnic exploration and belonging when interacting within different relational contexts. For instance, whereas traditional measures of ethnic belonging assess feelings of ethnic pride and connectedness, in general, relational ethnic belonging refers to the ethnic pride that individuals feel when they are interacting with people from various relationships (e.g., when with parents or peers). Levels of ethnic identity exhibited across relational domains may thus vary depending on the relationship, or alternatively appear more stable regardless of with whom one is interacting.

Relational Variation in Ethnic Identity

Much of the existing research on the dynamic properties of ethnic identity has largely focused on situational variation and contextual salience. For instance, Huang (1998) found that, situationally, Asian American youth felt more Asian at home, more American at school, and equally Asian and American with peers. In terms of salience, Yip and Fuligni (2002) found that adolescents from Chinese backgrounds felt more Chinese when engaging in cultural activities and interacting with ethnically diverse peers. Recent work has indicated that ethnic identity also varies more specifically at a relational level. Adults from Chinese American backgrounds expressed their ethnic identity at highest levels with parents, followed by Asian peers, and, lastly, European American peers (Kiang et al. 2007). Furthermore, average levels of ethnic identity expression and the patterns in which such expression was related to adjustment were most distinct among same-ethnic (parent, Asian peer) and different-ethnic contexts (European peers). Collectively, a burgeoning literature thus supports the idea that individuals feel and express themselves and their ethnic identity differently across different social contexts. Still, a number of meaningful questions remain unanswered.

Relational Variation in Exploration Versus Belonging and Among Specific Relationship Pairs

As found in prior work that has focused on the external expression of ethnic identity or ethnic identity salience (e.g., Kiang et al. 2007; Yip and Fuligni 2002), do specific dimensions of ethnic exploration and belonging vary across relationships? Furthermore, is ethnic exploration more susceptible to relational variation compared to ethnic belonging? Given that a primary component of ethnic exploration, by nature, involves social interactions and learning from others from within one’s ethnic community (Lee 2003), perhaps there are greater differences between the exploration found within same-ethnic and different-ethnic relationships, as opposed to the relational differences found in light of an affectively-based sense of ethnic belonging. What is more, perhaps differences in exploration between same-ethnic (e.g., parents, same-ethnic peers) and different-ethnic relationships are particularly striking given that young adults are more likely to explore their ethnic background with those who share a similar ethnic heritage. Ethnic belonging, which presumably reflects one’s private or internal thoughts and feelings, could thus reflect a more stable dimension of identity, both in comparison to ethnic exploration, as well as in terms of the relative degree of variation found when contrasting same-ethnic and different-ethnic relationships.

Ethnic Differences in Relational Variation

Another unanswered question is whether variation in relational ethnic identity reflects a universal process that can be generalized across individuals from diverse ethnic groups, or whether young adults from specific ethnic groups may be more or less susceptible to exhibiting relational differences. Research has indeed shown that individuals from ethnic minority backgrounds often find themselves negotiating and having to shift between their ethnic and mainstream contexts (Jones and Shorter-Gooden 2003; Saylor and Aries 1999; Umaña-Taylor 2004). Furthermore, the social demand for bicultural competence (LaFromboise et al. 1993) may lead those from ethnic minority backgrounds to feel exceptionally fluid and familiar with contextual adaptation. Hence, individuals in the ethnic minority may be particularly likely to differentiate across their social relationships, such as between same-ethnic and different-ethnic peers or between parents and different-ethnic peers, and perhaps do so to a greater degree than their European American counterparts.

Relational Variation and Adjustment

An additional unanswered question concerns the degree to which relational variation in young adults’ ethnic identity affects their adjustment and well-being, and whether such liabilities differ across ethnicity. In terms of research on the self, one important developmental goal is to integrate one’s multiple selves, which proliferate during adolescence, into a cohesive whole (Harter 1999). Empirical work, predominantly conducted with European Americans, indeed documents the importance of perceived self-consistency across relational domains (Harter et al. 1997). A lack of consistency or self-integration can be detrimental to self-esteem, and contribute to feelings of depression or confusion over one’s “true” self. Similarly, those who report a large degree of variation in their relational ethnic identity may potentially exhibit psychological difficulties and poor adjustment.

However, relational flexibility could be adaptive, or at least normatively found, in youth from ethnic minority backgrounds. Although research suggests that young adults, regardless of ethnicity, struggle to integrate their relational experiences across multiple “worlds” (Phelan et al. 1991; Cooper 1999), it is likely that those from ethnic minority backgrounds face relational demands of cultural adaptation more often than those from the European majority. Moreover, those with Latin American and Asian ancestry tend to be relationally oriented (Gaines et al. 1997; Greenfield et al. 2003) such that feeling different across relationships might represent a phenomenon that is more readily accepted and even expected. As a result, relational variation may not appear as foreign, and may even be considered a positive aspect of development. Indeed, acculturation research often finds that individuals who are biculturally savvy and comfortable in both their ethnic and mainstream environments exhibit favorable adjustment outcomes (LaFromboise et al. 1993). Hence, we argue that, compared to those from European backgrounds, young adults from ethnic minority backgrounds would be less vulnerable to any liabilities that stem from having a differentiated sense of ethnic identity.

Developmental Significance of Young Adulthood

Much of the existing work on ethnic identity and context has focused on adolescents (e.g., Yip and Fuligni 2002). Although developmentally salient during adolescence (Erikson 1968), identity is a continually evolving process and is by no means completely formed by the end of one’s teenage years. In fact, identity may be even more central during the post-secondary years when youth are “emerging” into the real world and have the opportunity to further explore who they are through college and work experiences (Arnett 2007). Research on multiple selves and their liabilities (e.g., Harter et al. 1997) also has been traditionally examined within adolescence. Yet, with new experiences that coincide with young adulthood, individuals may naturally broaden the contexts in which they interact and find themselves increasingly relating to others with different backgrounds from their own. Hence, using a relational framework to examine ethnic identity during the overlooked period of young adulthood provides a notable contribution to the field.

Summary and Hypotheses

The current study seeks to expand our knowledge and understanding of ethnic identity in context by addressing five key questions. We focused on three primary relationships that are generalizable across young adults from Latin, Asian, Filipino, and European American backgrounds, namely, parents, same-ethnic peers, and different-ethnic peers. First, do young adults exhibit significant differences in the average levels of ethnic exploration and ethnic belonging reported across relationships? Based on prior work (Kiang et al. 2007), we expected that the highest levels of relational ethnic identity would be reported with parents, and the lowest levels would be reported with different-ethnic peers. Ethnic identity with same-ethnic peers was expected to fall moderately in between. Second, is the amount of relational variation in ethnic exploration greater than the amount of relational variation in ethnic belonging? Given the contextual nature of exploration (Lee 2003), we expected that young adults would exhibit greater variation in ethnic exploration compared to belonging. Third, are there specific relationship pairs where greater variation is found? Drawing again from prior research (Kiang et al. 2007), we expected that, particularly for explorative aspects of ethnic identity, the amount of variation between same-ethnic (parent, same-ethnic peer) and different-ethnic contexts would be more striking than the variation between parent and same-ethnic peer contexts. Fourth, does the degree of relational variation in ethnic identity vary by ethnicity? Fifth, is relational variation related to adjustment, and do potential liabilities vary by ethnicity? Several outcomes were of interest. Global indicators of well-being included self-esteem and positive and negative affect. In addition, we included a daily assessment of relational competence which measured the extent to which young adults felt like a good person in their daily interpersonal relationships. Given the real-world demands for cultural adaptation that are particularly salient for ethnic minority youth (LaFromboise et al. 1993), as well as cultural differences in relational orientation (Greenfield et al. 2003), we expected that young adults from European American backgrounds would exhibit the least variability in their relational ethnic identity, but that they would be most vulnerable to any negative effects of such variation.

Method

Participants

Participants were drawn from a longitudinal study of youth from Latin, Filipino, Asian, and European American backgrounds in Northern California. Individuals were originally recruited from public schools that consisted of socioeconomically and ethnically diverse students. The community also is diverse in terms of families’ immigration status, educational attainment, and income level. Approximately 80% of the original sample was randomly selected to participate in the most recent wave of data collection, described as a smaller, but more intensive, examination of the transition to adulthood in young adults from ethnically diverse backgrounds. Of the 598 young adults we attempted to contact, 17% were not able to be reached. Of those contacted, approximately 59% agreed to participate. Due to our sampling procedure, ethnic and generational characteristics in the current study resembled those found in prior waves of data.

The final sample used in the current study included 223 young adults who returned all questionnaires and daily diary material (26% Latino, 14% Asian, 32% Filipino, 28% European). The participants’ average age was 25.2 years (SD = 3.9). Males comprised 40% of the sample and females comprised 60%. In terms of generational status, 27% were first generation (foreign born), 41% were second generation (U.S. born of at least one immigrant parent), and 32% were third generation (U.S. born of U.S. born parents). Generational status significantly varied by ethnicity (χ 2 (3, 218) = 45.05, p < .001) such that first generation young adults were mostly from Latin American and Filipino backgrounds, second generation young adults hailed mostly from Latin American, Asian, and Filipino backgrounds, and the majority of third generation young adults were of European descent. Approximately 50% of the sample reported individual incomes of $30,000 or less, 43% reported incomes between $30,000 and $60,000, and 7% reported incomes over $60,000. Although income did not vary by ethnicity, differences in education were found such that Asian American youth were significantly more likely than Latin American and Filipino youth to have obtained at least a Bachelor’s degree (χ 2 (3, 213) = 3.47, p < .05).

Procedure

Randomly selected participants received a letter through the mail describing the study. They also were called to ensure that they received the mailing. Respondents signed and returned a consent form to indicate their interest in participating. Upon receipt of their signed consent, participants were mailed a set of initial questionnaires, a 14-day supply of daily diary checklists, and a small electronic time stamper to help monitor daily diary completion. Upon completion of questionnaires and daily reports, all materials were returned through the mail in prepaid envelopes. Young adults were compensated for participation. For the current paper, relational ethnic identity, self-esteem, and positive and negative affect were derived from initial questionnaires. Daily reports were used to calculate young adults’ daily relational competence, averaged over the 14-day period.

Measures

Relational Ethnic Identity

Based on the Exploration/Achievement and Affirmation and Belonging subscales of the Multigroup Ethnic Identity Measure (Phinney 1992), a relational ethnic identity scale was created to assess ethnic identity within parent, same-ethnic peer, and different-ethnic peer contexts (see Appendix A). The Relational Exploration subscale consists of four items rated on a 5-point scale ranging from Almost Never to Almost Always. This subscale assesses ethnic behaviors and an active exploration into what it means to be a member of one’s ethnic group. The Relational Belonging subscale contains two items assessing ethnic belonging, affirmation, and pride. Participants responded to each of these six items with respect to each of the three relationships in question. Higher scores reflect higher levels of relational ethnic identity. Internal consistencies were calculated separately for each relationship (αs = .81–.86).

Self-esteem

Global self-esteem was assessed using the Rosenberg self-esteem scale (Rosenberg 1986). Ten items are rated on a 5-point scale ranging from Strongly Disagree to Strongly Agree, with higher values indicating higher self-esteem. Sample items include, “I feel that I have a number of good qualities,” and, “I take a positive attitude towards myself.” The internal consistency of this measure was .72 (M = 4.03, SD = .53).

Positive and Negative Affect

Utilizing a scale by Mroczek and Kolarz (1998), young adults were given a list of six positive affective states (e.g., full of life, satisfied) and asked to indicate on a 5-point scale ranging from Not at All to Almost All of the Time how much they experienced that feeling in the past 30 days (α = .89, M = 3.50, SD = .76). Negative affect was similarly assessed with six items tapping negative states (e.g., hopeless, worthless) (α = .75, M = 1.96, SD = .63). Higher scores correspond to higher positive affect and higher negative affect.

Relational Competence

For each of the 14 daily assessments, young adults indicated on a 7-point scale how much they felt like a good daughter/son and a good friend that day. Possible responses range from Not at All to Extremely. Feeling like a good daughter/son and feeling like a good friend were averaged each day, and then aggregated across the 14-day period to create an overall index of relational competence. Higher scores indicate higher relational competence.

Results

Bivariate Associations Among Study Variables

Table 1 illustrates bivariate correlations among study variables by ethnicity. As shown, subscales of relational ethnic identity were significantly correlated with each other regardless of ethnicity. With respect to adjustment, correlations suggest that outcomes were more consistently related to the affective component of ethnic identity reflected through ethnic belonging, compared to ethnic exploration. In addition, there appear to be ethnic differences in the patterns of association. For instance, for young adults from Latin American backgrounds, self-esteem was significantly correlated with relational exploration and belonging, particularly with parents. However, these correlations were not significant for youth from other ethnic groups. Further analyses regarding average levels of and variation in relational ethnic identity, to which we turn to next, should provide more information on how these constructs are connected.

Table 1 Correlations of study variables by ethnicity
Full size table

Mean Levels of Relational Ethnic Identity

As illustrated in Fig. 1, young adults, regardless of ethnicity, reported higher levels of ethnic exploration with parents, followed by same-ethnic peers, and, lastly, different-ethnic peers. Similar patterns were found for ethnic belonging. Paired samples t-tests within each ethnic group revealed that average levels of relational ethnic identity reported across all possible pairs of relationships were statistically different (see Table 2), confirming our first hypothesis.

Fig. 1
figure 1

Mean levels of relational identity: (a) Ethnic exploration, (b) Ethnic belonging

Full size image
Table 2 Means (SD) of relational ethnic identity
Full size table

Relational Variation in Ethnic Identity

Young adults’ actual degree of relational variation was quantified by calculating the absolute difference in ethnic identity reported across each possible pair of relationships. Hence, there were six indices of relational variation: absolute difference in ethnic exploration reported between parents and same-ethnic peers, between same-ethnic and different ethnic peers, and between parents and different-ethnic peers, and absolute difference in ethnic belonging reported between parents and same-ethnic peers, between same-ethnic and different ethnic peers, and between parents and different-ethnic peers. Larger values reflect larger relational variation.

Figure 2 depicts the amount of variation found among these six indicators, delineated by young adults’ ethnicity. Within each relationship pair, paired samples t-tests examined the statistical difference between variation in ethnic exploration versus ethnic belonging (e.g., the difference in ethnic exploration reported between parents and different-ethnic peers was compared to the difference in ethnic belonging reported between parents and different-ethnic peers). Results revealed that, regardless of young adults’ ethnicity, the variation in ethnic exploration was significantly higher than the variation in ethnic belonging when comparing parents with different-ethnic peers (t-range (31–70) = 2.20–5.21, p < .05). In addition, when comparing parents with same-ethnic peers, the variation in exploration was higher than the variation in belonging for Asian Americans (t (31) = 2.18, p < .05), Filipino Americans (t (70) = 3.86, p < .001), and European Americans (t (58) = 2.65, p < .01). Overall patterns thus generally confirm our second hypothesis that greater variation would be found among explorative aspects of identity as compared to a more affectively based component of belonging.

Fig. 2
figure 2

Variation in relational ethnic identity subscales depicted by absolute differences across relationship pairs

Full size image

Further examination of Fig. 2 suggests that the amount of relational variation in both ethnic exploration and belonging appears to be greatest when comparing parents with different-ethnic peers. Paired samples t-tests were used to compare the statistical difference between variation across specific relationship pairs (e.g., the difference in ethnic exploration reported between parents and different-ethnic peers was compared to the difference in ethnic exploration reported between parents and same-ethnic peers). Results revealed that, regardless of young adults’ ethnicity, the variation in exploration reported between parents and different-ethnic peers was greater than the variation in exploration reported between parent and same-ethnic peers and greater than the variation reported between same-ethnic and different-ethnic peers (t-range (31–70) = 2.58–6.26, p < .05). In addition, for Latin American youth, the variation in exploration reported between same- and different-ethnic peers was significantly greater than the variation in exploration reported between parents and same-ethnic peers (t (57) = 2.58, p < .05).

In terms of relational ethnic belonging, European American youth reported greater variation in ethnic belonging reported with parents and different-ethnic peers compared to the variation in belonging reported with parents and same-ethnic peers (t (58) = 3.42, p < .01), and same- and different-ethnic peers (t (58) = 2.80, p < .01). For Latin American youth, variation in belonging reported with parents and different-ethnic peers was significantly greater than the variation in belonging reported with parents and same-ethnic peers (t (57) = 2.64, p < .05). No other differences in the degree of relational variation reported across paired relationships were significant (t-range (31–70) = .25–1.53, ns). Taken together, comparisons of the degree of variation in relational ethnic identity reported across specific pairs of relationships suggest that the most striking differences were consistently found between parents and different-ethnic peers.

Ethnic Differences in Relational Variation

A multivariate analysis of variance with ethnicity as the independent variable and the six indicators of relational variation in exploration and belonging as dependent variables were used to determine whether relational variation differed by ethnicity. Results revealed only one significant univariate difference for the variation in ethnic exploration reported with same- and different-ethnic peers (F (3, 204) = 2.92, p < .05). Bonferroni post hoc tests demonstrated that Latin American youth reported greater variation than their European American counterparts. Given that only one significant contrast (out of a possible 18) emerged, results suggest that young adults’ actual amount of relational variation in ethnic identity is similarly found among diverse ethnic groups.

Relational Variation and Outcomes: Associations by Ethnicity

To determine whether relational variation in ethnic identity was associated with adjustment, we conducted a series of regressions with each outcome (self-esteem, positive affect, negative affect, daily relational competence) regressed on each index of relational variation. The effects of variation in ethnic exploration and ethnic belonging reported between all possible relationship pairs were tested in separate models. All models included gender as a covariate.

Results revealed that the variation in ethnic exploration and ethnic belonging reported between parents and same-ethnic peers was not significantly associated with any of the four adjustment outcomes; hence, are not included in our tables. Significant associations were found between adjustment and variation in ethnic identity reported with same- and different-ethnic peers (see Table 3), and with parents and different-ethnic peers (see Table 4). As expected, greater variation in ethnic belonging reported with same- and different-ethnic peers was related to lower self-esteem, lower positive affect, and lower relational competence, but only for young adults from European American backgrounds. Similarly, for these youth, greater variation in ethnic belonging reported with parent and different-ethnic peers was related to lower positive affect. In terms of ethnic exploration, variation reported between same- and different-ethnic peers was significantly associated with lower positive affect, again, only for those from European American backgrounds.

Table 3 Summary of regression analyses for relational variation (same—different ethnic peers) predicting outcomes
Full size table
Table 4 Summary of regression analyses for relational variation (parent—different ethnic peers) predicting outcomes
Full size table

Discussion

Given that we, as social beings, must interact with a variety of individuals with both similar and different ethnic backgrounds, it is important to understand how ethnic identity operates within the context of our diverse social relationships. Our results extend existing literature that has supported the dynamic nature of ethnic identity. Unique to the current study was our assessment of relational ethnic identity as an internally based construction, as opposed to situational salience or expression as defined in prior work (e.g., Kiang et al. 2007; Yip and Fuligni 2002). We focused on two dimensions of identity that are commonly used in the field, namely, ethnic exploration and ethnic belonging, and examined relational differences at mean levels as well as with regard to the absolute variation or discrepancy reported across pairs of relationships. Additional strengths were our use of an ethnically diverse sample of young adults from an often-overlooked developmental period, our focus on relevant relational contexts including parents, same-ethnic peers, and different-ethnic peers, and the inclusion of state and trait measures of global well-being as well as a relationally oriented measure of relational competence, assessed at the daily level.

We first confirmed that young adults report significantly different levels of ethnic exploration and belonging when interacting with parents, same-ethnic peers, and different-ethnic peers. Specifically, we found that young adults engage in ethnic activities, explore what it means to be a member of their ethnic group, and feel strongest amounts of ethnic pride and connectedness when they are with their parents. They report significantly lower ethnic exploration and belonging with same-ethnic peers, and lowest ethnic identity with different-ethnic peers. Interestingly, such differences in average levels of relational ethnic identity were found regardless of ethnicity, suggesting that ethnic identity is a relationally dynamic construct for young adults from both ethnic minority as well as majority backgrounds.

We quantified young adults’ actual degree of variation by calculating the difference between ethnic identity reported in one relational context versus another. These variation scores were used to determine whether greater relational variation would be found with explorative aspects of ethnic identity as compared to a more internal, affective component of identity. Overall patterns suggest that young adults exhibit significantly greater variation in terms of ethnic exploration. These patterns, which were most striking within the comparisons between parent and different-ethnic peer contexts and between parent and same-ethnic peer contexts, can be expected given that ethnic exploration involves an active process of learning more about one’s ethnicity. An individual who is engaged in exploring his or her ethnicity and culture would likely be doing so in the company of his or her parents since parents logically provide much assistance in this process. Indeed, research in the area of ethnic socialization suggests that parents are a primary source of knowledge about one’s cultural history and background (Hughes et al. 2006). Hence, there may be greater variability in explorative aspects of ethnic identity reported across parent and peer contexts, as opposed to more affiliative aspects of ethnic belonging.

Collectively, these results speak to the idea that ethnic exploration is more contextual in nature than the psychological connectedness that stems from ethnic pride and belonging (Lee 2003). Given that ethnic exploration involves an active learning process whereby adolescents may seek specific individuals and social contexts to further explore who they are, it is reasonable to expect that exploration is particularly relationally driven. Results thus confirm that it is important to consider ethnic identity as a truly multidimensional construct (Phinney 2003); some dimensions of identity might be especially vulnerable to relational influences, while other dimensions might be better characterized as relationally stable, to a certain extent. However, it is important to recall that, although ethnic belonging appears to convey a more relationally stable phenomenon relative to exploration, young adults still exhibited significant differences in average levels of relational ethnic belonging.

Relational variation scores also were used to determine whether differences in ethnic identity reported across specific relationship pairs were more or less distinct (e.g., does the difference in ethnic exploration reported between parents and same-ethnic peers differ from the difference in ethnic exploration reported between parents and different-ethnic peers). We generally found that the greatest variation in ethnic identity, particularly in terms of exploration, was found when comparing parent and different-ethnic peer relationships. One explanation for the relative concordance between relational ethnic identity reported with same-ethnic peers and parents is that the similarly of ethnicity in these relationships may dilute the contrast between them. Furthermore, the difference in ethnic identity reported between same-ethnic and different-ethnic peers also may be less distinct given the equal status and camaraderie that these peer relationships share. Interestingly, as discussed shortly, it is the amount of variation in ethnic identity reported between same-ethnic and different-ethnic peer relationships that appears to have the most consistent associations with outcomes.

In terms of ethnic group differences, the absolute levels of ethnic identity variation reported across relationship pairs were generally similar regardless of ethnic background. Hence, young adults from European American backgrounds exhibited similar levels of relational variation in ethnic exploration and belonging as compared to young adults from ethnic minority backgrounds, suggesting that relational differentiations in ethnic identity reflect a process that is common to diverse youth in this society. Notably, our conceptualization of ethnic identity follows the tradition of some researchers (e.g., Phinney 1992) who provide a widely applicable assessment of ethnic identity by allowing respondents to decide for themselves what their ethnic background is when answering questions about their ethnic identity. Although the current debate over ethnic identity itself representing a similar or dissimilar construct among individuals from European and ethnic minority backgrounds is beyond the scope of this paper, it would be worthwhile in future research to explore this issue further.

Although the tendency to exhibit variation in relational ethnic identity was similar across young adults with different ethnicities, group differences were found in terms of liabilities associated with variation. Interestingly, all of the liabilities found were significant only for young adults from European American backgrounds. For these youth, greater relational variation in ethnic identity was associated with lower self-esteem, lower positive affect, and lower relational competence. It thus appears that, particularly for European American youth, it is important to feel relationally consistent in one’s sense of ethnic identity. These patterns were generally not found for young adults from ethnic minority backgrounds. In fact, if anything, some of the patterns that emerged resembled a trend in the opposite direction. Perhaps due to interdependent values or the reality of having to often navigate across different cultural contexts (e.g., Greenfield et al. 2003), young adults from Latin American, Asian, and Filipino backgrounds did not appear to be significantly or negatively affected by relational variation.

These findings call into question the applicability of prior work on self-consistency and multiple selves to youth from ethnic minority backgrounds. For instance, prior work has pointed to the fundamental need for adolescents and young adults to exhibit a coherent sense of self (see Swann et al. 2003). However, while this developmental goal may be crucial for European Americans, perhaps it is less important for individuals from ethnic minority backgrounds. Perhaps, instead, there is something unique about being relationally oriented or contextually adaptive that is not a liability for ethnic minority youth and that might even serve as a positive resource. Hence, one suggestion for future research is to revisit whether the general developmental goal of integrating one’s multiple selves or forming a sense of self-coherence (as described by Harter 1999; Swann et al. 2003) is truly applicable to youth from more collectivistic and ethnic minority backgrounds, in which cultural and contextual adaptation, and identity flexibility, may be beneficial skills.

Several limitations to the current study should be noted. We did not obtain data regarding the ethnic diversity of young adults’ immediate environment. It is thus possible that some youth reported low levels of ethnic identity with different-ethnic peers because they simply interact with very few different-ethnic peers on a daily basis. Similar patterns due to exposure to same-ethnic peers also may be true. In addition, we delineated our diverse sample by using broad, panethnic labels. Hence, there may be even more variation within these groups than that suggested by our analyses. Likewise, our assessment of relational ethnic identity focused on broad relationships such as “same-ethnic” peers. It is unclear whether, for instance, a Mexican American participant interpreted “same-ethnic” to mean a more general group such as “Latino,” a more specific “Mexican” group, or even a group simply considered “American.” Did European American youth take “same-ethnic peers” to refer to “White” peers or to other “American” peers? Clearly, individual differences in how these broad terms were defined could exist. Future work could build upon our findings by focusing on more specific relationships and by obtaining a diverse and large enough sample to probe whether certain sub-ethnic groups exhibit similarities or differences in the processes shown here.

Developmentally, it would be important to extend our findings and to examine issues of relational ethnic identity across youth from different periods of the lifespan. For instance, we found that young adults from ethnic minority backgrounds did not experience significant liabilities with regard to variation. However, it is possible that feeling comfortable with varying levels of ethnic identity is a critical skill that is acquired earlier in development. That is, for younger adolescents from ethnic minority backgrounds who are only just beginning to establish a sense of self and identity, relational consistency may be more of an issue in terms of adjustment. Similarly, it would be important to understand whether, in later adult years, one’s relational identities eventually become psychologically integrated into a more stable, cohesive whole.

In future research, it also would be worthwhile to examine how relational ethnic identity interacts more specifically with the interpersonal opportunities that are immediately available in one’s environment. For instance, relational ethnic identity may have important implications in terms of individuals’ same- and cross-ethnic friendships and social interactions. As described by Tatum (1997), adolescents’ peer relationships and social support seeking may be highly dependent on their perceptions of how their same- and different-ethnic peers understand and commiserate with their cultural perspectives and experiences. Another fruitful area of research is to better integrate daily diary methodology (e.g., Bolger and Zuckerman 1995) with the construct of relational ethnic identity. That is, does relational ethnic identity determine the types of activities and social interactions that adolescents and young adults engage in on a daily basis? Does relational variation in ethnic identity affect youths’ fluctuating levels of daily well-being?

To further broaden the implications of our findings, it would be interesting to determine why individuals report such varying degrees of fluctuation. Are there individual differences that predict the tendency to be more or less dynamic in one’s view of the self? Recent work has pointed to ethnic socialization or support as one contributor to the relational expression of ethnic identity (Kiang et al. 2007). As an alternative perspective, perhaps the flipside of positive ethnic socialization, namely, perceived discrimination can impede one’s ethnic identity in certain relational contexts. To add to a growing line of work examining the interaction between ethnic identity and discrimination, Rivas-Drake et al. (2008) found that ethnic identity was a protective buffer against the negative effect of discrimination on psychological well-being. Perhaps the effect of discrimination can be further teased apart in consideration of relational ethnic identity. For example, perceived discrimination stemming from European American peers could lead one to develop lower levels of ethnic pride when interacting with European American peers, but not necessarily when interacting with one’s parents. Furthermore, in this example, perhaps a strong sense of ethnic identity, especially with regards to one’s European American peers, would be an important resource to have. Or, perhaps a strong sense of identity reported in another relationship can sufficiently compensate for the negative effects of discrimination that may be perceived elsewhere.

Development is embedded within social relationships, and a relational approach to ethnic identity recognizes the importance of context in self and identity development. Our results add to a growing literature that supports the utility of explicitly examining the dynamic nature of ethnic identity, not only in theory but also in practice. Relational variation in ethnic identity appears to be quite prevalent among emerging adults from a variety of ethnic backgrounds, having significant associations with adjustment. What is more, the actual degree of ethnic identity variation reported across relational contexts was related to adjustment for those from European American backgrounds, suggesting that achieving relational consistency in ethnic identity is a developmentally important task for these youth. Further exploration into the processes and implications of relational ethnic identity could aid in a more practical and real-world understanding of the construct of identity, including how our ethnic exploration and ethnic pride and belonging interact with our everyday interpersonal experiences and with our overall social and psychological well-being.