Impact of Youth Cultural Orientation on Perception of Family Process and Development among Korean Americans

Abstract

Objectives

This study examined how cultural orientations influence youth perception of family processes in Korean American families, and how these family processes in turn predict depressive symptoms and antisocial behaviors among youth. Family processes were examined separately for maternal and paternal variables.

Methods

This study used survey data from Korean American families living in the Midwest (256 youth and their parents) across two time periods, spanned over a year. At the time of the first interview, the average age of youth was 13 (SD=1.00). Using Structural Equation Modeling, this study tested the hypothesized associations concurrently, longitudinally, and accounting for earlier outcomes.

Results and Conclusion

Results show that identity and behavioral enculturation in one’s heritage culture are predictors of bonding with parents, which is notably protective for youth. The results highlight the critical effect of enculturation in enhancing youth perception of the parent-child relationship. Behavioral acculturation to mainstream culture, in contrast, predicts youth problems, although the effect may not necessarily always be via family processes. Similarly, Korean and English language proficiencies predict fewer youth problems, but not always by way of family processes. A few differences emerged across maternal and paternal variables, although there was much commonality in the hypothesized relationships.

Acculturation has emerged as a critical area of research in the last decade (Gupta, Leong, Valentine, & Canada, 2013; Yoon et al., 2013). The field now commonly conceptualizes acculturation as bilinear (i.e., involving both acculturation, learning and adopting the mainstream culture, and enculturation, maintaining one’s heritage culture) and multidimensional (i.e., across multiple components such as language, identity, values, and behaviors) (Yoon et al., 2013). However, with a few exceptions (e.g., Benner & Kim, 2009; Choi, Tan, Yasui, & Pekelnicky, 2014), it is still uncommon to empirically examine acculturation and enculturation simultaneously across multiple domains (Yoon et al., 2013). It is even less common to examine bilinear and multidimensional acculturation in a single model. This lagging practice may contribute to the largely inconclusive findings to date. Furthermore, much less is known about the intervening mediating mechanisms that could specifically point to how acculturation variables influence mental health and behavioral outcomes (Yoon, Hacker, Hewitt, Abrams, & Cleary, 2012). Without such an understanding, policy and intervention efforts are likely to be less effective. This study aims to begin to fill this gap in the literature.

The fastest growing racial-ethnic group in the United States—Asian Americans—remains understudied (Choi, 2008). Contrary to the prevailing “model minority” stereotype, Asian American youth show both positive behaviors, such as good academic performance and behaviors, and negative mental health outcomes, such as high rates of anxiety, depression, and self-harming behaviors (Choi & Lahey, 2006; Hahm, Gonyea, Chiao, & Koritsanszky, 2014; Jang, 2002). Korean American youth, in particular, reflect this mixed pattern of both positive and negative outcomes (J.-S. Lee & Koeske, 2010; Yeh, 2003). Yet, why these outwardly contradictory outcomes exist is an open question. Moreover, Korean Americans, adults and youth alike, are known to be culturally insular (Pew Research Center, 2015; S. J. Lee, 1996; Min, 2010). Nonetheless, Korean American youth, unlike their parents’ generation, are also likely to develop an orientation toward the mainstream culture, especially in language, even if they can preserve their traditional culture (Min, 2006). However, we know little about how Korean youth are responding to acculturation demands and how acculturation variables explain their development. Different dimensions of acculturation, e.g., high ethnic pride but relatively low retention of one’s heritage language among Korean youth (Min, 2006), influence youth outcomes differently (Choi et al., 2014; Yoon et al., 2013). A more sophisticated understanding of the relation between cultural orientations and developmental outcomes and, further, the mediating variables between them, may help explain the mixed pattern of development among Korean American youth. This understanding can, in turn, inform how Asian American subgroups of youth who may share similar circumstances are adapting.

From an array of possible mediating variables between cultural orientations and youth development, this study focuses on family processes, specifically youth perceptions of the parent-child relationship. Family is one of the main developmental contexts for children and also the most culturally determined. Culture shapes family processes, including childrearing goals, parental beliefs and behaviors, and parent-child interactions and relations (Super & Harkness, 1986). However, ethnic minority immigrants living in the United States, including Asian Americans, straddle the line between at least two cultures, the mainstream Western culture and their own heritage. Yet a dearth of studies has investigated how cultural orientations shape the family processes of ethnic minority immigrants. In immigrant families, a typical scenario is such that parents adhere to the culture of origin to rear children, while children resist because they are more acculturated (Choi, He, & Harachi, 2008). The process of developing cultural orientations, especially among developing children, is dynamic and may vary across different dimensions. This study investigates the ways in which each dimension of cultural orientation influences how youth perceive family processes, and subsequently youth outcomes.

Asian American Family Process

Western theories of family process characterize Asian American parenting as more controlling and emotionally distant than the idealized Western authoritative parenting styles (Kagitçibasi, 2007; Vinden, 2001). In authoritative parenting, parents establish firm and clear rules, employ inductive reasoning and expressive warmth, and allow autonomy, active exploration, and risk-taking. It is thought to yield close parent-child relations and positive youth outcomes. Authoritarian parenting, in contrast, is strict, restrictive, coercive, punitive, and emotionally detached and is associated with poor parent-child relations and negative developmental outcomes. When viewed from the Western perspective, Asian parenting is often perceived as more authoritarian than authoritative (Chao & Tseng, 2002).

Recent studies, however, have provided a much more nuanced picture of Asian American parenting. For example, the Chinese American parenting style guan and qin (Chao, 1994; Wu & Chao, 2011) and Korean American parenting ga-jung-kyo-yuk (Choi, Kim, Pekelnicky, & Kim, 2013) blend authoritative and authoritarian styles. For example, Choi and colleagues (2013) found that authoritative and authoritarian styles were positively correlated among Korean Americans, suggesting that the two are not clearly distinctive or negatively related, as is the case in European American families (Deater-Deckard et al., 2011). In other words, parental control, as practiced by Asian parents, is not as coercive, punitive, or dominating as when practiced by Western parents. Although more directive and restrictive, Asian parental control is practiced with reasoning, praise, and warmth (Chao & Tseng, 2002; Kagitçibasi, 2007).

In addition, Asian parents may also express affection differently than Western parents. A traditional Asian parenting virtue is sternness, with fewer overt expressions of love and less use of praise (K. Kim, 2006). Thus, Asian parental affection is not as expressive, physical, or verbal as Western parental warmth and affection, but rather, indirect and expressed through instrumental support (Lisman, Wu, & Chao, 2001; Wu & Chao, 2011). A recent study confirms this virtue among Korean Americans, as more than 90% of Korean immigrant parents in the study reported indirect expression of affection (e.g., working and sacrificing for their child’s education) (Choi, Kim, Pekelnicky, et al., 2013).

Cultural Orientation and Youth Perception of Family Processes

This study examines cultural orientations of youth as two distinct processes (i.e., acculturation and enculturation) in three discrete dimensions of cultural orientation: 1) language competence (English vs. Korean proficiency), 2) identity (as American or Korean), and 3) behavioral cultural participation (activities in mainstream vs. heritage cultures). These dimensions are most commonly used to assess cultural orientation and they are particularly pertinent to Asian American youth (Choi et al., 2014). Although these dimensions are often used interchangeably or in combination, they are distinct and independent aspects of cultural orientation (Ward, 2001). For example, one’s inability to speak a heritage language does not necessarily mean a low level of racial-ethnic identity. Racial-ethnic identity in fact indicates a much more conscious endorsement of race-ethnicity than the other dimensions such as language and behaviors (Tsai, Chentsova-Dutton, & Wong, 2002). Distinctions among dimensions of cultural orientation are particularly salient for Asian American youth because language retention, for example, is not likely linearly correlated with the level of behavioral acculturation or racial-ethnic identity because U.S.–born Asian American youth and those who immigrated as young children largely do not retain their heritage language (Portes & Rumbaut, 2001; Tsai et al., 2002).

The main question of the study is whether Asian youth who are growing up in the United States with different levels of cultural orientation can perceive parenting behaviors as they are intended. Youth appreciation of the culture of their parents will likely influence their perceptions of family processes, particularly in those aspects that may culturally vary from the mainstream culture (Chao & Aque, 2009). In other words, youth who successfully retain their heritage culture (i.e., are highly enculturated) are likely to better understand parental values, intentions, and behaviors that may differ from mainstream culture (Tseng & Fuligni, 2000). Further, youth who understand their parents’ culture, even if they are highly acculturated (i.e., youth who are bicultural), are more likely to avoid confrontations with parents, and thus have less parent-child conflict. Hence, youth appreciation of parental culture can also mitigate the intergenerational cultural gap that is a common source of tension in immigrant families (Choi et al., 2008). On the other hand, youth who mainly endorse mainstream culture are likely to view their parents’ behaviors from a perspective of mainstream culture. As a consequence, parents and their children may argue more, which would strain parent-child relationships (Choi et al., 2008).

This study uses youth reports of family processes, because youth perception of family processes may influence youth development more than their parents’ views would. In addition, youth and parents’ perceptions of parent-child conflict are often not significantly correlated and reveal different pathways of influence on youth development (Choi et al., 2008). For example, while parent-child conflict can weaken parent-child bonding, which in turn increases unfavorable behaviors and psychopathology among children and youth (S. Y. Kim, Chen, Wang, Shen, & Orozco-Lapray, 2013), the impact is most pronounced when youth (but not parents) perceive the conflict (Choi et al., 2008). Thus, it is central to examine how youth perceive their parents’ behaviors and the nature of their relationships with them.

In addition, although understanding youth perception of family processes and its influences is uniquely valuable, few studies have focused on the impact of youth cultural orientation on youth-reported family processes (for an exception, see Dinh & Nguyen, 2006). Empirical findings on the relations between cultural orientations and family processes have largely focused on the parent-child acculturation gap (but not cultural orientations per se) and its impact on relationships, typically parent-child conflict (e.g., Choi et al., 2008; Dinh & Nguyen, 2006; Farver, Narang, & Bhada, 2002; Lau et al., 2005; Ying & Han, 2007) or parenting practices (e.g., S. Y. Kim, Shen, Huang, Wang, & Orozco-Lapray, 2014). Among the few available, (Tseng & Fuligni, 2000) found that, among Chinese American adolescents, youth heritage language use led to improved communication and understanding in parent-child interactions and (Dinh & Nguyen, 2006) found that, among Asian American adolescents as a whole, youth ethnic identity encouraged positive family relations, which supports the idea that enculturation positively impacts family processes. However, for a majority of the youth, English would be their first language, and it would make sense that linguistic acculturation to English would have either no effect or, if any, a positive effect on family processes. For example, Choi and her colleagues (2014) found that language proficiency in both heritage and mainstream culture was related to fewer depressive symptoms among Korean Americans. Although they did not examine the mediating mechanisms, we speculate that while linguistic and identity enculturation had positive effects on family processes, so too did linguistic acculturation. The effect of behavioral acculturation and enculturation is less clear. In the same study, (Choi et al., 2014) found that behavioral acculturation had a negative and concurrent association with mental health, while behavioral enculturation had a negative effect later but not concurrently. This study may help untangle how and why these patterns occurred.

Family Processes and Youth Outcomes

A vast body of parenting literature on the role of parent-child bonding and conflict is unequivocal: family conflict is consistently shown as a risk among adolescents, including Korean American and other Asian American adolescents (Choi et al., 2008; E. Kim & Cain, 2008; Park, Kim, Cheung, & Kim, 2010; Qin, Rak, Rana, & Donnellan, 2012) and close parent-child bonding is related to better development (Choi et al., 2008; Willgerodt, 2008). However, the impact of parental behaviors is less clear. For example, a handful of studies on Asian American families have found mixed, and sometimes paradoxical, effects. For example, strict parental control is positively associated with fewer externalizing problems and better academic outcomes (Ho, Bluestein, & Jenkins, 2008; Kao, 2004; Li, Costanzo, & Putallaz, 2010). At the same time, others have found that a lack of maternal warmth and high maternal control, for example, were predictive of depressive symptoms among Korean American adolescents (E. Kim & Cain, 2008). Authoritarian style was also correlated with depressive symptoms among Vietnamese American youth (Nguyen & Cheung, 2009).

Considered together, it may be conceivable that firm control and less warmth reduces external problems and promotes academic performance. Conversely, the two may be detrimental to youth mental health. It is also plausible that in Asian American families, parental control does not negatively influence parent-child relations and youth outcomes because, unlike white youth, Asian American youth may interpret parental control as a sign of parental care and love (Russell, Crockett, & Chao, 2010). However, if Asian American youth adopt the Western perspective and fail to understand parental culture, they may see their parents’ emphasis on academic excellence as excessive, or view their parents as restrictive, coercive, or emotionally distant (Park et al., 2010; Pyke, 2000), much the same way as the Western parenting theories portray Asian parenting. In such a case, Asian parental control and parental implicit affection may result in negative outcomes.

Present Study and Hypotheses

This study empirically examines whether and how youths’ bilinear and multidimensional cultural orientations relate to their views of parenting behaviors (explicit and implicit affection, firm control) and parent-child relations (bonding and conflict). The study further examines how these family process variables predict two distinct youth outcomes, depressive symptoms and antisocial behaviors (Figure 1). The two outcomes were selected to explain the mixed pattern of youth behaviors that has been found among Asian American youth, that is, high internalizing but low externalizing problems. In addition to the mediating relations, the study examines the direct relations between cultural orientations and youth outcomes given that certain family process variables may not completely mediate the relationships between them.

Figure 1.

Conceptual Model

This study examines whether family process variables influence youth development in a way that is commonly conceptualized even after accounting for the impact of youth cultural orientation. In addition to the conventional measure of parental affection, the study included a measure that was developed specifically to capture Asian parental responsiveness and warmth (Lisman et al., 2001). This measure allows us to determine whether youths’ cultural orientations similarly influence both conventional and culturally-specific family processes. We labeled the conventional measure as explicit affection and the culturally tailored one as implicit affection.

Hypotheses reflect findings that adopting an ethnic orientation (in particular, language and identity enculturation) produces positive family processes and youth outcomes and, on the other hand, acculturation to mainstream, with an exception of language, has no effect or may be detrimental to family processes and youth outcomes (Yoon et al., 2013). We expect the contrasts to be more evident among culturally-specific aspects, such as implicit affection and parental control. We further hypothesize that, after controlling for youth cultural orientations, parental affection and parent-child bonding will reduce depressive symptoms and antisocial behaviors, while parent-child conflict will increase them. However, firm control is hypothesized to reduce antisocial behaviors but increase depressive symptoms. Some relationships of the model are exploratory, specifically the association between behavioral acculturation/enculturation and family processes and youth outcomes.

The data used here measured youth outcomes across two time points over the span of a year. Thus, we examine both concurrent point-in-time relations and longitudinal relations. We also examine longitudinal relations after accounting for earlier youth outcomes because earlier behaviors are one of the strongest predictors of youth behaviors (Moffitt, 1993). The hypotheses concerning longitudinal relationships are exploratory, except that language acquisition is expected to have lasting and positive impacts on youth outcomes (Choi et al., 2014). Finally, the hypothesized models are tested with maternal and paternal variables, both together and separately. Fewer studies focus on fathers than mothers, although there is a growing research interest in paternal parenting among Asian Americans (e.g., Choi, Kim, Kim, & Park, 2013). There is very limited information from which to generate a set of explicit hypotheses regarding parent differences by gender. Thus, the maternal and paternal differences in the hypotheses of this study are exploratory.

Method

Data are from the Korean American Families (KAF) project. The KAF is a survey of Korean American youth and their parents living in a Midwest metropolitan area. The first wave of data collection in 2007 surveyed 291 families (220 youth, 272 mothers, and 164 fathers, N = 656), and a follow-up survey a year later collected data from 247 families (220 youth, 239 mothers, and 146 fathers, N = 605)¹. Korean American and immigrant families with early adolescents (ages 11–14) were eligible to participate. The research team recruited participants from three sampling sources: phonebooks, public school rosters, and Korean church or temple rosters. Approximately equal numbers from each source participated, and further analyses revealed no statistical differences across the three sources in demographics or socioeconomic status (SES). Each parent received $40 and each youth received $20 for participating in the survey. Trained bilingual interviewers recruited and interviewed each family.

Sample Characteristics

In Wave 1, the average age was 12.97 years (SD=1.00) for youth, 43.4 for mothers (SD=4.57), and 46.3 for fathers (SD=4.69). Nearly 64% of mothers and 70% of fathers reported having attained some college either in Korea or the United States. All parents were born in Korea and had lived in the United States for an average of 15.44 years (SD=8.36). Well over one-half (61%) of youth were U.S.–born, and those who emigrated had lived in the United States for an average of 10.44 years (SD=4.14). About one-half (47%) of the families reported an annual household income between $50,000 and $99,999. Nearly one-fourth (23.6%) had incomes between $25,000 and $49,999, and 22% had incomes greater than $100,000. The remaining 7.4% made less than $25,000. Fifteen percent of mothers reported having received public assistance, including food stamps or free/reduced-price school lunch. Nearly all fathers and approximately 40% of mothers reported being currently employed outside the home. Overall, the survey sample was predominantly urban, middle class, Protestant (76.7%), and small business owners (40%), which is fairly comparable to the Korean immigrant profile in the United States (Min, 2006) and in representative data sets such as the National Longitudinal Study of Adolescent to Adult Health (Add Health).

Measures

Unless noted otherwise, response options for all measures were an ordinal Likert scale, ranging from 1 (rarely or not at all) to 5 (very often or strongly).

Youths’ Cultural Orientation

Language enculturation/acculturation (Korean vs. English)

We used two sets of four parallel items (8 total items) from the Language, Identity, and Behaviors (LIB, Birman & Trickett, 2002) scale to measure youth language competency in Korean and English. The questions include “How would you rate your overall ability to speak Korean (or English)?” and “How well do you understand Korean (or English)?” (α=.86 for Korean and α=.91 for English).

Identity enculturation/acculturation (Korean ethnic identity vs. American identity)

Similar to the language scales, we used 14 questions from the LIB (Birman & Trickett, 2002) to ask about the extent to which youth identified themselves as Korean or Americans; for example, “I think of myself as being Korean (or American)” and “I have a strong sense of being Korean (or American)” (α=.88 for Korean identity and α=.91 for American identity).

Behavioral enculturation/acculturation

Adopted from the LIB, we used 18 items to measure youth participation in either Korean or American cultural activities. Topics included peer composition, participation in social clubs or parties, media use, and food; for example, “How often do you usually listen to Korean (or American) songs?” (α=.76 for enculturation and α=.77 for acculturation).

Perception of Family Processes

Unless noted otherwise, we assessed youth perception of parental behaviors separately for mothers and fathers with a parallel set of items. In the combined analyses, we used the sum of the items for single-parent families and the average of summed scores for two-parent families.

Explicit affection

We used 18 items from a short version of the Parental Acceptance and Rejection Scale (PARQ, Rohner, 2004) to assess parents’ caring, attentive, and comforting behaviors (nine items each for maternal and paternal behaviors). Examples include: saying nice things about me, being really interested in what I do, and paying a lot of attention to me. (α=.91 for maternal explicit affection, α=.90 for paternal explicit affection, α=.94 for combined).

Implicit affection

To assess parent-child relations specific to Asian American families, we used 4 items developed by Lisman, Wu, and Chao (2001). The items ask how youth perceive parental sacrifices and indirect ways of expressing affection; for example, “My mom (or dad) will put my needs before her/his own needs,” and “My mom (or dad) does not often say it, but does things that show me she/he loves me” (α=.45 for maternal, α=.75 for paternal, α=.65 for combined).

Firm control

We used eight items from the Children’s Report of Parent Behaviors (CRPB, Schludermann & Schludermann, 1988) to assess controlling behaviors of mothers and fathers; for example, “My mom (or dad) believes in having a lot of rules and sticking with them,” and “My mom (or dad) gives hard punishments when I misbehave” (α=.78 for maternal, α=.79 for paternal, α=.82 for combined).

Bonding with parents

We used 12 items to assess youth bonding with parents (Armsden & Greenberg, 1987; Hawkins & Catalano, 1990). Questions include, “How close do you feel to your mom (or dad)?” and “How much do you think she (or he) cares about you?” and “How often do you share your thoughts and feelings with her (or him)?” (α=.84 for maternal, α=.92 for paternal, α=.91 for combined).

Mother-child conflict

Adopted from the Conflict Behavior Questionnaire (CBQ, Prinz, Foster, Kent, & O'Leary, 1979), we used four items to assess youth perception of mother-child conflict. The questions include, “My mom and I get angry at each other a lot” and “My mom never listens to my side of the story” (α=.80). Items measuring father-child conflict were not available in the survey.

Youths Outcomes

Depressive symptoms

Fourteen items from the Children’s Depression Inventory (Angold, Costello, Messer, & Pickles, 1995) and the Seattle Personality Questionnaire for Children (Kusche, Greenberg, & Beilke, 1988) assess the frequency of depressive symptoms during the two weeks prior to the survey. For example, “I found it hard to think properly or concentrate,” “I felt miserable or unhappy,” and “I feel like crying a lot of the time” (α=.91 in both waves).

Antisocial behaviors

We adopted 34 items from the DSM criteria for conduct disorder as well as several measures of antisocial behaviors frequently used in research, such as the Seattle Social Development Project (Hawkins & Catalano, 1990) and the Seattle Personality Questionnaires (Kusche, Greenberg, & Beilke, 1988), including measures of delinquent and aggressive behaviors. As only a few youth reported committing the listed behaviors frequently, we dichotomized each item as either 0 for no incidence or 1 for any incidence of the behavior, and we summed answers for the items (from 0 to 34) (α=.83 in Wave 1 and α=.90 in Wave 2). Because a high rate of the respondents reported having no antisocial behaviors, we further dichotomized the scale into binary variable, 0 for no and 1 for any antisocial behaviors.

Demographic Controls

We included age and gender of the child and a parent report of family SES and single-parenthood as controls. For example, older age, male gender, lower SES and single-parenthood predict a higher rate of antisocial behaviors and female gender predicts a higher rate of depressive symptoms. Family SES was coded 1 (Yes) if they had received public assistance and 0 (No) if they had not. Although a few items accessed parental SES in the survey, receiving public assistance stood out as the most viable SES variable. Other variables lacked enough variance (e.g., education) or they did not significantly relate to other SES variables or youth outcomes as expected. In fact, standard parental SES measures often do not predict youth outcomes among minority groups in the same way they do for whites (Gavin et al., 2010) and this study confirmed this pattern. Single-parent family was coded 1 (Yes) and 0 (No) and the gender variable was coded 1 (boys) and 0 (girls). In addition, we added youth’s place of birth and the number of years living in U.S. as controls, which enables the examination of the effect of cultural orientations above and beyond the demographic acculturation variables.

Analysis Plan

We constructed the data around youth participants (n=256), with youth data matched to parents’ data. Before testing the conceptual model, we conducted several univariate and bivariate descriptive analyses, including means and standard deviations at the individual item and construct levels, item-total correlations, and reliability of each construct and pair-wise correlations among main study constructs. We examined descriptive statistics by gender as well as with a combined group of boys and girls. We also examined mothers and fathers, across both waves whenever available.

We conducted a structural equation modeling to test relations in the hypothesized model. This approach enables a simultaneous investigation of direct and indirect effects. In addition, to test both current and predictive relationships between predictors and mediators (i.e., youth cultural orientations and family process variables) and youth outcomes, we used depressive symptoms and antisocial behaviors from both Waves 1 and 2. We first ran a model with independent and mediating variables predicting youth outcomes at Wave 1. We used independent and mediating variables from Wave 1 to predict youth outcomes at Wave 2; this allowed us to examine predictive relationships over one year. To examine whether youth cultural orientation and family process variables predict youth outcomes above and beyond the effect of the prior behaviors, we added youth outcomes at Wave 1 as one of the independent variables in the longitudinal model. We used M Plus 7.4 (Muthén & Muthén, 2013) to examine the path models.

We assessed the fit of all models by examining model chi-square (χ²), the Comparative Fit Indices (CFI, Bentler, 1990), and the Root Mean Squared Error Approximation (RMSEA, Browne & Cudeck, 1993). A good fit is indicated by CFI values of greater than .90 (Byrne, 1994). Values of less than .05 for the RMSEA are considered a good fit, values between .05 and .08 indicate a fair fit, and values greater than .10 represent a poor fit (Browne & Cudeck, 1993). We examined the statistical significance of the estimated parameters with z statistics and a .05 level of statistical significance. We also used modification indices function (MI) to investigate whether data suggest dropping any particular path in the model. We used maximum likelihood for estimating the models and handling missing data because the outcome variables include both continuous (depressive symptoms) and binary (antisocial behaviors).

Results

Descriptive Statistics

Table 1 summarizes the descriptive statistics, sorted by parent gender. Notable findings are that youth reported significantly higher Korean identity than American identity, but endorsed behavioral acculturation and enculturation about equally. Overall, youth reported higher rates of positive family processes, e.g., explicit affection, implicit affection, and bonding to parents, than negative ones, e.g., conflict and firm control. Bivariate correlations showed that enculturation and acculturation variables were moderately and inversely correlated with one another, while enculturation was in general positively correlated with positive family processes.

Table 1.

Mean, standard deviations and correlations of the study constructs

	Constructs	1	2	3	4	5	6	7	8	9	10	11	12	13	14	15
1	Language Enculturation	--	0.17*	0.47**	−0.34**	−0.37**	−0.36**	0.15*	−0.07	−0.15**	0.12	−0.07	−0.17*	−0.10	−0.05	−0.06
2	Identity Enculturation	0.17*	--	0.38**	0.11	0.00	0.00	0.33**	0.25**	0.02	0.34**	−0.15*	−0.18*	0.01	−0.15*	−0.17*
3	Behavioral Enculturation	0.47**	0.38**	--	−0.16*	−0.19**	−0.23**	0.24**	0.10	−0.01	0.23**	−0.10	−0.12	0.06	−0.08	−0.01
4	Language Acculturation	−0.34**	0.11	−0.16*	--	0.44**	0.52**	0.19**	0.22**	0.06	0.16**	0.08	−0.08	−0.07	−0.20**	−0.14
5	Identity Acculturation	−0.37**	0.00	−0.19**	0.44**	--	0.57**	0.12	0.10	0.03	0.02	0.05	0.08	−0.01	−0.10	−0.00
6	Behavioral Acculturation	−0.36**	0.00	−0.23**	0.52**	0.57**	--	0.11	0.12	0.18*	0.08	0.11	0.17**	−0.01	0.08	0.11
7	Explicit Affection	0.15*	0.38**	0.28**	0.09	0.15*	0.04	--	0.62**	−0.06	0.77**	--	−0.40**	−0.12	−0.25**	−0.09
8	Implicit Affection	−0.01	0.18**	0.11	0.10	0.13	0.08	0.46**	--	0.15*	0.63**	--	−0.19**	0.04	−0.15*	0.02
9	Firm Control	−0.12	0.10	−0.02	0.15*	0.07	0.14*	0.04	0.17*	--	0.05	--	0.13	0.08	0.07	−0.03
10	Bonding to Mom or Dad	0.12	0.39**	0.29**	0.05	0.09	0.04	0.77**	0.38**	0.08	--	--	−0.39**	−0.12	−0.22**	−0.04
11	Mother-Child Conflict	−0.07	−0.15*	−0.10	0.08	0.05	0.11	−0.45**	−0.23**	0.29**	-0.51**	--	0.37**	0.13	0.28**	0.15*
12	Depressive Symptoms (W1)	−0.17*	−0.18**	−0.12	−0.08	0.08	0.17*	−0.47**	−0.23**	0.02	−0.50**	0.37**	--	0.37**	0.36**	0.13
13	Depressive Symptoms (W2)	−0.10	0.01	0.06	−0.07	−0.01	−0.01	−0.04	−0.01	−0.00	−0.10	0.13	0.37**	--	0.12	0.32**
14	Antisocial Behaviors (W1)	−0.05	−0.15*	−0.08	−0.20**	−0.10	0.08	−0.33**	−0.15*	−0.02	−0.35**	0.28**	0.36**	0.12	--	0.61**
15	Antisocial Behaviors (W2)	−0.06	−0.17*	−0.01	−0.14	−0.00	0.11	−0.15	−0.09	−0.08	−0.19*	0.15*	0.13	0.32**	0.61**	--
	Mean (Maternal)	3.12	4.37	3.51	4.47	3.35	3.87	4.03	4.11	3.19	4.21	2.29	1.55	1.72	1.12	1.11
	(Paternal)							3.91	4.08	3.30	3.93
	Standard Deviations (Maternal)	0.89	0.61	0.65	0.69	0.97	0.65	0.70	0.78	0.75	0.61	0.84	0.56	0.63	0.14	0.18
	(Paternal)							0.86	0.93	0.89	0.88

Note: The results with maternal variables are below the diagonal and those with paternal variables are above the diagonal.

^*p < .05;

^**p < .01;

^***p < .001

Path Models

We tested the model with (1) youth and parent variables (i.e., maternal and paternal variables combined), (2) youth and maternal variables, and (3) youth and paternal variables.² Mother-child conflict was not included in the model with paternal variables. Thus, there were six predictors of acculturation/enculturation and five mediators (four for the paternal family process model). The model was fully saturated to examine both mediating and direct relations. Figure 2 summarizes the results of the point-in-time path models for mothers and Figure 3 for fathers. We focus reporting on the respective models for mothers and fathers, because the model with the mothers and fathers combined was an addition of significant paths from the two separate models. The control variables were adjusted in the model but not shown in the figures for simplicity. We describe notable results from the longitudinal models and the longitudinal models that accounted for outcomes during the prior year, though the results are not reported in the figures, because they were largely similar across the models, except that the model fits were the best in the initial point-in-time models.

Figure 2.

Final point-in-time path model for mothers

Fit indices: χ²(30) = 48.980, p = 0.0158, CFI = 0.964, RMSEA = 0.050

Path coefficients are standardized, *p < .05; ** p< .01; ***p < .001

Several paths are not shown for simplicity, e.g. six controls for the outcome variables (age, gender, family SES, single parenthood, youth’s place of birth, and the number of years living in U.S.) and covariations among six independent variables on the far left hand side and among mediating variables.

Figure 3.

Final point-in-time path model for fathers

Fit indices: χ²(24) = 67.163, p = 0.0000, CFI = 0.915, RMSEA = 0.084

Path coefficients are standardized, *p < .05; ** p< .01; ***p < .001

Several paths are not shown for simplicity, e.g. six controls for the outcome variables (age, gender, family SES, single parenthood, youth’s place of birth, and the number of years living in U.S.) and covariations among six independent variables on the far left hand side and among mediating variables.

Maternal Variable Models

First, in the models with maternal variables, fit indices were χ²(30) = 48.980, p = 0.0158, CFI = 0.964, RMSEA = 0.050 (R²=39% for depressive symptoms, R²=24% for antisocial behaviors) for the initial point-in-time model. For the longitudinal model, they were: χ²(30) = 49.156, p = 0.0152, CFI = 0.955, RMSEA = 0.050 (R²=11% for depressive symptoms, R²=10% for antisocial behaviors). After accounting for prior outcomes, fits were: χ²(42) = 133.793, p = 0.0000, CFI = 0.825, RMSEA = 0.092 (R²=25% for depressive symptoms, R²=10% for antisocial behaviors).

In the point-in-time model with maternal variables, identity enculturation was positively associated with explicit affection (and implicit affection at p =.052) and bonding. Behavioral enculturation was also positively related to explicit affection and bonding. These findings are largely in line with the study hypotheses. However, contrary to the expectation, identity acculturation was positively related to explicit affection. In regard to the relations between maternal family processes and youth outcomes, explicit affection was related to fewer depressive symptoms, bonding was associated with fewer depressive symptoms as well as fewer antisocial behaviors, and mother-child conflict was associated with more antisocial behaviors. Implicit affection and firm control did not predict either youth outcome. In addition, after accounting for the mediation by maternal family processes, language acculturation was associated with fewer depressive symptoms, on the other hand, behavioral acculturation was related to more depressive symptoms and more antisocial behaviors.

Paternal Variable Models

In the models with youth and paternal family process variables, fit indices were χ²(24) = 67.163, p = 0.0000, CFI = 0.915, RMSEA = 0.084 (R²=31% for depressive symptoms, R²=26% for antisocial behaviors) for the initial point-in-time model. For the longitudinal model, they were: χ²(24) = 67.385, p = 0.0000, CFI = 0.899, RMSEA = 0.084 (R²=11% for depressive symptoms, R²=8% for antisocial behaviors). After accounting for prior outcomes, fits were: χ²(34) = 121.860, p = 0.0000, CFI = 0.824, RMSEA = 0.100 (R²=24% for depressive symptoms, R²=8% for antisocial behaviors).

In the point-in-time model with paternal variables, language enculturation was positively associated with explicit affection and negatively associated with firm control. Identity enculturation was positively related to paternal affection (both explicit and implicit affection) and bonding. Behavioral acculturation was positively related to firm control. Similar to the maternal family process model, explicit paternal affection was associated with fewer depressive symptoms and bonding was associated with fewer depressive symptoms and fewer antisocial behaviors. Paternal implicit affection and firm control did not predict either youth outcome. Also largely similar to the maternal model, after accounting for paternal family processes, direct paths emerged mainly between mainstream orientation variables and youth outcomes, i.e., language acculturation was related to fewer depressive symptoms and behavioral acculturation was associated with more depressive symptoms and more antisocial behaviors. Unique to the paternal model, identity acculturation was related to fewer antisocial behaviors.

Point-in-time and Longitudinal Models

The point-in-time models had the best fit among the models. Because the predictors (i.e., cultural orientation variables) and mediating variables (i.e., family process variables) were used only from Wave 1 data, the relationships among them did not change across the models. The changes were mainly due to youth outcomes in Wave 2. Notable findings in the maternal model were that language enculturation predicted fewer depressive symptoms longitudinally (β=−.216, p <.05), and remained significant after accounting for prior depressive symptoms (β=−.184, p=.053). Language enculturation also predicted fewer antisocial behaviors longitudinally (β=−.219, p<.05), and after accounting for prior antisocial behaviors (β=−.218, p<.05). Language enculturation did not predict either outcome in the point-in-time maternal model, although it is significantly and negatively correlated with depressive symptoms in Wave 1. In the paternal longitudinal models, language enculturation predicted fewer antisocial behaviors (β=−.213, p <.05), both with and without accounting for earlier youth outcomes. The relationships between language acculturation, behavioral acculturation, and youth outcomes were not significant in any of the longitudinal models, i.e. the maternal, paternal or combined models. Finally, none of the family process variables remained significant in the longitudinal models.

Parent Models (Maternal & Paternal Variables Combined)

In the combined model (not shown in figure), fit indices were χ²(30) = 53.662, p = 0.0050, CFI = 0.959, RMSEA = 0.056 (R²=39% for depressive symptoms, R²=25% for antisocial behaviors) for the initial point-in-time model (with Wave 1 outcomes). For the longitudinal model (Wave 2 outcomes), the fits were χ²(30) = 53.530, p = 0.0052, CFI = 0.950, RMSEA = 0.055 (R²=11% for depressive symptoms, R²=9% for antisocial behaviors). For the longitudinal model accounting for the prior youth outcomes (Wave 2 outcomes accounting for Wave 1 outcomes) the fits were χ²(42) = 139.885, p = 0.0000, CFI = 0.828, RMSEA = 0.095 (R²=25% for depressive symptoms, R²=10% for antisocial behaviors). In this parent model, it was noted that the relationship between identity enculturation and implicit affection was positive and significant (β=.194, p<.01).

Indirect Effects

We further examined indirect effects of cultural orientations via family process variables. We found that, in the point-in-time models, identity enculturation was related to fewer depressive symptoms via maternal explicit affection (β=−0.057, p<.05), bonding to mom (β=−0.098, p<.05), paternal explicit affection (β=−0.069, p<.05), and bonding to dad (β=−0.081, p<.05). Identify enculturation was related to fewer antisocial behaviors via bonding to mom (β=−0.102, p<.05) and bonding to dad (β=−0.119, p<.01). Behavioral enculturation was related to fewer depressive symptoms via bonding to mom (β=−0.058, p<.05) and to fewer antisocial behaviors via bonding to mom (β=−0.060, p<.05) and bonding to dad (β=−0.066, p<.05). There was no statistically significant indirect effect between mainstream orientation and youth outcomes. In addition, no significant indirect effect was found in the longitudinal models.

Sensitivity Analyses

Although the family process mediators in the models were specified to correlate with one another to account for correlations among them, we conducted additional sensitivity analyses with one mediator at a time to see whether the statistically significant paths hold. When considered one mediator at a time, one or two additional paths emerged as significant in each model, but all of the significant paths between family process variables and youth outcomes in the multivariate full model remained significant. In fact, the size of significant path coefficients in the full model became larger when one mediator was modeled. For example, in a single mediator model, maternal explicit affection was negatively associated with antisocial behaviors (β=−0.251, p<.001) and its relation with depressive symptoms became larger, i.e., β=−0.488, p<.001. Similarly, maternal implicit affection (β=−0.202, p<.01) and paternal implicit affection (β=−0.155, p<.05) were significantly related to less depressive symptoms. Mother-child conflict was also significantly related to more depressive symptoms (β=0.345, p<.001), in addition to its positive relation to antisocial behaviors. Also, when explicit affection was not included in the model (except when bonding was a mediator), language enculturation was directly and significantly related to fewer depressive symptoms.

Discussion

In recent years, numerous scholars have advocated for immigrant groups to retain their heritage culture to better facilitate parent-child relations (e.g., Booth, Crouter, & Landale, 1997) and strengthen a child’s sense of self (e.g., Downey, Eccles, & Chatman, 2005). Many immigrant parents have also feared the loss of traditional values and behaviors if their children “Americanize” too rapidly or too fully. Several empirical studies find a positive role for retaining one’s heritage culture, though the research often lacks precision, leaving the question unanswered about which aspects of culture youth should be encouraged to retain or discard. Acculturation is unavoidable among youth and, in fact, may be essential for children to succeed in American society. Further, acculturation and enculturation occur concurrently, thus it is imperative to examine them simultaneously rather than studying either in isolation. Examining a bilinear and multidimensional acculturation in a single model, the findings of this study extend existing knowledge by enhancing specificity in the roles of cultural orientations, and further delineating the mediating pathways of family process by parent gender.

Cultural orientations and family processes

One of the main findings in the current study is that, as hypothesized, ethnic cultural orientations among youth (i.e., language, identity, and behavioral enculturation) may enhance positive perceptions of family processes. Specifically, when Korean American youth report being able to speak Korean, they perceive their fathers as more loving and less restrictive. This result supports a previous finding that youth heritage language is beneficial in parent-child communication (Tseng & Fuligni, 2000) and expands that it is especially useful in father-child relation. In addition, when youth have a stronger sense of Korean identity, they perceive their parents as loving and responsive and bond more strongly with them. Youth involvement in Korean culture (e.g., having co-ethnic friends, consuming Korean media, and eating Korean foods) is also associated with enhanced mother-child bonding and increased perception of explicit affection from mothers. It is noteworthy that Korean ethnic identity seems to increase youths’ perception of parental affection, not only as measured by conventional items, but also as measured by the items that were specifically tailored for Asian Americans. Using the Asian-American-specific measures, Korean ethnic identity was especially linked to an increase in youths’ perception of paternal affection. Youth endorsement of Korean ethnic identity helps youth perceive traditional parenting behaviors as parents intend them, for example, such youth recognize that fathers show their love through their actions rather than verbally expressing it. The findings of this study augment empirical evidence for the positive role of ethnic identity in family relations (e.g., Dinh & Nguyen, 2006) and, more important, provide a much detailed picture of family processes that are facilitated by youth enculturation.

Conversely, the impact of mainstream cultural orientation on family processes is complex. Language and identity acculturation are not associated with paternal family processes. Language acculturation does not influence maternal family processes as expected, either, but identity acculturation is a positive factor, increasing youth perception of mother’s explicit affection. These findings provide mixed support for the study hypotheses. The positive association between identity acculturation and maternal explicit affection may be explained by the interactive nature of parent-child relationships. Acculturating youth may be more sensitive to their mother’s demonstration of explicit affection and, to accommodate their children’s acculturation, mothers may adopt Western ways of affection (e.g., verbal and physical affection), being reinforced by the interactions with their children (Choi & Kim, 2010). It remains curious why this may be the case only for identity acculturation, although identity is a more conscious endorsement of a culture than the other dimensions such as language and behaviors (Tsai et al., 2002) and may be more noticeable to mothers. Only behavioral acculturation plays a negative role in family processes and only in relation to fathers. When Korean American youth are more involved in the mainstream culture (i.e., behavioral acculturation), they report perceiving their fathers as more controlling. This finding seems to support existing notions that adopting mainstream culture may increase youth’s negative view of parental cultural behaviors, but this study shows that the adolescents’ negative view is toward only their fathers’ behaviors and not their mothers’.

Family processes and youth outcomes

Among family process variables, parent-child bonding (both maternal and paternal) is the most noticeable predictor of both depressive symptoms and antisocial behaviors in the point-in-time models. Explicit affection is a significant and negative predictor of depressive symptoms and conflict is a significant and positive predictor of antisocial behaviors. Notably, the positive effect of ethnic cultural orientation on youth outcomes is predominantly through increased explicit affection and parent-child bonding. Implicit bonding and firm control are not predictive of youth outcomes when all family process variables were considered simultaneously in a multivariate model. Implicit bonding, only when considered alone, is a significant predictor. This may be a measurement issue, as the scale is rather weak for mothers. Korean immigrant mothers report feeling more comfortable than fathers to adopt Western styles of explicit affection (e.g., Choi, Kim, Kim, et al., 2013), which may explain a low reliability of the scale among mothers. Or it may be that explicit affection is simply a stronger predictor among Korean American youth when various forms of parental affection are considered together. Interestingly, firm control is not a negative influence on youth outcomes, partially supporting the existing notion that firm control may not be as negative among Asian American families as it is the case for white families.

Although a few differences are noted in how cultural orientations influence family process by parent gender, identity enculturation enhances both maternal and paternal family processes. In addition, explicit affection and bonding are a significant protective factor for youth outcomes regardless of whether they are specific to mothers or fathers. The participating fathers in the survey show high rates of democratic parenting and close relationships and interactions with their children, even more so than mothers, in some cases. These characteristics may not apply generally to Korean immigrant fathers, who are often characterized as the opposite (Hahm et al., 2014; Min, 1998). Nonetheless, the findings of this study should serve as a springboard to additional and in-depth investigations of differences by parent gender in Asian American family processes and cultural orientations.

Cultural orientations and youth outcomes

As expected, language proficiency in both Korean and English plays a positive role in reducing youth problems, both indirectly and directly, suggesting the equal importance of heritage and mainstream language. The importance of English proficiency is related to fewer depressive symptoms concurrently, while speaking one’s heritage language has current (but mediated) and lasting benefits, even after accounting for prior problems. The benefit of bilingualism has been noted in previous studies (Adesope, Lavin, Thompson, & Ungerleider, 2010; Han & Huang, 2010; Schachter, Kimbro, & Gorman, 2012). Yet the pathways through which language proficiency positively impacts youth outcomes are not clear from the research (Choi et al., 2014). This study provides evidence for heritage language helping youth perceive parents as more affectionate and less controlling. Nonetheless, this question, especially in regard to longitudinal effect, is not yet fully answered. In addition, existing research often conflates heritage language proficiency with heritage cultural awareness (e.g., J. S. Lee, 2002), but the findings of this study indicate that the effects of each are distinct. Additional studies are warranted to confirm this finding and to further investigate the specific mechanisms through which language proficiency, apart from cultural proficiency, enhances youth development. This question is particularly important, given the strong and lasting positive role of heritage language proficiency.

The results also show that the impact of mainstream cultural orientation, specifically behavioral acculturation, on youth development may not be due to increasingly negative family processes. Instead, the results allude to a negative and direct influence. Specifically, the findings show that a higher level of behavioral acculturation is directly associated with higher rates of depressive symptoms and antisocial behaviors. Much of the existing literature shows that youth mainstream orientation negatively impacts youth outcomes, primarily due to the differing cultural orientations of youth and parents (e.g., Hwang & Wood, 2009). The results highlight that the negative impact of mainstream orientation postulated in the literature may in fact be dimension-specific. In other words, not all aspects of acculturation may have a negative impact on youth outcomes. The results of this study show behavioral acculturation to be the primary strong negative influence on youth outcomes (not language and identity acculturation), although this influence is not always mediated by family processes and is not lasting.

Language acculturation must be beneficial as it provides youth access to institutions, resources, and opportunities necessary to survive and thrive in society. In addition, identity acculturation may provide youth with a sense of belonging to the mainstream society, which has been linked to improved developmental outcomes (Faircloth & Hamm, 2005; Gillen-O'Neel & Fuligni, 2013). It should be also noted, however, that Korean American youth of this study reported a significantly higher rate of Korean identity (mean = 4.37 out of 5 point scale) than American identity (mean = 3.35) and the significant and positive impact of American identity was after accounting for Korean identity. In additional analyses, we did not find youth who reported mainly high American identity in this sample. In fact, Benner and Kim (2009) reported that youth with greater identity acculturation were more vulnerable, e.g., more susceptible to the negative impact of discrimination. Thus, it is possible that identity acculturation may have a positive effect on family processes and youth outcomes only in the presence of higher ethnic identity, supporting the benefit of bicultural identity.

It is conceivable that youth who are more involved in the mainstream culture, i.e., behaviorally acculturated, are exposed to the negative portrayal of minorities, thus harming their self-image and interfering with positive identity development (bivariate correlation between behavioral acculturation and Korean ethnic identity was −.36), which may in turn be detrimental to mental health. Or it may be that behavioral acculturation increases opportunities for less desired behaviors. It is posited that behavioral acculturation without identity acculturation (i.e., participating without a sense of belonging) may indicate a superficial participation in culture (Birman, 1994), which may leave youth more vulnerable both psychologically and behaviorally (Tosh & Simmons, 2007). Further studies are needed to unpack how and why behavioral acculturation may be damaging to youths’ outcomes and whether or not identity acculturation alone can be beneficial to youth development.

Study limitations

The study has its limitations. First, the sample size is not large enough to run the conceptual model as a latent construct SEM, which would have enabled testing of each construct for measurement quality. However, the model did enable us to examine bilinearity and multidimensionality of cultural orientations in a single hypothesized model, which allowed for greater precision in the findings. By doing so, we were able to identify unique contributions of each dimension to family processes and youth outcomes. Also, we were able to isolate that behavioral participation in the mainstream culture is unfavorable to youth development, while language and identity acculturation are not. Language, identity, and behavioral enculturation also uniquely promote youth outcomes. These findings point to the specific factors to target in interventions to minimize or reverse the negative impact of the acculturative process on youth.

Another limitation is the directionality of the model. The relationships among the constructs are likely to be interactive. For example, although parenting behaviors and parent-child relations may determine youth outcomes, it is also possible that youth behaviors in turn influence family processes (Kerr & Stattin, 2000). In other words, when children behave badly and are not happy, children and their parents are more likely to have conflicts and feel less bonded to each other. We need more longitudinal data to establish temporal sequences and account for the interactive nature of family processes and youth development.

Some measures have limitations. For example, the measure of implicit affection shows only moderate reliability of .45 for mothers, although it shows a reliability of .75 for fathers. Even so, this measure reveals several significant relationships. Additional relations might have surfaced with a better scale. Continued efforts are warranted to develop measures that can capture culturally-specific family processes among non-Western families. We used the LIB scale extensively because it captures both bilinearity and multidimensionality of cultural orientations. However, LIB did not have value measures and future studies should include value items to examine how value acculturation or enculturation influence family processes.

We mainly focused on youth reports of family processes when parents’ reports are important as well. Given the importance, we ran the same model using parents’ reports. With the exception of parent-child conflict, we found no relation between youth cultural orientations and parent-reported family processes. Specifically, Korean ethnic identity decreased parent-reported conflict that predicted youth negative problems. A strong sense of Korean identity among children may help parent perceive less conflict, implying the positive effect of youth enculturation in parent perception of family process. However, when considered together with youth reported parent-child conflict, the predictive power of parent-reported conflict disappeared. The results do not necessarily mean parents’ perceptions are unimportant; rather, they highlight the importance of youth perceptions. In addition, parents’ perceptions may have a different pathway in affecting youth development, as shown in other studies (e.g., Choi et al., 2008).

Further, the results may not be generalizable to other Asian American groups. Although Asian Americans share a common culture, they vary widely in family processes (Choi, Kim, Pekelnicky, et al., 2013). Although additional comparative analyses across Asian subgroups are warranted to discern both common and unique mechanisms in family processes, the findings of this study on Korean Americans can build understanding of other subgroups who may share some cultural and SES similarities. Relatedly, although the models account for family SES and single parenthood, the parent sample is largely urban, middle class, and married. It is plausible that the effect of family SES is pervasive in family processes and potentially more powerful than youth cultural orientation. Additional data with greater SES variability are needed to better understand this role.

Conclusions

The number of racial-ethnic minorities and immigrants in the United States is growing, and they must navigate a racialized society while juggling multiple cultural contexts and conflicting demands. This study provides more detailed empirical evidence to guide the development of effective interventions for Asian American youth. Acculturation and enculturation are no longer tasks unique to immigrant youth as today’s youth are exposed to multiple cultures and the opportunities to adopt different cultures abound. Our findings should be expanded to other global contexts, which can in turn inform studies with children of immigrants in the United States.