Upholding Familism Among Asian American Youth: Measures of Familism among Filipino and Korean American Youth

Abstract

Although it is one of the core cultural values of Asian American families and an influential determinant of youth development, familism remains under-studied among Asian Americans and, despite crucial within-group heterogeneity, lacks subgroup specificity. This study describes the ways in which two major Asian American subgroups of youth, i.e., Filipino Americans and Korean Americans, maintain traditional familism. Specifically, this study constructed six self-report subscales of familism utilizing underused and new survey items and tested their psychometric properties. Using data collected from Filipino American (n=150) and Korean American (n=188) adolescents living in a Midwest metropolitan area, the measures were examined for validity and reliability for each group and, when appropriate, for measurement invariance across the groups. The main findings are that the finalized scales demonstrated solid reliability and validity (e.g., content and construct) in each group and some invariance and that core traditions, in the form of familism values and behaviors, persevere among second-generation Asian Americans, although familism was more evident among Filipino American youth than in Korean American youth. In both groups, subdomains of familism were not as discrete as found among their parents, who were predominantly foreign-born first-generation immigrants. The finalized familism scales were associated differently with several correlates including acculturation variables and youth outcomes. The findings are discussed with a call for further empirical research of diverse ethnic groups and immigrant generations to more accurately account for how family process interacts with cultural origin and acculturation.

Introduction

Familism, broadly defined as family-centered over individualist values, is regarded as a core trait of many non-Western immigrant families in the U.S. and an influential factor that may facilitate positive adolescent development (Fuligni & Masten, 2010). Despite its distinct presence and critical implications for Asian American youth development, familism remain under-investigated among Asian American families. Because familism has come to be most closely associated with Latinx families (e.g., Schwartz et al., 2010), the prevailing conceptualization of familism may not be entirely applicable to Asian American families. Moreover, Asian Americans, like other communities subsumed in broad racial and ethnic categorizations, are highly heterogeneous, and its subgroups may ascribe to different subdomains of familism (Choi, Kim, Noh, Lee, & Takeuchi, 2018), which may further contribute to inconsistent findings. Without an enhanced conceptualization of and appropriate measures for assessing endorsement of familism specific to Asian Americans and their subgroups, research efforts are hampered in understanding how familism operates in Asian American families. Addressing these gaps in the current literature, this study develops a model of familism specific to two major subgroups of Asian Americans, i.e., Filipino American and Korean American youth, through scale development and psychometric testing of six domains of familism on adolescents (i.e., traditional manners and etiquette; respect for adults; caring for aging parents; centrality of family; harmony and sacrifice; and parental expectation of family obligation). This study also examines interrelations between the finalized subdomains of familism and several youth outcomes and acculturation variables to add construct validity of the scales and to establish a foundation to produce a better understanding of how familism among Asian American ethnic groups hinder or facilitate positive youth development.

Familism and Adolescent Development

Familism is a cultural value that emphasizes family unity and loyalty, prioritizes the family over individual needs, and expects support and commitment to and from the family (Schwartz, 2007). This multifaceted construct is composed of structures, attitudes, and behaviors and is characteristic of collectivist cultures (S. Y. Kim, 2010). The term familism was first introduced to describe the organizing values of traditional peasant-based societies that honored the preeminence of the family, contra to those of individualistic modem urban societies (Burgess & Locks, 1945) and initially gained traction as a central tenet of Latinx communities (Zinn, 1982). Zinn’s (1982) explication of Arce’s (1978) four domains of familism remains influential in contemporary scholarship. The demographic domain accounts for family size and makeup; the stmctural domain examines the geographic and social boundaries of kinship; the normative domain examines beliefs regarding the importance of family and family life; and the behavioral domain examines practices of providing emotional, financial, and physical support to kin. A global conception of familism that takes into account all four domains has rarely been applied. In particular, unlike the demographic and structural domains that may be externally evident, the normative and behavioral domains can be vague and subject to group variability. Among immigrant and/or cultural minorities in the U.S., such as Asian American adolescents and their families, acculturative pressure further complicates the normative and behavioral domains of familism. Accordingly, the study of familism in American scholarship has been ensconced in a broader narrative of acculturation of immigrants from collectivist cultures to an individualistic culture (Schwartz, 2007).

Often known as familismo among Latinx families, the significance of familism among Asian Americans remains understudied despite the continued centrality of familism to Asian American families (Toyokawa & Toyokawa, 2013). Researchers have noted the primacy of familism among Asian communities in the U.S. (e.g., Yee, DeBaryshe, Yuen, Kim, & McCubbin, 2007). For example, Asians in the U.S. are more likely to live in multigenerational households (Cohn & Passel, 2018) and are also more likely to assist in caring for or financially supporting family members than are Whites, Blacks, or Latinxs (Yee et al., 2007). In addition, Asian American college students place greater attitudinal emphasis on family interdependence and spend more time assisting their families than their White, Latinx, or African American peers do (Tseng, 2004). Similarly, Asian American adolescents, along with those from Latinx American families, were more likely than White American adolescents to believe that they should make sacrifices for the family and consider family impact when making important life decisions (Fuligni, 2001; Fuligni, Tseng, & Lam, 1999).

A preponderance of relevant scholarship focuses on familism as a protective resource for Latinx adolescents (e.g., Stein, Gonzales, Cupito, Kiang, & Supple, 2015). A similarly positive role of familism has been identified in Asian American adolescent development (e.g., Liu et al., 2012). For example, family obligation, one aspect of familism, protects against the negative effects of financial stress on academic outcomes among Asian American high schoolers (Kiang, Andrews, Stein, Supple, & Gonzalez, 2013). A study, using a composite scale of familism, found a positive relation between familism and emotional adjustment (Juang & Cookston, 2009). Likewise, family assistance, which consists of concrete behaviors related to family obligation values, was found to be associated with greater well-being among Asian American ninth-graders (Telzer & Fuligni, 2009). Familism has also shown to enhance parenting that promotes youth development and acts as a protective factor against adversity such as chronic poverty (Jocson, 2020) and to mitigate the negative effects of major stressors, such as racial and cultural discriminations (Corona, Campos, & Chen, 2017). These protective effects may extend into young adulthood (Fuligni & Masten, 2010).

However, several recent qualitative as well as quantitative studies with Asian American adolescents and young adults throw into question previous findings of the protective effects of familism by revealing mixed and harmful effects among Asian Americans. For example, recent qualitative studies (e.g., Nadal, 2011) have illuminated a much more complex process of how familism unfolds among Asian American families. Emphasis on family obligation, for instance, can take a significant psychological toll (Le Espiritu, 2003). More specifically, familism, especially when coupled with gendered expectations, can lead to greater familial care burdens and inhibitions placed on daughters and can cause serious mental distress among Asian American young women (Hahm, Gonyea, Chiao, & Koritsanszky, 2014). Inconclusive findings may also be attributed to familism being a multifaceted construct (Schwartz, 2007), its nuances not yet fully explored (Alampay, 2014).

As a result, different subdomains of familism are examined in different studies, producing inconsistency. In accord with this point, a set of recent empirical studies also adds complexity to the landscape of findings. For example, among Filipino American and Korean American families, familism when measured as participation in family-centered activities (e.g., spending time together as a family) was indeed beneficial for adolescents (Choi, Park, Lee, & Lee, 2020). However, a high endorsement in several other domains of familism (e.g., respect for adults, importance of maintaining harmony within the family, sacrifice for the family, and family obligation) was associated with higher depressive symptoms among Filipino American adolescents (Choi, Park, Lee, Yasui, & Kim, 2018). Moreover, contrary to other findings that familism promotes good parenting (Jocson, 2020) and strengthens family ties (Kennedy & Ceballo, 2013), familism and adherence to traditional values were significantly correlated with negative parenting behaviors (e.g., psychological control, parental self-worth based on children’s performance) among Filipino American and Korean American parents (Choi, Park, Lee, Kim, & Noh, 2019), which can engender parent-child conflict. Intergenerational cultural conflict and subsequent parent-child conflict (Choi, He, & Harachi, 2008) are a major etiology of mental distress among Asian American adolescents and young adults (Choi, Park, Lee, Takeuchi, & Noh, 2020), and certain subdomains of familism may be a source of such conflict. These studies collectively question the utility of familism as a global and composite indicator.

Multidimensionality and Subgroup Differences

The diversity of the U.S. Asian population poses difficulties in generalizing the influence of familism among Asian Americans. Despite being a core family value in variant groups with some commonalities, familism, a multidimensional construct, may emphasize different subdomains in their respective cultures. For example, familism was positioned as a distinctly Latinx value, and filial piety as distinctly Asian, but it was found that the two cluster (along with communalism) into a single latent factor (Schwartz et al, 2010). Contrastingly, Latinx familism centers on an open exchange of emotional and instrumental support to and from the family, while more distinct among Asian American families are respect for and obedience to parents and elders, and attendance to parents’ needs (Schwartz et al., 2010). Building on Schwartz et al. (2010), a study found that Latinxs reported higher mean values of familism than did Asians (Campos, Ullman, Aguilera, & Dunkel Schetter, 2014). However, a study that examined whether Asian and Latinx familism can be measured using the same scale, did not find support for construct invariance between Asian American and Latinx family values (Toyokawa & Toyokawa, 2013). A closer examination of familism anchored among Asian Americans can clarify this inconsistency. A few researchers have measured and analyzed the normative and behavioral domains of Asian American familism, using measures such as the Asian Values Scale (B. S. K. Kim & Hong, 2017), the Dual Filial Piety Model (Yeh & Bedford, 2003), the Family Obligation scale (Fuligni et al., 1999), and Korean ga-jung-kyo-yuk (“family socialization”) measures (Choi, Kim, Pekelnicky, & Kim, 2013). These scales, nevertheless, assess different aspects of familism or measures of familism values and behaviors in a single composite scale. A comprehensive but domain-specific approach is needed to sharpen current understanding of Asian American familism.

Notwithstanding the growing scholarship on familism among Asian Americans, the literature rarely addresses familism among subgroups of Asian American adolescents. As several studies (e.g., Chao, 1994; Choi et al., 2013; Wu & Chao, 2011) have established, scales validated with specific subgroups of Asian Americans are necessary to capture ethno-specific attitudes and behaviors of familycentric concepts. A refined understanding of within-group differences is complementary to studies on pan-ethnic differences between Asian Americans and other racial/ethnic groups; together, they support further understanding of the mechanisms and interactions of acculturation in Asian American youth development.

To elucidate ethnic subgroup differences of familism among Asian Americans, this study focuses on Filipino Americans and Korean Americans. The selection was purposeful, based on the overlapping and contrasting sociocultural profiles of these two groups. Largely post-1965 immigrants, Filipino Americans and Korean Americans are among the top five major subgroups that comprise more than 80% of Asian Americans (Census, 2017). These two subgroups share similar socioeconomic status, i.e., comparable median income and college education level (Pew Research Center, 2017), diminishing a confounding class effect. Both groups also share high rates of religiosity (Lien, 2004), which can influence endorsement of familism (Chen & Jeung, 2012). However, they differ notably in acculturation and family process. For example, Filipino Americans are thought to be most assimilated among Asian American groups and exceed Korean Americans on key measures of acculturation, characterized by fluency in English (82% English proficiency compared to 47% in Korean Americans), a greater residential assimilation and less reliance on their coethnic community, more professional occupations in the mainstream society, and pre-immigration acculturation, due to their colonial history (Oh & Min, 2011). Conversely, Korean Americans, especially immigrant adults, are arguably the most socio-culturally segregated of Asian American subgroups (Pew Research Center, 2013), remain largely monolingual, socialize primarily with coethnics, and reside in areas with a high concentration of Koreans (Oh & Min, 2011).

Although Filipino Americans appear more acculturated than Korean Americans in these external indicators, both groups highly endorse familism (Choi, Kim, Noh, et al., 2018; Fuligni & Masten, 2010). In fact, Filipino American parents have been found to express more familism and are more likely to reinforce them in their children, specifically scoring higher on their expectation of family obligation from daughters (Choi, Kim, Noh, et al., 2018). Similarly, Filipino American adolescents, along with Latinx adolescents, were more likely than East Asian and White adolescents to place a strong value on family obligation (Fuligni & Pedersen, 2002). This stronger endorsement of family obligation accounted for a significant portion of Filipino participants’ greater tendency to financially support their families and to live with their parents, in comparison to their East Asian and White counterparts. These findings refine previous profiles of Filipino Americans demonstrating more acculturated behaviors and attitudes than East Asian Americans. In addition, while both groups similarly endorse the centrality of family, Filipino American familism may emphasize family care obligation and providing support and assistance to the family, while Korean American familism underscores greater formality in the relationship (Fuligni & Masten, 2010).

Current Study

To better comprehend the dynamic nature of familism as reported by Asian American adolescents, this study develops a model of familism among Filipino American and Korean American youth through the development and psychometric testing of six subdomain measures of familism. Without such careful consideration of the subjective context in which familism occurs, i.e., taking into account subgroup-specific nuances, as well as the perspective of the child along with that of the parent, familism’s contributions to Asian American youth development cannot be fully understood. Indeed, the gap between children’s endorsement of familism and that of their parents’ may be pivotal to understanding interactions between familism and youth outcomes. To date, very few studies have considered ethnic- and generation-specific measures of familism among Asian American families. More nuanced measures of familism were recently developed with parent samples (Choi, Kim, Noh, et al., 2018), but it cannot be assumed that familism measures validated with parent samples can be applied with succeeding generations without validation. This study on familism at the youth level among Filipino and Korean Americans, along with that of parents (Choi, Kim, Noh, et al., 2018), may offer important insight into different outcomes among Asian American subgroups broadly, and between Korean American and Filipino American youth specifically. This study also examines familism as a multifaceted construct. While surveying individual domains of familism is instructive and necessary, simultaneously testing multiple domains of familism values and behaviors in a single paper allows for inclusive and distinctive expressions of familism. Understanding familism among Filipino American and Korean American adolescents holds important implications for family socialization, acculturation, and acculturation gaps (Park, Kim, Cheung, & Kim, 2011).

Scale Development

This study set out to identify aspects of familism that are particularly relevant to Filipino American and Korean American families. Several steps of both etic and emic approaches were taken to generate a series of familism items and scales, including (1) extensive literature review that included a search for existing Asian familism scales, (2) focus groups, (3) generation of nearly 100 preliminary items, (4) review of those items by expert panels and the research team, and (5) pretest of the items. This process produced a total of six scales and 27 items to be tested for psychometric properties reported in this study.

Based on a comprehensive review of existing literature on Filipino Americans and Korean Americans, some of which are described earlier, several subdomains were considered to capture essential traits of familism common across cultures, as well as those that are more salient among Asian Americans (e.g., characteristics of filial piety) and/or in each subgroup. For example, the study started with several constructs such as centrality of family [a likely universal trait of familism (S. Y. Kim, 2010)], respect for adults and cultural behaviors that symbolize respect for adults and family hierarchy [likely more emphasized among Korean American families that stress formality in the family) (Choi et al, 2013)], values and behaviors around caring for aging parents and family care obligation [supposedly more salient among Filipino American families (Fuligni & Masten, 2010)], and expectations of maintaining harmony within the family and making sacrifices for the family [deeply rooted cultural ideals among Filipino American families (e.g., Nadal, 2011)].

Along with measures of Latinx familismo (e.g., Lugo-Steidel & Contreras, 2003; Sabogal, Marin, Otero-Sabogal, Marin, & Perez-Stable, 1987), existing measures were instructive in capturing pan-ethnic constructs of Asian American familism, such as the Family Obligation Scale, assessing adolescents’ sense of obligation to support, assist, and respect the family (Fuligni et al., 1999); the Dual Filial Piety Model, measuring attitudes and behaviors around how children should treat their parents (Yeh & Bedford, 2003); and the Asian Values Scale, which has been validated for Chinese, Filipino, Japanese, and Korean American college students (B. S. K. Kim, Yang, Atkinson, Wolfe, & Hong, 2001). In addition, a few underused scales were located that were judged to specifically measure Filipino and Korean familism. Three Filipino family values and parenting scales were identified: the Panukat ng Pagkataong Pilipino (PPP) and the Panukat ng Ugali at Pagkatao (PUP), both developed in the Philippines (Enriquez & Guanzon-Lapeña, 1985), and the Enculturation Scale for Filipino Americans (ESFA; del Prado & Church, 2010). These three measures assess idealized Filipino character traits, attitudes, and core family values. The ga-jung-kyo-yuk measures for Korean Americans (Choi et al., 2013) were also selected for investigation; these recently developed measures assess family socialization and enculturation processes of Korean American families in which traditional parenting ideology, such as a hierarchical family order, age veneration, and respect for elders and parents, prevails. Through a review of the literature, it was confirmed that Filipino Americans and Korean Americans share much commonality but also notable or nuanced differences in familism, and that many familism constructs can be applicable in both groups.

To identify aspects of familism that may be missing in existing literature and scales, a total of six focus groups for Filipino Americans (three parent groups and three youth groups) and nine focus groups for Korean Americans (five parent groups and four youth groups) were conducted. The participants were asked to identify traits of family process as uniquely Filipino or Korean. Nearly 100 additional items were generated through the analysis of qualitative data obtained via these focus groups as well as an extensive literature review. Items were rendered in English, Tagalog, and Korean as appropriate. In addition, two five-member panels composed entirely of Korean Americans or Filipino Americans were recruited on the basis of bilingual/bicultural capacity, experience as a parent or working with parents and youth in the community, and an understanding of the research process. The panels reviewed the generated scales and items for the etic/emic nature of the questions, the applicability of the situational context of the questions, and the accuracy of translation. Each item was then examined for redundancy, length, level of difficulty, double-barreling, and ambiguity (DeVellis, 1991). Only those items that were believed to be central to the construct of Asian American familism were retained. Scale items, including translated versions, were pilot-tested using five Korean American parent-child dyads and five Filipino American dyads. The items were further edited, refined, or removed entirely based on the results of these pilot tests.

The resulting 27-items, mapped onto six scales of Asian American familism, are shown in Table 2. This study tests multiple aspects of psychometric properties, including reliability and validity (content, construct, discriminant, and divergent) as well as measurement invariance.

Intercorrelations were examined with the finalized subdomain scales, youth outcomes, and acculturation variables to add construct validity and demonstrate the effectiveness of the scales. Familism is subject to ongoing processes of acculturation and enculturation, especially among immigrant families (Russell, Crockett, & Chao, 2010). Acculturation occurs unequally across various domains. That is, among immigrant parents, core values such as familism may not easily change but peripheral behaviors, especially those that draw legal sanctions (e.g., traditional disciplinary practices), can change more readily (for details, see Choi & Kim, 2010; Choi et al., 2013). Nevertheless, child generations of immigrants are acculturating much faster than parent generations, and a general trend was expected in which acculturation indicated by English proficiency, American identity, and mainstream cultural practices would be negatively correlated with subdomains of familism, while enculturation (i.e., heritage language proficiency, ethnic identity, and heritage cultural practices) would be correlated in the opposite direction. In regard to youth outcomes, based on a handful of existing studies, scales of familial care responsibility and family obligation were expected to be positively associated with mental distress, but not with other outcomes such as GPA and antisocial behaviors. To stay within a reasonable scope for this paper, the associations were examined in bivariate intercorrelations, not in multivariate analyses. This was also appropriate because the data that was used for this study were collected mainly to pretest new and underused measures, rather than to fully develop multivariate models.

Measurement Invariance

Empirical testing of measurement invariance, although uncommon and particularly rare across Asian American subgroups, is a prerequisite to conduct comparative studies of different cultural groups. This study tested measurement invariance of the scales, both conceptually and empirically. Conceptual invariance, the most basic invariance, is exhibited when concepts represented in measures hold equivalent intrinsic meanings across groups. In this study, conceptual invariance was established through extensive literature review and review of items and scales by each of the Korean American and Filipino American focus groups and expert panels. Many of the existing items, even if developed for one group, can be applicable to both (i.e., pan-ethnic items), and were included in this study to be tested for their appropriateness. Some modifications of existing items were made, such as specifying example behaviors that are relevant to each group.

The meaning of constructs, however, even if similar in terms of face validity and conceptually invariant, may still differ across groups. Thus, all items that met criteria for validity and reliability, respectively, in each group were then tested for other types of invariance, i.e., functional, item, and scalar invariance (Hui & Triandis, 1985). Functional (or structural) invariance establishes the equivalence of factor scores across groups and is tested in this study through examining the variance of factor scores across groups, the factor intercorrelations across groups, and the latent mean score on each factor across groups. Tests of structural invariance instruct on whether individual items are related to the latent variables in a consequential way across groups.

Item invariance tests whether the individual item parameters are consistent across groups. Items are subject to sequential tests of configural, metric, scalar, and strict variance (Widaman & Reise, 1997). Configural invariance is supported when the pattern of fixed and free parameters is equivalent across groups (Widaman & Reise, 1997). In the present study, a finding of configural invariance would suggest that the latent concept can be discussed with both Korean American and Filipino American youth. However, it is possible that the youth may answer items differently across groups, because factor loadings may vary. Thus, metric (or weak) invariance requires that the magnitudes of factor loadings are similar across groups; only then can relations between the scale and other variables be compared across groups. Scalar (or strong) invariance is established when both item intercepts and item loadings are equivalent – representing the same magnitude – across groups. Scalar invariance is necessary to compare mean values across groups. Lastly, strict invariance is the hardest to establish in practice, and implies that the error variance of scores across groups is equivalent.

Although all of the types of invariance were tested, the present study aims to ascertain at least metric invariance of the measures used with Filipino American and Korean American youth to ensure compelling comparative analyses between the two groups. Measures that do not exhibit at least metric invariance, but which otherwise have sound psychometric properties in each respective sample, can still be used with each group, but their latent means cannot be compared across Filipino and Korean American youth.

Methods

Overview of the Project

This study uses data from the Midwest Longitudinal Study of Asian American Families (MLSAAF). The primary goal of MLSAAF is to formulate an Asian American family process model to explain Asian American youth development. The data used for this paper was collected in 2013, the first year of the study, to develop and pretest a series of existing and new measures that capture culturally specific family processes among Filipino American and Korean American families with children between the ages of 12 and 17. Families were eligible to participate if mothers were of Filipino or Korean heritage and had at least one adolescent child. A total of 680 individuals, comprised of 188 Korean American youth, 186 Korean American parents, 155 Filipino American youth, and 151 Filipino American parents living in Chicago and the surrounding Midwest areas, were surveyed. Of those, 183 Korean American families were parent-child dyads, while 133 Filipino American families were parent-child dyads. This paper used youth data only.

The self-administered questionnaires (available both in paper and online formats) were distributed to eligible participants and collected either in person, by mail, or via web. The survey was available in English, Korean, and Tagalog; the English version of the survey was translated into Korean and Tagalog using a committee translation (Epstein, Osborne, Elsworth, Beaton, & Guillemin, 2015) in which multiple translators made independent translations of the same questionnaire and, at a consensus meeting, reconciled discrepancies and agreed on a final version. The initial version of the survey was pilot-tested with several parents and youth from each subgroup and further revised for clarity before being administered to the family. The majority of youth samples used the English version and less than 3% of samples (n=5 Korean youth and n=4 Filipino youth) used their heritage language version of the survey.

Sample Characteristics

The youth samples consisted of equivalent proportions of girls and boys, with a mean age of 15.42 years. The Filipino American sample was slightly older, with 78.1% of Filipino American youth in high school, compared to 65.6% of Korean Americans. Filipino Americans were also more likely to be U.S.-born (70% Filipino Americans vs. 57.2% Korean Americans), and less likely to receive free/reduced-price school lunch.

Parent samples consisted mostly of foreign-born middle-age mothers. Most Korean parents were currently married, but about one-fourth of Filipino parents were unmarried at the time of survey. Foreign-born (first-generation immigrant) Filipino parents (90% of the sample) had resided in the U.S. for an average of 19 years, while all Korean parents were foreign-born and had been in the U.S. for an average of 16 years. A majority (over 90%) of parents of both groups were employed, with the exception of Korean mothers. One-third of Korean mothers were not employed at the time of survey. Additional details are reported in Table 1. In general, the parental characteristics of the study samples are consistent with the socio-demographic profiles of Filipino American and Korean American families as reported in the Census or national surveys such as the Add Health (Harris, 2009).

[Table 1].

Demographic Characteristics of the Samples

		Filipino Americans	Korean Americans	Total2
Youth
	Age	15.60 (1.77)	15.28 (1.81)	15.42 (1.79)
	Girls	49.0%	52.2%	50.7%
	High School	78.1%	65.6%	71.2% (FA>KA*)3
	U.S.-born	69.3%	57.2%	62.6% (FA>KA*)
	Years in U.S.1	6.42 (4.92)	8.08 (4.28)	7.45 (4.59) (FA<KA*)*
	Free-lunch	11.3%	17.2%	14.5%
Parents
	Age	46.72 (6.81)	46.56 (4.32)	46.63 (5.55)
	Married	76.0%	92.4%	85% (FA<KA***)
	Foreign-born	90.0%	100.0%	95.5% (FA<KA***)
	Years in U.S.1	19.43 (11.78)	16.11 (9.01)	17.52 (10.39) (FA>KA**)
	College Ed
	Mothers	95.6%	84.2%	89.1% (FA>KA*)
	Fathers	76.5%	84.4%	80.3%
	Unemployed
	Mothers	7.0%	33.8%	22.4% (FA<KA***)
	Fathers	5.6%	9.7%	7.5%

Note:

¹Foreign-born subjects only

²Statistical significant differences of mean or proportion across the two groups.

³FA Filipino Americans KA Korean Americans

^†p < .10

^*p < .05

^**p < .01

^***p < .001

Analysis Strategy

Psychometric properties.

Using SPSS (v.22) and Mplus (v 7.4), various components of psychometric properties were tested to establish content and construct validity, including validation of a sound factor structure. The measures were first examined separately for each group, then compared across Filipino and Korean subgroups. The content validity of each scale was tested by examining mean and standard deviations of each item and of the entire scale, internal consistency within the scale, and item-total correlations among items in the scale (Nunnally & Bernstein, 1994). To assess the factorial structure and content validity at scale level, confirmatory factor analyses (CFA) were conducted, modeling each scale as a single factor with multiple indicators. In CFA, factor loadings were examined, as well as the fit of the measurement model, indicated by χ² statistics, Comparative Fit Indices [CFI >.90 indicating a good fit (Bentler, 1990)], Root Mean Square Error of Approximation [RMSEA<.05 good fit, between .05 and .10 a fair to mediocre fit, and >.10 a poor fit (MacCallum, Browne, & Sugawara, 1996)]. Items with near zero endorsement, item-total correlation less than .3 (Nunnally & Bernstein, 1994), and factor loading less than .4 (Floyd & Widaman, 1995) were to be excluded. If items were excluded from the initial scale, another series of analyses were to be executed to obtain the new scale mean, internal consistency reliability, and CFA for the construct.

Multi-factor CFA was run for the six familism scales in a single CFA model, with each scale specified as a discrete factor (Nunnally & Bernstein, 1994). CFA fit indices, modification indices (MI), and correlations among the six scales were used to examine whether each item was loaded to its designated factor and whether each scale was discrete from others but also reasonably convergent. Given that each scale is a subdomain of familism and shares latent traits with the other scales, correlations among scales were expected to be statistically and positively significant (exhibiting convergent validity), but not too high (exhibiting divergent validity if r < .85) (Campbell & Fiske, 1959). Based on the results, a few additional steps were considered, such as excluding additional items from scales, merging factors, or loading items to a different factor. When modifications were made, another series of analyses were executed to obtain a new scale mean, retest for internal consistency reliability, and obtain a single-factor CFA for the modified scale. With the finalized sets of scales, interrelations were examined among Filipino and Korean samples respectively. Pair-wise bivariate correlations were run among factors, factors and youth outcomes, and factors and acculturation variables.

Measurement invariance.

When a scale showed a configural invariance (i.e., common items, a comparable factor structure, and a good model fit in both groups), metric, strong, and strict invariance testing were conducted to investigate measurement invariance across Filipino American and Korean American youth (Wang & Wang, 2012). The purpose of the multiple-group CFA was to test if the structure of the measures was equivalent across Filipino American and Korean American youth, empirically testing whether measurement parameters (e.g., factor loadings, intercepts, and error terms) were invariant across groups. The analysis was conducted in a stepwise approach—from metric, strong, and strict invariance. First, in the unconstrained model, all of the parameters (e.g., factor loadings, intercepts, and error variances) were set free across the two groups and no equality restrictions were imposed on any parameter across the groups. The factor loadings and thresholds were allowed to differ across the groups with the scale factors fixed at one, and the factor means fixed at zero. This serves as a baseline model. In the next step, factor loadings were constrained to be equal to test metric invariance. The modification indices (MI) were used to identify the factor loadings to be constrained. Strong invariance was tested by adding equality constraints on intercepts. Finally, strict invariance was examined by further constraining error variances to be equal. At each addition, the differences in χ² statistics (Δχ²/Δdf) were examined to test statistical differences between unconstrained model and the constrained model at the next level. In addition, the change in CFI was considered; ΔCFI < .01 indicates invariance (Cheung & Rensvold, 2002).

Measures

Response options for items are mostly a 5-point Likert Scale, unless noted. For example, they are: (1) “Not at all,” (2) “Not much,” (3) “Moderately,” (4) “Much,” and (5) “Very much,” or, (1) “Strongly disagree,” (2) “Disagree,” (3) “Neutral,” (4) “Agree,” and (5) “Strongly disagree.”

Familism scales.

Traditional manners and etiquette.

This scale assesses how much parents of the youth participants emphasize practicing a set of traditional manners and etiquette that symbolize respect for elders. Four items from the Important Traditional Korean Etiquette scale (Choi et al., 2013) were adopted but revised to add examples appropriate to Filipinos. For instance, to greet adults/elders properly, Koreans bow to adults, saying appropriate greeting words for adults (an-nyung-ha-se-yo), while Filipinos gently place the back of one’s hand on elders’ forehead and say, “monopo.” Similar changes were made to other items that describe Korean/Filipino manners and etiquette to show respect to adults and elders. Based on literature and focus groups with Filipino Americans, one additional item was created that asks the extent to which parents emphasize the importance of acknowledging authority figures.

Respect for adults.

This four-item scale adopted one item from Panukat ng Ugali at Pagkatao (PUP, Enriquez & Guanzon-Lapeña, 1985) (“not fight or talk back to older person out of respect”), and two items from the Latino Familism Scale (Lugo-Steidel & Contreras, 2003) that assess absolute obedience to and respect for older persons regardless of one’s contrary views. One additional new item was created based on a literature review (de Guzman, 2011; Wolf, 1997) that highlights the importance of upholding parental wishes over the child’s.

Caring for aging parents.

Children’s sense of obligation to care for aging parents was assessed with a total of five new items constructed from focus group interviews and extant literature (e.g., Le Espiritu, 2003; Lim, 2011; Nadal, 2011). Both Korean American and Filipino American youth in focus group and individual interviews viewed the tradition of caring for aging parents as something particularly strong in their culture. Some of the items resemble the items of the Latino Familism scale (Lugo-Steidel & Contreras, 2003), but new items intentionally use verbiage from interviews with Filipino American and Korean American families to better capture cultural nuances.

Centrality of the family.

Similar to Caring for Aging Parents, the centrality of the family surfaced as one of the most distinctive feature of Filipino American and Korean American families both in interviews and in literature reviews (Enriquez & Guanzon-Lapeña, 1985). Youth in focus group interviews said that the degrees to which family occupies central importance in one’s life is unique to their culture. Filipino American youth, in particular, stated that they maintain close ties with family members across generations, despite adverse personal circumstances, and even if their relatives live far away. They also thought that Filipinos are unusually willing to share their homes with relatives in need, indicating close family relations. One item (“It is acceptable that several generations of a family share one household”) from ESFA was included in this scale because it further highlights the cultural norm of sharing the home with multiple generations.

Harmony and sacrifice.

This five-item scale measures the degree of harmony and sacrifice made by an individual to the benefit of both family and non-family members. Four items were newly developed, ascertaining the importance of maintaining harmony at the expense of one’s own needs and desires, and the degree of how much one should sacrifice for the greater familial good. In addition, one item (“A person should support members of the extended family if they are in need, even if it is a big sacrifice for me”) was adopted from the Latino Familism Scale (Lugo-Steidel & Contreras, 2003) because it echoed the sentiments of Filipino American focus group participants and, to a lesser degree, Korean American participants as well.

Parental expectation of family obligation.

This scale was developed to assess youth’s perception of the familial obligations expected of them from their parents, i.e., how much their parents want them to live close, to help out, or to live close to help out, and/or to take care of them when they age. The Caring for Aging Parents taps into the general understanding and endorsement of the essential family value of caring for aging parents, but the Parental Expectation of Family Obligations items ask how youth are socialized in the family and perceive of parental expectations to perform those values. A high level of family obligation, particularly among Filipino American families, is noted in the literature (e.g., Nadal, 2011) and was corroborated in focus groups.

Acculturation variables.

Language competency.

Adopted from the Language, Identity, and Behavior (LIB) acculturation measure (Birman & Trickett, 2002), two sets of one parallel item (two total) measured youth language competency in speaking both their heritage language (Filipino or Korean) and host language (English).

Behavioral cultural participation.

Also adopted from the LIB (Birman & Trickett, 2002), 12 items asked about participation in either heritage or American cultural activities such as social gatherings, media use, and peer composition [α=.80 (Filipino American) and .77 (Korean American) for heritage cultural participation; α=.74 (Filipino American) and .78 (Korean American) for host cultural participation].

Identity.

Ethnic and American identity were assessed, respectively, using 10 questions from LIB (Birman & Trickett, 2002), asking the extent to which youth identified themselves as Filipino/Korean or American [α=.79 (Filipino American) and .82 (Korean American) for ethnic identity; α=. 77 (Filipino American) and .84 (Korean American) for American identity].

Youth outcomes.

Depressive symptoms.

Mental distress was assessed based on 14 depressive symptom items from the Children’s Depressive Inventory (Angold, Costello, Messer, & Pickles, 1995) [α=.94 (Filipino American) and .92 (Korean American)].

School grade.

GPA was computed based on grades in English, math, social studies, and science.

Antisocial behaviors.

Behavioral problems were measured with a 19-item list adopted from DSM-IV conduct disorder criteria (Gelhorn et al., 2009) and the Add Health, including bulling, physical fights, hurting others, and stealing. The variable was constructed to 0 for none and 1 for any antisocial behavior.

Results

Psychometric Properties

Results of psychometric properties are summarized in Table 2. Overall, both Filipino American and Korean American youth endorsed items in each of the six domains of familism. Exceptions are found in the responses of Korean American youth: means of item #4 of Respect for Adults, and item #1 of Parental Expectation, were lower than 3 – which indicates “moderate” endorsement of the item. On all subscales, Filipino American youth scored higher compared to Korean American youth (e.g., Respect for Adults (3.74 vs. 3.37, p<.001), Caring for Aging Parents (4.15 vs. 4.02, p<.05), Centrality of the Family (4.14 vs. 3.78, p<.001), Harmony and Sacrifice (3.90 vs. 3.68, p<.01), and Parental Expectation of Family Obligation (3.77 vs. 3.22, p<.001). The only exception was that Korean American youth reported higher means of Traditional Manners and Etiquette (4.34 vs. 4.04, p<.001). All six scales showed acceptable to good Cronbach’s α reliability coefficients ranging from .66 to .83 in both Korean American and Filipino American youth.

[Table 2].

Measures of Familism and Single-Factor CFA

Constructs	Mean (SD)		Alpha1 Item-Total2			Factor Loading		Factor Loading (Modified)
Items	Korean	Filipino	Korean	Filipino		Korean	Filipino	Korean	Filipino
F1. Traditional Manners and Etiquette	4.34 (0.67)	4.04 (0.87) ***	0.81	0.81	χ2 CFI RMSEA	4.78*** 1.000 0.001	4.64*** 1.000 0.001	4.78*** 1.000 0.001	4.64*** 1.000 0.001
1. Recognizes authority figures	4.16 (0.91)	4.24 (0.88)	0.52	0.39		0.582	0.444	0.582	0.444
2. Properly greetings	4.63 (0.69)	3.96 (1.28)***	0.66	0.73		0.749	0.826	0.749	0.826
3. Social norms/etiquette toward adults	4.28 (0.98)	4.27 (1.04)	0.64	0.71		0.713	0.806	0.713	0.806
4. Uses proper addressing terms	4.26 (1.00)	4.42 (1.02)	0.59	0.59		0.655	0.640	0.655	0.640
5. Traditional way of speaking to adults	4.37 (0.86)	3.31 (1.51)***	0.61	0.61		0.712	0.677	0.712	0.677
F2. Respect for Adults	3.37 (0.69)	3.74 (0.71) ***	0.72	0.77	χ2 CFI RMSEA	19.53*** 0.894 0.217	7.03* 0.971 0.128	19.53*** 0.894 0.217	7.03* 0.971 0.128
1. Shouldn’t fight or talk back	3.59 (0.95)	3.90 (0.92)**	0.50	0.61		0.655	0.722	0.655	0.722
2. Treat adults with respect	4.14 (0.84)	4.47 (0.69)***	0.56	0.49		0.716	0.542	0.716	0.542
3. Children should obey	3.06 (1.00)	3.48 (1.00)***	0.62	0.71		0.720	0.868	0.720	0.868
4. Parental wishes are more important	2.69 (0.94)	3.12 (1.03)***	0.38	0.52		0.446	0.612	0.446	0.612
F3. Caring for Aging Parents	4.02 (0.63)	4.15 (0.62) *	0.81	0.77	χ2 CFI RMSEA Alpha	12.09 * 0.975 0.087	14.170* 0.960 0.109	12.09* 0.975 0.087	142.521 *** 0.831 0.141 0.87
1. Caregiving for aging parents a duty	3.98 (0.88)	4.41 (0.83)	0.65	0.52		0.760	0.629	0.760	0.672
2. Take care of my aging parents	4.38 (0.72)	4.44 (0.70)	0.66	0.72		0.757	0.842	0.757	0.735
3. Disturb to place parents in nursing home	3.98 (0.95)	4.26 (0.88)**	0.50	0.38		0.556	0.440	0.556	0.400
4. Stay close to take care of my parents	3.57 (0.85)	3.72 (1.00)	0.51	0.50		0.569	0.585	0.569	0.633
5. Important to help parents financially	4.16 (0.85)	4.27 (0.86)	0.67	0.68		0.735	0.782	0.735	0.776
F4. Centrality of the Family	3.78 (0.67)	4.14 (0.63) ***	0.68	0.66	χ2 CFI RMSEA Alpha	13.90*** 0.896 0.181	11.98** 0.904 0.177	110.187*** 0.818 0.128 0.80	11.98 ** 0.904 0.177
1. The most important above all	4.25 (0.94)	4.51 (0.81)**	0.42	0.42		0.527	0.645	0.562	0.645
2. Maintain close ties regardless	4.15 (0.87)	4.32 (0.76)†	0.49	0.53		0.596	0.740	0.651	0.740
3. Sharing home an indication of closeness	3.38 (1.01)	4.02 (0.97)***	0.53	0.53		0.696	0.602	0.600	0.602
4. Generations can share one household.	3.34 (0.95)	3.72 (1.02)***	0.41	0.30		0.544	0.339	0.539	0.339
F5. Harmony and Sacrifice	3.68 (0.61)	3.90 (0.68) **	0.70	0.80	χ2 CFI RMSEA	46.294*** 0.794 0.210	47.28*** 0.827 0.234	Combined with F4	Combined with F3
1. Harmony with family	3.91 (0.87)	4.02 (0.87)	0.60	0.60		0.779	0.707	0.603	0.635
2. Harmony with non-family	3.39 (0.88)	3.54 (0.94)	0.52	0.60		0.713	0.707	0.530	0.551
3. Sacrifice for the greater good	3.34 (1.01)	3.68 (1.01)**	0.34	0.53		0.446	0.610	0.413	0.497
4. Support family	4.09 (0.77)	4.28 (0.83)*	0.45	0.51		0.510	0.581	0.676	0.735
5. Support the extended family	3.66 (0.87)	3.98 (0.95)***	0.40	0.66		0.453	0.717	0.691	0.734
F6. Parental Expectation of Family Obligation	3.22 (0.76)	3.77 (0.86) ***	0.75	0.83	χ2 CFI RMSEA	24.38*** 0.884 0.246	19.65*** 0.928 0.243	24.38*** 0.884 0.246	19.65*** 0.928 0.243
1. Stay close after high school	2.95 (0.99)	3.61 (1.09)***	0.43	0.59		0.562	0.712	0.562	0.712
2. Help out the family	3.65 (0.99)	4.04 (0.98)***	0.62	0.68		0.688	0.704	0.688	0.704
3. Live close to help	2.90 (0.68)	3.48 (1.13)***	0.68	0.77		0.836	0.910	0.836	0.910
4. Take care of old parents	3.40 (1.03)	3.94 (1.06)***	0.46	0.59		0.547	0.638	0.547	0.638

Notes:

Some items are shortened to fit the table. Lead-in questions are mostly not included, also to fit the table. The original items including lead-in questions are available from the first author.

¹Cronbach alpha for scale

²Item-total correlations

^†p < .10

^*p < .05

^**p < .01

^***p < .001

Specifically, in the Traditional Manners and Etiquettes measure, Korean American youth reported a significantly higher endorsement than Filipino American youth in two items: “greeting adults/elders properly” and “using Korean/Filipino way of speaking to adults to show respect.” The rest of the items did not statistically differ across the two groups. All items showed overall acceptable item-total correlations, factor loadings, as well as fit indices. Although the item “To recognize and acknowledge authority figures within and outside the family” among Filipino American youth has a low item-total correlation (.39) and factor loading (.444), still around or above the cutoff criteria, the removal of this item did not improve measurement fit. In both samples, model fit tests indicated strong fit of the factor (CFI of 1.0 and RMSEA of .001).

All four individual items for the Respect for Adults measure were endorsed consistently and significantly at a higher rate among Filipino American youth than Korean American youth. The item “Parental wishes are more important than the child’s desires” was endorsed at a lower rate than other items in both groups. This item also had low item-total correlation (.38) and factor loading (.446) among Korean American youth but still around or above the cutoff criteria. The removal of this item did not improve the goodness of fit. Cronbach alphas were good (.72 for Korean Americans and .77 for Filipino Americans). CFI value was fair among Korean American youth (.894) and very good among Filipino Americans (.971). However, the RMSEA values were relatively poor (.217 for Korean Americans,. 128 for Filipino Americans).

The mean of Caring for Aging Parents was significantly higher among Filipino Americans than Korean Americans, but at the item level, statistical difference was found only in the item, “It will disturb me if I place my aging parents in a nursing home” (4.26 for Filipino Americans vs. 3.98 for Korean Americans). Although this item showed relatively low item correlation (.38) and factor loading (.44), as was the case above, it met the criteria, and the removal of this item did not change the overall fit of the scale. Model fits were overall good in both samples with a of .81 and .77 and CFI of .975 and .960, for Korean American and Filipino American youth, respectively. The model fit indicated by RMSEA was at the acceptable level (.087 for Korean Americans and .109 for Filipino Americans).

All four items of the Centrality of the Family measure were endorsed significantly higher by Filipino Americans than by Korean Americans. The item “It is acceptable that several generations of a family share one household” showed low item correlation (.30) for Korean Americans and poor factor loading (.339) for Filipino Americans. However, again, the removal of this item did not improve the measurement fit. The overall measurement fit was not desirable, as indicated by an alpha (α) of .68 and .66 as well as CFI of .896 for Korean Americans and .904 for Filipino Americans. The RMSEA test (.181 and .177, for Korean Americans and Filipino Americans, respectively) further suggested a poor fit of the measurement.

The mean of three items of Harmony and Sacrifice were statistically significantly higher among Filipino American youth, i.e., “It is important to sacrifice individual(s) for the greater good (e.g., family or group)” (3.68 vs. 3.34), “I should support my family, even if it is a big sacrifice for me” (4.28 vs. 4.09), and “A person should support members of the extended family (e.g., aunts, uncles, and in-laws) if they are in need, even if it is a big sacrifice for me” (3.98 vs. 3.66). Although Cronbach alphas were reasonable (.70 and .80), the measurement fit as a latent construct was poor in both samples (CFI of .794 and .827 and RMSEA of .210 and .234).

Filipino American youth endorsed all individual items for Parental Expectation of Family Obligation significantly higher than Korean American youth. Model fits were better for Filipino American youth (α=.83, CFI=.928) than for Korean American youth (α=.75, CFI=.884). Again, the fit of the model was poor according to RMSEA (.234 for Filipino Americans and .246 for Korean Americans).

Modifications

The fit indices of the initial multi-factor CFA model were χ²=615.938, p<.001, CFI=.821, RMSEA=.073 for Korean Americans and χ²=539.844, p<.001, CFI=.863, RMSEA=.069 for Filipino Americans. For the modification of scales, high correlations among items (>.7), significant BY statements that suggest loading an item to a different factor, significant WITH statements (i.e., high correlation among factors or items), and low factor loading (<.4) were considered. Each significant MI was ordered by the size of χ² and modification was made in that order. In each modification, multi-factor CFA model was run to examine model fits and changes in significant Mis. There were several modifications and their model fit and each modification steadily improved the model fits. The model fits of the finalized measurement are reported in this paper.

Modifications were made in each group. Among Korean Americans, Centrality of the Family and Harmony and Sacrifice were merged because of high correlation between the two (>.90). Some of the error terms of items were specified to covary (e.g., “harmony with family” and “harmony with non-family” in Harmony and Sacrifice). These modifications significantly improved the model fit (CFI=.866, RMSEA=.063). For Filipino Americans, Caring for Aging Parent and Harmony and Sacrifice were also merged because of high correlation between them (>.90). Similar to Korean American samples, the same sets of the error terms of items were specified to covary. This also improved the model fit (CFI=.890, RMSEA=.062).

A single-factor CFA for each merged scale was run to confirm their fit as a single scale. For the combined scale of Centrality of the Family and Harmony and Sacrifice for Korean Americans, CFI was .818 and RMSEA .128. Although the fit of Centrality of the Family as a single scale had a better model fit, (e.g., CFI=.896) and the fit of Harmony and Sacrifice was not good (e.g., CFI=.794), high correlation warranted merging the two. Similarly, the merged scale of Caring Aging Parents and Harmony and Sacrifice showed 0.831 of CFI and 0.141 of RMSEA. The model fit of Caring Aging Parents worsened from CFI=.960 and the model fit of Harmony and Sacrifice was fair (e.g., CFI=.827), but with their very high correlation, the two scales had to be merged.

Intercorrelations

The results of pair-wise correlations among the finalized five scales are presented in Table 3 for Korean Americans (F4 and F5 merged) and Table 4 for Filipino Americans (F3 and F5 merged). From the correlations among factors, discriminant and divergent validity of the scales were tested. Overall, the scales were significantly and positively correlated with one another with a few exceptions. Respect for Adults and Parental Expectation of Family Obligation were not significantly correlated among Filipino American youth, while Traditional Manners and Etiquette was not significantly correlated with Respect for Adults and Parental Expectation of Family Obligation. Otherwise, the remaining correlations were significant and positive, and the magnitudes of the correlations ranged from .184 to .602 but none above .850 (Campbell & Fiske, 1959), which supports both discriminant and convergent validity of the scales.

[Table 3].

Correlations among Factors for Korean Americans

Factors	F1	F2	F3	F4/F5	F6
F1 Ethnic Manners and Etiquette	--
F2 Respect for Adults	0.143	--
F3 Caring for Aging Parents	0.373***	0.380***	--
F4/F5 Centrality of the Family/Harmony and Sacrifice	0.285***	0.537***	0.529***	--
F6 Parental Expectation of Family Obligation	0.014	0.226***	0.248**	0.207**	--

Notes:

^*p < .05

^**p < .01

^***p < .001

[Table 4].

Correlations among Factors for Filipino Americans

Factors	F1	F2	F3/F5	F4	F6
F1 Ethnic Manners and Etiquette	--
F2 Respect for Adults	0.380***	--
F3/F5 Caring for Aging Parents/Harmony and Sacrifice	0.348***	0.537***	--
F4 Centrality of the Family	0.299***	0 494***	0.602***	--
F6 Parental Expectation of Family Obligation	0.256**	0.152	0.435***	0.184*	--

Notes:

^*p < .05

^**p < .01

^***p < .001

In regard to the interrelations between the finalized scales and correlates (i.e., youth outcomes and acculturation variables), summarized in Tables 5 and 6, the results show that in both groups, contrary to the expectation of the study, acculturation variables (English proficiency, American identity, and mainstream cultural practices) were not associated with familism. Conversely, enculturation, especially ethnic identity and heritage cultural practices in both groups but particularly among Filipino American youth, had extensively positive correlations with familism subdomains. In line with the expectations, Parental Expectation of Family Obligation was positively correlated with depressive symptoms in both groups. In contrast, while familism subdomains did not have significant associations with GPA and antisocial behaviors among Korean American youth, Respect for Adults and Caring for Aging Parents/Harmony & Sacrifice were negatively correlated with both depressive symptoms and antisocial behaviors among Filipino American youth. Centrality of the Family was also negatively correlated with antisocial behaviors among Filipino American youth.

[Table 5].

Correlations between Familism, Youth Outcomes and Acculturation for Korean Americans

	F1 (Manners Etiquette)	F2 (Respect for Adults)	F3 (Caring for Parents)	F4/F5 (Centrality/Harmony)	F6 (Family Obligation)
Acculturation Variables
English Proficiency	.001	−.012	.032	−.066	.118
American Identity	−.105	−.017	.003	−.024	.003
Mainstream Cultural Practices	−.042	−.012	.002	−.022	−.078
Heritage Language Proficiency	.283**	−.015	.168*	.151*	−.086
Ethnic Identity	.338**	.328**	.380**	.416**	.167*
Heritage Cultural Practices	.414**	.064	.225**	.219**	−.003
Youth Outcomes
Depressive Symptoms	−.029	−.019	−.007	.026	.161*
GPA	.121	.101	.122	.130	.020
Antisocial Behaviors	−.093	−.135	−.124	−.070	.076

Notes:

^*p < .05;

^**p < .01

^***p < .001

[Table 6].

Correlations between Familism, Youth Outcomes and Acculturation for Filipino Americans

	F1 (Manners Etiquette)	F2 (Respect for Adults)	F3/F5 (Caring/Harmony)	F4 (Centrality Of Family)	F6 (Family Obligation)
Acculturation Variables
English Proficiency	−.138	−.106	−.001	−.001	.070
American Identity	−.122	.030	.023	.118	−.048
Mainstream Cultural Practices	−.071	.020	.007	−.010	−.029
Heritage Language Proficiency	.485**	.042	.180*	.146	.144
Ethnic Identity	.485**	.235**	.300**	.326**	.180*
Heritage Cultural Practices	.556**	.283**	.307**	.268**	.215**
Youth Outcomes
Depressive Symptoms	.005	−.253**	−.203*	−.100	.183*
GPA	−.031	−.048	−.008	.077	−.024
Antisocial Behaviors	−.128	−.302**	−.360**	−.311**	−.075

Notes:

^*p < .05

^**p < .01

^***p < .001

Measurement Invariance

Factorial invariance was tested for three scales, Traditional Manners and Etiquette, Respect for Adults, and Parental Expectation of Child’s Family Obligation, that indicated configural invariance (Table 7). Traditional Manners and Etiquette and Respect for Adults did not attain metric invariance, but Parental Expectation of Family Obligation attained metric, strong, and strict invariance. Lastly, functional invariance among these scales was indicated by a similar pattern of intercorrelations as described above, except three non-significant associations described above.

[Table 7].

Factorial Invariance Tests

Model	Δχ2	Δdf	CFI	RMSEA
Traditional Manners and Etiquette
1 Configural	--	--	.999	.024
2 Metric	29.485***	4	.949	.113
3 Strong	113.919***	4	.737	.224
4 Strict	80.15***	5	.591	.244
Respect for Adults
1 Configural	--	--	1.000	.000
2 Metric	10.199*	3	.983	.083
3 Strong	6.979	3	.971	.085
4 Strict	9.558*	4	.954	.087
Parental Expectation of Family Obligation
1 Configural	--	--	.914	.334
2 Metric	2.52	3	.915	.210
3 Strong	6.084	3	.908	.173
4 Strict	4.728	4	.907	.142

Notes:

No asterisk means invariance.

^*p < .05

^**p < .01

^***p < .001

Discussion

Familism is one of the most distinctive characteristics of Asian American culture (Yee, Su, Kim, & Yancura, 2009). Specific aspects of familism, such as family obligation or filial piety, have garnered significant scholarly attention, while familism as a multidimensional construct has been more infrequently studied among Asian Americans. This study examines how two large subgroups of Asian American youth, largely U.S.-born or having immigrated at an early age, understand familism by developing and testing the psychometric properties and measurement invariance of six scales of familism. Several scales of this study showed high quality (i.e., moderate to high endorsement, solid reliability, and good content and construct validity including discriminant and convergent validity) and suggested that core traditions, in the form of familism values and behaviors, endure among both Filipino American and Korean American youth samples, despite high acculturation to the mainstream society. Consistent with previous studies (e.g., Fuligni & Masten, 2010; Fuligni & Pedersen, 2002) and similar to the pattern found among their parents’ generation (Choi, Kim, Noh, et al., 2018), Filipino American youth endorse more traditional aspects of familism than do Korean American youth. This is not surprising given the finding that Filipino American parents reinforce traditional familism beliefs and behaviors with their children at a greater rate than Korean American parents (Choi, Kim, Noh, et al., 2018); the present study reflects this difference in socialization. However, as discussed further below, there are also indications that Filipino American and Korean American youth hold a shared sense of familism that is distinct from that of their parents; differences in familism across Asian American populations may diminish over time across generations.

Psychometric Properties

Starting with an extensive literature review, focus groups, and the filter of rounds of investigative analyses and pretests, this study set out with a total of 27 items categorized into six scales of familism. All of them showed fair to excellent measurement fit as a single scale including modest to high endorsement, item-total correlation, and reliability, which are minimum requirements for a good scale. When put through more rigorous testing such as confirmatory factor analysis as a latent construct, most of the scales demonstrated fair to excellent fit of the measurement model. However, through further modifications and merges to enhance the fit, a slightly different factorial structure of familism emerged in each group. Significantly, this study finds that Traditional Manners and Etiquette, previously developed and tested for Korean Americans, worked well with Filipino Americans. Respect for Adults and Parental Expectation of Family Obligation also worked well in both groups. The measures that showed configural invariance between the two groups, such as Caring for Aging Parents, and the two measures that merged differently in two groups (i.e., Centrality of the Family/Harmony and Sacrifice for Filipino Americans and Caring for Aging Parents/Harmony and Sacrifice and Centrality of the Family for Korean Americans) also worked well with each group. These scales should be robust enough to be used, including via structural equation modeling as latent constructs, which typically requires thorough measurement testing. RMSEA is higher than desired, but it is common to use items as a scale (summed or averaged) or, in analyses modeling latent constructs, to parcel items, which likely reduces residual covariance among items and, subsequently, improves fit (Kline, 2010). No item was discarded in the process of testing, and the scales are ready for use and further enhancement.

However, the differential merging of constructs in each group needs further discussion. These scales, while clearly discrete among the parent generation (Choi, Kim, Noh, et al., 2018), may not be as discrete among the youth generation. Specifically, although all scales remain distinct from one another, Harmony and Sacrifice was merged with Centrality of the Family among Korean American youth, and with Caring for Aging Parents for Filipino American youth. Furthermore, some of the finalized subdomains, especially those related to familial obligations, were highly correlated, suggesting additional merging. However, when merged, the measurement fit was poor, as some of the items did not load well to the omnibus factor. This finding underscores the dynamic nature of familism, and further points to the acculturative process at work in immigrant families. Even as immigrant children uphold the familism values of their parents, subsequent studies may find that subdomains of familism continue to merge across subgroups of immigrant youth.

It remains unclear how high endorsement of core values of familism would serve Asian American youth when they leave home and in the long run. Asian American young adults may experience more difficulty in navigating new social settings because of a lack of independent and explorative experiences during adolescence. Likewise, the cultural pressures of familism may complicate the social expectation to achieve autonomous adulthood (Shibusawa, 2008). Asian collectivism via familism expects grownup children to maintain interdependence and fulfill their family obligations (Kagitîbasi, 2007). Western individualism, in contrast, emphasizes individual goals and views parent-child tension as generative and needed for autonomy (Trommsdorff, 2006). Additional, especially longitudinal, research is needed to understand how the mismatch between social and familial norms is related to adjustment among Asian American young adults.

Ethnic Group Similarities and Differences

The intercorrelations among familism subdomains were fairly high in both Filipino American and Korean American youth (with a couple of exceptions) and slightly more extensively among Filipino American youth. With consistently higher endorsement in most subdomains of familism among Filipino American youth, it could be concluded that Filipino American parents are more proactive in instilling familism values in their children than Korean American parents, which were indicated in the parents of this study’s samples (Choi, Kim, Noh, et al, 2018). It is plausible that Filipino American parents counter their high rates of acculturation in domains such as language and residence with greater adherence to traditional values and socialization of their children (B. S. K. Kim et al., 2001). Alternatively, Filipino American parents, who have a higher household income and are more likely than Korean American parents to be employed in professional occupations in the mainstream society, may be asked more often to assist their struggling extended families both in the U.S. and in the Philippines, which may enhance preservation of familism in the family. In each case, it is important to keep in mind that the difference in scores is relative; both Filipino American and Korean American families scored highly on each measure of familism. Moreover, with an exception of Traditional Manners and Etiquette, subdomains of familism were significantly correlated with one another among Korean American youth. Interestingly, Parental Expectation of Family Obligation, which was not correlated with other aspects of familism among Korean American parent samples, was correlated with Caring for Aging Parents, Centrality of the Family/Harmony and Sacrifice among Korean American youth samples. This may be an indication that even if Korean American parents do not explicitly instill traditional core values of helping out the family and caring for aging parents, children may still be socialized to internalize these values, which is consonant with what Korean American adolescents shared in focus groups preceding the present study. These youth revealed that, even if parents may not expressly communicate their expectations, they are well aware of their parents’ values and implicit desire that they also adopt these values; many youth stated that they know their parents would be pleased if they, as children, have internalized their parents’ values.

As found among Korean American parents (Choi, Kim, Noh, et al., 2018), Traditional Manners and Etiquette among Korean American youth was not correlated with Parental Expectation of Family Obligation and even Respect for Adults that such manners and etiquette are designed to embody, confirming that traditional manners and etiquette among Korean Americans are more akin to behavioral codes than values. However, unlike among their parents, Traditional Manners and Etiquette was correlated with the remaining subdomains of familism among Korean American youth. It could be theorized that, while treated as behavioral codes, Korean manners and etiquette help maintain core values, such as caring for aging parents, expressing the centrality of family, and reinforcing the importance of harmony and sacrifice. Thus, Korean American parents may not need to expressly communicate familism values. Korean behaviors and manners are organized to symbolize family hierarchy, and the plain act of engaging in these behaviors and manners may serve to transmit and reinforce core Korean familism values. For example, a gesture of bowing to adults or using honorifics may instill respect for adults in those children that practice those behaviors, even if respect for adults is not a value that is specifically spoken about and reinforced by parents. Future research may examine the ways in which these traditional manners and etiquette can actually help maintain core values.

In contrast to Korean American youth, Traditional Manners and Etiquette among Filipino American youth was integral to familism, which was also the case with Filipino American parents (Choi, Kim, Noh, et al., 2018). Still, the differences between Korean American and Filipino American families were more evident among the parent generation than the youth generation. This is true in regard to measurement invariance as well. Notably, Parental Expectation of Family Obligation attained configural, metric, strong, and strict invariance among Korean American and Filipino American youth, although as a scale, it was more endorsed, and had higher reliability and better fit among Filipino American youth. The limited invariance of the measures found in this study cautions against liberally making comparative analyses. In other words, constructs like Traditional Manners and Etiquette are valid to use in each group but what they mean in each group is likely to differ. Similarly, the concept of harmony and sacrifice is associated disparately with unique aspects of familism in each group. However, Parental Expectation of Family Obligation has an equivalent meaning across Filipino American and Korean American youth, and its predictive relations to youth development can be compared with conceptual and empirical confidence.

Correlates: Acculturation and Youth Outcomes

The results that none of the youth acculturation indicators, in contrast to enculturation, were associated with the finalized familism subdomains are noteworthy. Immigrants, especially those who immigrated as an adult, can be resistant to change in core values like familism, and acculturation to the mainstream culture may not make a significant impact on their core values (Choi, Kim, Lee, et al., 2018). Unsurprisingly, literature on intergenerational cultural conflict among immigrant families has conceptualized that youth acculturate at a faster speed and adopt mainstream values, leading to culture clash (Lee, Choe, Kim, & Ngo, 2000). More recently, however, scholars have challenged such conceptualizations as too simplistic and demonstrated that parent-child conflict and cultural gaps among immigrant and cultural minority families are far more complex (Suh, Choi, Park, & Yoon, 2020; Telzer, 2010). The present study adds to this complexity by providing evidence that various aspects of familism, as a core Asian American value, may persevere among the second generation, despite their high level of acculturation. The positive relations between familism and enculturation indicators, notably ethnic identity in both groups, also stand out, suggesting that they may go hand in hand, solidifying enculturation processes among second-generation Asian American youth.

The correlations between the finalized scales and youth outcomes demonstrate how distinctive subdomains of familism may have differential impacts on youth outcomes, e.g., parental expectation of family obligation but no other subdomains may have a negative impact on mental distress. Interestingly, there are subdomains of familism that are potentially protective among Filipino American youth in both psychological health and externalizing problem behaviors. These results together highlight the importance of more nuanced and domain-specific understandings of familism and distinct effects of each subdomain on youth development. The varying, sometimes divergent, associations may be related to qualitative differences between youth endorsement of values and youth perceptions of how much they are expected to fulfill these values. In fact, youth endorsement of familism such as caring for aging parents, respect for adults, and making sacrifices for the family, may be an indication of positive parent-child relationship quality, thus yielding favorable influence on youth outcomes, specifically among Filipino American youth. Conversely, when it is perceived as an external expectation, i.e., parental expectation of family obligation, it may feel burdensome and negatively relate to psychological health. Moreover, when a child (or adult child) actually needs to provide support to parents, those circumstances may entail a myriad of factors that can complicate parent-child relationships and the child’s mental health (e.g., economic hardship among parents or relatives). Indeed, among Filipino American parent samples of the study, the centrality of family values did not converge with the centrality of family behaviors (e.g., providing financial assistance to family and relatives) (Choi, Kim, Noh, et al., 2018).

Conclusion

Heritage cultural values of non-Western families in the U.S. may dwindle in the process of acculturation, notably among succeeding generations of immigrants (Alba et al., 2000). This study, however, provides evidence that core family values such as familism are upheld among the second generation of Asian Americans and that Asian American youth are significantly socialized to the values and behaviors of familism. Youth sample of the study was mostly bom in the U.S. or immigrated at a young age and had a high level of linguistic assimilation. Both Filipino American and Korean American youth in the focus groups of the study averred that their parents did not explicitly impart traditional core values, such as an obligation to care for the family or aging parents. Yet, these youth are well aware of and even continue to endorse familism. This finding may bode well for child-family relationships, to the extent that discrepancies in familism values can lead to intergenerational conflict. On the other hand, if youth’s experience of familism is characterized by an obligation to maintain harmony at the cost of their individual desires, and fulfilling family obligations is the main conduit of familism for youth, it is not clear whether familism is supportive of their wellbeing. Filipino American and Korean American youth are particularly vulnerable to mental distress (Choi, Park, Lee, Takeuchi, et al., 2020). Upholding traditional values may have mixed effects on youth mental health (Hahm et al., 2017) and additional multivariate, subgroup-specific, and longitudinal analyses should pinpoint aspects of familism that may be beneficial or harmful, further examine why, and identify how to strengthen beneficial aspects of familism to maximize positive youth development.