Recent estimates by demographers suggest that while 11% of the U.S. population is comprised of immigrants, over 20% of children under the age of 18 are in immigrant families (Capps et al. 2005). By the year 2020, it is expected that over 30% of children in the U.S. will have foreign-born parents. Given these dramatic and rapid demographic trends, there is a pressing need for research on both the normative and possibly unique aspects of development for youth in immigrant families. While such efforts are increasing, the majority of research focuses on adolescents in families from Latin America and Asia. Research examining family relationships and associated outcomes is necessary for all immigrant groups, including those that have traditionally been less prominent within social science research. The main goal of our study, consequently, is to focus on family dynamics and youth outcomes for Armenian immigrant adolescents, an ethnic minority group that has been largely understudied.

For children in immigrant families, the period of adolescence can be a difficult transition to negotiate. There is often encouragement for most American adolescents to individuate from parents by engaging in more independent decision making, developing emotional autonomy from parents, and increasing focus on individual success within academic and occupational pursuits (Steinberg and Silk 2002; Steinberg and Silverberg 1986). Such emphases are frequently in contrast to collectivistic socialization practices in many immigrant households where adolescents are encouraged to value interdependence, group harmony, and activities that benefit the greater community (Triandis 1989). As a result, adolescents from immigrant families must balance traditional roles and responsibilities in the family with competing pressures from American society to value individuality, self-expression, and independence from parents (Phinney et al. 2000). Such contradictory cultural influences are likely demonstrated in adolescent attitudes towards familism.

Conceptualized as the degree to which adolescents feel an obligation to support and assist family members, place family well-being and interests above those of the self, and maintain respect for parental desires and expectations, familism typically is characterized as a cultural feature of collectivism (Fuligni et al. 1999; Peterson et al. 2004b; Phinney et al. 2000). Research conducted both cross-culturally, and among ethnic minority adolescents residing in the U.S., suggests that family members from collectivistic cultures strongly encourage youth to emphasize interdependence and to place family needs before their own. As a result, adolescent attitudes in reference to familism may serve as an indication of how well youth retain cultural value systems taught to them by parents. Research completed by Fuligni et al. suggests that youth from immigrant families report stronger attitudes towards family obligation compared to their European American counterparts and that these young people are able to retain cultural values oriented towards collectivism even as they adjust to life in an individualistic society.

The concept of familism is particularly salient for Armenian immigrant families, as they tend to maintain close contact with immediate and extended family members on a regular basis, place emphasis on upholding family honor, and maintain that family is an important aspect of daily life (Bakalian 1993; Dagirmanjian 1996). Armenian immigrants, residing in the U.S. as a result of war, political unrest, and forced deportations, began arriving in the U.S. at the end of the 19th century, with a large influx arriving in the U.S. between 1900 and 1914 (Bakalian 1993). Although research on Armenian Americans is limited, available research suggests that Armenian adolescents may hold some of the highest orientations towards familism compared to other ethnic minority groups (Phinney et al. 2000). Consequently, Armenian American youth who endorse familism are considered to be retaining the collectivistic orientation that they are socialized to hold. To date, however, research has not considered how maintaining such a cultural orientation influences adolescent development and parent–adolescent relationships for Armenians coming of age in the U.S.

Previous research with diverse samples of non-Armenian ethnic minority adolescents suggests positive associations between familistic attitudes and youth educational aspirations, time spent studying, self-esteem, and emotional well-being (Fuligni et al. 1999; Fuligni and Pedersen 2002). Additionally, adolescents who identify more strongly with the culture of their family are more likely to have higher academic motivation (Fuligni et al. 2005). Although these studies suggest that high levels of familism might have a positive influence on the development of adolescents in immigrant families, previous research also suggests that immigrant adolescents who adopt a bicultural strategy to “fitting in” to American society are more likely to experience positive outcomes such as higher academic competence, positive peer interactions, adaptive achievement styles, lower stress levels, and decreased feelings of loneliness (Berry et al. 1987; Coatsworth et al. 2005; Gomez and Fassinger 1994; Suarez et al. 1997). In reference to familism, therefore, adolescent immigrants who ascribe to moderate levels of familism might be best prepared to deal with the challenges of adolescence. Put another way, the association between familism and youth outcomes might be best described as curvilinear. Support for this argument was provided by Fuligni and colleagues (1999) who reported that adolescent grades were highest at moderate levels of familism for Asian American, Latin American, and European American youth. Based upon this previous research, we hypothesize a positive direct association between familism and both self-esteem and academic motivation, but also examine potential curvilinear effects of familism with these outcomes for Armenian American youth.

Adolescent attitudes towards familism may also influence the nature of parent–adolescent relationships. According to Fuligni et al. (1999), immigrant adolescents who embrace familism (providing assistance to parents, respecting parental authority, and feeling a sense of obligation to parents) reported greater feelings of cohesion in the family and discussed issues more frequently with parents. Previous research also suggests that adolescents whose attitudes run contrary to parental expectations for collectivistic attitudes more often experience conflict with parents, reduced respect for parental authority, and difficulties associated with identity and self esteem (Farver et al. 2002; Fuligni et al. 2005; LaFromboise et al. 1993; Verkuyten 2003). Taken together, these results suggest that immigrant adolescents who embrace familism, possibly as a means to remain consistent with parental value orientations, are more likely to experience closeness with parents, may have increased respect for parental authority, and may help reduce negative parent–adolescent interactions. Based on this previous research, we hypothesize that adolescents who endorse high valuation of familism are more likely to report greater conformity to parental wishes, respect for parental authority, and increased parental knowledge of youth activities and whereabouts.

Given that familism is expected to influence the manner in which adolescents relate to their parents, and that parent–adolescent relationships are expected to influence adolescent outcomes (Bush 2000; Bush et al. 2004; Peterson et al. 2004a), we also evaluate the potential for parental conformity, authority, and knowledge to act as mediators in the association between familism and youth outcomes of self-esteem and academic motivation. Parent–adolescent interactions consistent with Armenian cultural upbringing might provide partial explanations for how familism is associated with developmental youth outcomes. Thus, we propose that immigrant adolescents who report higher levels of familism are more likely to experience positive outcomes, in part, due to experiencing higher levels of conformity, authority, and knowledge in parent–adolescent interactions.

Previous research with immigrant youth suggests that parents place greater constraints on adolescent daughters compared to sons due to concerns over girls’ exposure to American peer culture (Suárez-Orozco and Suárez-Orozco 2001). Although minimal research exists concerning gender differences in Armenian American parent–adolescent relationships, Bakalian (1993) and Dagirmanjian (1996) contend that Armenian immigrant parents are more likely to monitor adolescent daughters and expect girls to be more obedient towards parental desires. Given that associations between study variables might differ for sons versus daughters in Armenian American families, we examine the potential for gender to moderate associations between parental conformity, authority, and knowledge with adolescent outcomes.

The purpose of our research is to examine how adolescent attitudes regarding familism influence parent–child relationships and developmental outcomes for youth in Armenian immigrant families. Our study contributes to the literature by: (1) studying Armenian Americans, an immigrant group that has rarely been the focus of empirical research; (2) considering a “child effects” model where characteristics of the young person are hypothesized to influence parent–adolescent relationships, self-esteem, and academic motivation; (3) employing a within group strategy to focus on the diversity of experiences in reference to collectivistic attitudes in a specific cultural group; and (4) considering direct, indirect, and curvilinear influences of familism.

Method

Procedures

Students at one ethnically diverse high school in the Greater Los Angeles area provided self-report data for our study. Written permission was obtained from school administration to collect data in a required course. The research process was explained to students, adolescent and parent consent forms were distributed, and students were given an opportunity to ask questions about the project. Students with completed parent and adolescent consent forms were allowed to participate in the study and were given extra credit for participation. Questionnaires were distributed in class, and students were given 1 week to return the completed questionnaires to their teacher. Although most students completed the questionnaire in class on the day of distribution, some students took the questionnaire home for completion. During the initial visit, our research team emphasized the importance of providing honest answers to survey questions and discussed issues pertaining to confidentiality of student responses. Our research group returned after 1 week to pick up completed questionnaires from teachers.

Participants

A total of 112 Armenian American adolescents returned the study questionnaire. However, 15 adolescents did not complete the entire survey, and thus were excluded from final analyses. Demographic variables (e.g., language spoken in the home, age, grade level, gender, youth religious preference, parent education level, adolescent and parental birth country, length of time living in the U.S.) were examined to determine if excluded youth differed from the remaining 97 Armenian American adolescents. Results demonstrated no demographic differences between groups, suggesting that exclusion of participants due to missing data did not increase the likelihood for biased results. Missing values analyses were then completed on the remaining sample of 97 Armenian American adolescents. Results from these analyses demonstrated an average of 3.76% missing data across all study variables. Summary variables were created using all available data for each respondent.

The final sample consisted of 97 Armenian American adolescents. 59% of the sample was female, the mean age was 14.87 (SD = .88), and 65% of adolescents reported that their mothers have some college education (58% for fathers). The majority (67%) of the adolescents were born in Armenia (15% were born in Iran, and 12% born in the U.S.), and 85% reported that Armenian is the language most frequently spoken in the home. Among those adolescents born outside of the U.S., the median years living in the United States was 7.5 years (range 1–18; SD = 4.33). Parents of youth participants had been living in the U.S. for an average of 8.96 years (range 1–30; SD = 5.72) for mothers and 8.97 years (range 1–22; SD = 5.49) for fathers.

Measures

Familism

The 5-item Bardis Familism Scale (Bardis 1959) assessed the extent to which adolescents are oriented towards family values and loyalty (e.g., “Family responsibilities should be more important than my career plans in the future”). Response options ranged from 1 (strongly disagree) to 4 (strongly agree). Previous research with non-Armenian samples has demonstrated adequate reliability of this scale, with Cronbach’s alpha ranging from .61 to .80 (Bush et al. 2004; Peterson et al. 1999). Cronbach’s alpha for the current sample was .67.

Parent–Adolescent Relationship

Three scales from the Parent Behavior Measure (PBM; Peterson et al. 2004a) assessed the extent to which adolescents demonstrated collectivistic orientations in reference to conformity to parental expectations, respect for parental authority, and parental knowledge of adolescent activities. The conformity to parental expectations measure consisted of nine items (e.g., “If this parent did not want me to go to a particular movie, then I believe that I would not go”). Respect for parental authority was measured with nine items (e.g., “This parent has a right to influence my decisions about the friends I choose”). Parental knowledge of adolescent activities was measured using five items (e.g., “I tell this parent who I am going to be with when I go out”). Adolescents completed items to assess relationships with mothers and fathers separately. Response options for all items ranged from 1 (strongly disagree) to 4 (strongly agree).

Previous research using the PBM with non-Armenian samples has demonstrated adequate reliability for all subscales. Cronbach’s alphas ranging from .76 to .80 have been reported for parental authority, and coefficients ranging from .74 to .80 have been reported for parental conformity (Peterson et al. 1999; Peterson et al. 1985). Previous research using the parental knowledge scale have reported Cronbach’s alphas ranging from .83 to .89 (Ingoldsby et al. 2003). For the current sample, all measures demonstrated a high degree of internal consistency, evidenced by Cronbach’s alpha coefficients greater than .80. Strong correlations between reports of conformity to mothers and fathers (r = .94, p < .05), respect for authority for mothers and fathers (r = .87, p < .05), and maternal and paternal knowledge (r = .74, p < .05) were found. Consequently, individual scores were averaged for ratings of adolescent conformity, respect for authority, and disclosure of free-time activities across mothers and fathers.

Self-Esteem

Global self-esteem was assessed with the 10-item Rosenberg Self-Esteem Scale (Rosenberg 1979). Preliminary analyses suggested that the items were best represented by two constructs, consistent with previous findings using the same measure of self-esteem (Owens 1994). One dimension of self-esteem centered on the items representing negative feelings towards the self (“I certainly feel useless at times”) and one dimension represented positive valuations of self (“I feel that I have a number of good qualities”). Consequently, two summary variables were created by averaging the five negative valence items (labeled self-derogation) and the five positive valence items (labeled self-esteem) separately. Response options for all items ranged from 1 (strongly disagree) to 4 (strongly agree). Previous research using this scale has reported Cronbach’s alpha coefficients ranging from .61 to .90 (Schmitt and Allik 2005; Whiteside-Mansell and Corwyn 2003). Cronbach’s alpha for the current sample was .74 for self-derogation and .83 for self-esteem.

Academic Motivation

Adolescents responded to five items assessing their motivation for academic achievement (e.g., “I try hard in school”; Plunkett and Bámaca-Gómez 2003). Response options for each item ranged from 1 (strongly disagree) to 4 (strongly agree). Previous research with non-Armenian samples has demonstrated adequate reliability of this measure with Cronbach’s alpha coefficients ranging from .71 to .85 (Ingoldsby et al. 2003; Plunkett and Bámaca-Gómez 2003). Cronbach’s alpha was .73 for this sample.

Analytic Strategy

Analyses related to missing values, internal consistency, sample means, standard deviations, and correlations were conducted using SPSS. Research questions examining direct paths, mediating models, and potential moderating effects were all conducted using structural equation modeling (AMOS 6). Due to the relatively small sample size for our study, measurement models for each variable were examined in preliminary analyses. Confirmatory factor analytic methods were used to examine factor structures for all study variables. Items with factor loadings less than .30 were excluded from subsequent analyses. Summary variables were constructed by taking the average of all completed items in each measure. Manifest indicators were then used for hypothesis testing in the structural model (see Fig. 1). Model fit was assessed based on the goodness-of-fit index (GFI), the comparative fit index (CFI), the normed fit index (NFI), and the root mean square error of approximation (RMSEA). Models with a GFI, CFI, and NFI statistic close to .95 suggested an adequate model fit (Hu and Bentler 1999). Models with an RMSEA statistic less than .05 suggest a good model fit, and statistics between .06 and .08 suggest an acceptable model fit (Browne and Cudeck 1993). To assess whether associations between familism and adolescent outcomes were curvilinear a quadratic familism term was also included.

Fig. 1
figure 1

Standardized path coefficients for direct and indirect effects. GFI = .98, CFI = .99, NFI = .97, RMSEA = .06. * p < .05, ** p < .01, *** p < .001

Full size image

While no definitive standard for minimum sample size exists when using SEM, typical recommendations suggest either a minimum of 100–200 cases, or 5–10 cases per parameter estimate, or 10–20 cases per observed variable in the model (Quintana and Maxwell 1999). By any of these standards, the sample size for our study is slightly smaller than typically recommended. The most important issue regarding sample size, however, is whether or not this sample has adequate statistical power to detect poor fit in the model. As a means to address this issue directly, supplemental analyses were conducted using the Satorra and Saris (1985) method. Results suggested that power to detect a reduction in model fit ranged from .90 to .99 for path coefficients that were moderate in size (and that were found statistically significant). That is, if statistically significant paths were removed from this model, this sample has ample statistical power to find that those changes would adversely affect model fit. As a result, while this sample is small, there is adequate statistical power for SEM purposes.

Mediation was examined according to the structural equation modeling approach described by Holmbeck (1997). To demonstrate mediation, it was necessary to demonstrate significant associations between familism and outcome variables; between familism and parent–adolescent interactions of conformity, authority, and knowledge; and between parental conformity, authority, and knowledge and the outcome variables. Subsequently, the baseline model was compared to a model where paths from familism to outcome variables were constrained to zero. According to Holmbeck, mediation is present if the constrained model demonstrates a better fit compared to the baseline model. A significant change in chi-square across the two models demonstrates that inclusion of conformity, authority, and knowledge in the model provides potential explanations (mediation effects) for associations between familism and outcome variables. The Sobel test was also used to indicate whether or not indirect associations between independent and outcome variables are statistically significant (Baron and Kenny 1986).

Multi-group analyses were conducted to examine potential moderator effects of adolescent gender (Byrne 2001). Using the baseline model, two groups were created (males and females) to examine whether associations between parent–adolescent relationships and outcome variables differ for male versus female adolescents. Two models were compared: (1) a fully constrained model, where all parameters are constrained to be equal for boys and girls, and (2) a model where paths from parental conformity, authority, and knowledge to outcome variables were freely estimated. A significant change in chi-square across the two models provided evidence of gender differences and path coefficients were examined to determine the specific location of gender differences.

Results

Descriptive statistics and correlations for all study variables are reported in Table 1. An examination of significant correlations revealed that familism was positively associated with increased conformity, respect for authority, and parental knowledge. These results suggest that adolescents who embrace the values of familism also demonstrate a more collectivistic orientation towards parent–adolescent relationships by conforming more to parents’ wishes, respecting parental authority, and disclosing activities to parents. Correlations also suggested that adolescents with greater endorsement of familistic attitudes reported higher self-esteem and greater effort in school. Parental conformity was also correlated significantly with self-derogation, parental knowledge was correlated with self-esteem, and parental authority and knowledge were correlated with academic motivation. Results from correlation analyses suggested the need to examine these relationships further.

Table 1 Correlations and descriptive statistics (n = 97)
Full size table

The baseline model demonstrated an acceptable fit to the data (GFI = .98, CFI = .99, NFI = .97, RMSEA = .06). In reference to main effects, adolescents reporting higher levels of familism reported higher conformity to parental wishes (β = .35, p < .001), greater respect for parental authority (β = .54, p < .001), and greater disclosure to parents (β = .42, p < .001). In addition to being associated with increased deference to parents, familism was positively associated with self-esteem (β = .28, p < .05) and negatively associated with self-derogation (β = −.22, p < .05). Familism was positively associated with academic motivation as well, but only at a marginally significant level (β = .18, p = .10). Taken together, these results suggest that familism is a positive aspect of development for these Armenian American youth as it positively relates to more openness to parental influence, and may lead to greater self-esteem and lowered self-derogation.

Examining associations between parent–adolescent interactions and outcome variables, parental knowledge of adolescent activities was positively related to academic motivation (β = .32, p < .01) and negatively associated with self-derogation (β = −.26, p < .01). Adolescents who disclosed information concerning their whereabouts to parents on a consistent basis were more likely to remain motivated in their academic pursuits and less likely to experience self-derogation. Contrary to expectations, conformity to parental expectations was positively related to self-derogation (β = .50, p < .001), suggesting that Armenian American youth who engaged in higher levels of parental conformity were more likely to experience higher levels of self-derogation.

Main effects demonstrated significant associations among familism, parental conformity, and parental knowledge with self-derogation. Initial criteria of significant relationships among these predictor and outcome variables were satisfied, suggesting the possibility that conformity and knowledge would mediate the association between familism and self-derogation (Baron and Kenny 1986; Holmbeck 1997). Comparing the baseline model to a model with constraints on the path from familism to self-derogation produced a significant change in chi-square [Δχ 2 (1, N = 97) = 4.25, p < .05]. However, examination of fit statistics for both models suggested that the baseline model fit the underlying data better than the constrained model. Reducing the path between familism and self-derogation to nonsignificance (by constraining the path to zero) does not improve the fit of the model, suggesting that parental conformity and knowledge do not act as mediators in the current study. Significant Sobel test statistics provided evidence, however, for indirect effects of the association between familism and self-derogation through parental conformity (z = 2.90, p < .01) and knowledge (z = −2.33, p < .05). Results from Sobel test statistics also provided evidence for indirect effects in the association between familism and academic motivation through parental knowledge (z = 2.61, p < .01). Taken together, although complete mediation is not present with these data, results suggest that familism might be indirectly associated with self-derogation through parental conformity and knowledge. Additionally, although familism was not associated directly with academic motivation, results suggest an indirect association among familism, parental knowledge, and academic motivation for Armenian American youth.

Multi-group analyses comparing competing models for boys and girls (model with paths freely estimated between mediators and outcome variables versus a model with paths constrained to equality) revealed a statistically significant difference in chi-square between the two models [Δχ2 (15, N = 97) = 25.70, p < .05]. Results from comparison of specific structural parameters revealed that only one path in the model varied by adolescent gender: the association between respect for parental authority and academic motivation. Although respect for parental authority was not related to any outcomes in the baseline model, the multi-group analysis suggests that this association is strong and positive for boys (B = .86, p < .05) and nonsignificant for girls (B = −.06, p > .10). Armenian American boys who reported high respect for parental authority were more likely to remain motivated regarding academic pursuits, but parental authority was not a salient predictor of academic motivation for Armenian American girls with the current data. Given these findings, indirect effects of associations between familism, parental authority, and academic motivation were examined for Armenian American boys in the current sample. Significant Sobel test statistics provided evidence for indirect effects of the association between familism and academic motivation through parental authority for boys (z = 2.61, p < .01).

Examination of potential curvilinear effects revealed that the associations between familism and self-esteem variables are best described as linear only, while a statistically significant quadratic coefficient from familism to academic motivation suggested that the association between familism and academic motivation was nonlinear. To clarify this effect we examined mean levels of academic motivation at three levels of familism (based on tertile splits) and found a significant difference in motivation between low (M = 3.01) and medium (M = 3.44) levels of familism, but no difference between medium and high (M = 3.43) levels of familism. This suggests that Armenian immigrant youth who reported at least moderate levels of familism were more likely to remain motivated with respect to academic endeavors compared to youth who reported low levels of familism, but that there was no additional benefit with respect to academic motivation for youth who reported the highest levels of familism. Nonsignificant coefficients on paths from the quadratic to self-derogation and self-esteem suggest that the nature of those associations is linear rather than curvilinear.

Discussion

The results of our study contribute to the knowledge base on adolescents in immigrant families by demonstrating that: (1) within-group variation in familism is a predictor of developmental outcomes for youth in immigrant families, (2) familism may influence adolescent outcomes both directly and indirectly by altering parent–adolescent relationships, (3) there may be gender differences in the manner by which respect for parental authority influences school outcomes, and (4) the association between familism and youth outcome variables might be nonlinear. In addition, our study provides one of the few empirical analyses of a sample of Armenian American adolescents.

Although endorsing collectivistic cultural beliefs might be a challenge for immigrant youth residing in individualistic U.S. society, our findings suggest that maintaining a collectivistic cultural orientation is adaptive for Armenian immigrant youth and increases the likelihood for positive psychosocial outcomes. One potential explanation for this finding is that adolescents who ascribe to a collectivistic cultural orientation are endorsing cultural attitudes consistent with family values and beliefs. Given the close knit nature of most immigrant families, youth who remain loyal to collectivistic beliefs and customs are less likely to feel at odds with the family socialization environment potentially allowing youth to feel more comfortable with their developing identity and self-esteem. As a result, Armenian immigrant youth who report higher levels of familism are more likely to experience higher self-esteem and lower self-derogation.

Results also suggest that parent–adolescent relationships likely play a key role in how adolescent perceptions of familism are associated with youth outcomes. According to our results, immigrant youth who ascribe to collectivistic cultural orientations are more likely to experience positive outcomes, such as higher academic motivation and lower self-derogation, when they conform to parent expectations, provide parents with knowledge of youth whereabouts, and defer to parental authority on a consistent basis. Adolescents in immigrant families who engage in these specific behaviors might feel they are contributing towards the goal of upholding family honor and maintaining group harmony. These notions of solidarity might then help to partially explain how familism is associated with academic motivation and self-derogation. However, results also suggest that different aspects of immigrant parent–adolescent relationships play distinct roles in explaining associations between familism and youth outcomes. Parental conformity appears important for psychosocial outcomes, but parental knowledge seems important for both psychosocial and academic outcomes. High levels of parental knowledge in immigrant parent–adolescent relationships might provide some youth with a sense of coherence between their cultural orientation and behaviors, thereby contributing to more positive feelings about themselves. These positive self-images might then contribute to positive developmental outcomes such as lower self-derogation and higher academic motivation. Parental conformity, however, appears to be more closely associated with individual self-concept.

Contrary to hypotheses, results demonstrated that increased levels of conformity predicted higher self-derogation, suggesting that Armenian American adolescents who alter their behaviors to suit parental expectations experience increased negative feelings toward the self. These findings lead to a somewhat paradoxical conclusion: that dimensions of familism might both positively and negatively influence negative aspects of self-esteem. Perhaps these results demonstrate an ambivalence felt by adolescents in immigrant families as they simultaneously try to respect and adhere to the cultural traditions of the family while learning how to function in school and peer groups that emphasize individual goals and desires. Armenian American adolescents who report higher levels of conformity in parent–adolescent relationships might feel at odds with the majority culture and thus develop more negative feelings towards the self. Furthermore, these results suggest that for this sample of Armenian American youth these measures of familism and conformity may represent distinct and divergent elements of collectivism. Put a different way, Armenian American adolescents may benefit psychologically from keeping identities that fit with their cultural attitudes towards familism. But the behaviors of acting in a conforming way contradicts the norms and values of the American peer group which may result in youth feeling out of place and negatively evaluating themselves.

Only the association between parental authority and academic motivation produced a significant moderator effect in reference to adolescent gender. While respect for the authority of parents may increase effort and motivation in school for Armenian boys, respect for parental authority was unrelated to academic motivation for the girls in this sample. One possible interpretation is that Armenian adolescent girls are more strictly controlled by parents and thus might acknowledge parental authority more readily. As a result, parental authority might not be a salient motivator for academic outcomes among girls. Boys, on the other hand, are argued to have more freedom to make their own decisions and might be less accustomed to high levels of parental authority (Suárez-Orozco and Suárez-Orozco 2001). Although Bakalian (1993) and Dagirmanjian (1996) suggest that gender differences with respect to parental control and youth outcomes might exist for Armenian American adolescents, for the most part these differences were not seen in our study.

With respect to the association between familism and academic motivation, Armenian American youth who reported moderate levels of familism were more likely to demonstrate higher academic motivation compared to youth who reported low levels of familism. Although these findings require replication, it appears important for Armenian immigrant youth to ascribe to at least moderate levels of familism to experience higher academic motivation. However, high levels of familism do not afford these adolescents any additional benefit with respect to academic motivation. This suggests that immigrant youth struggling to balance socialization from collectivistic home environments and individualistic peer cultures can maintain moderate attitudes towards familism and still experience adaptive academic outcomes. Although complete endorsement of familism might afford immigrant youth greater opportunities for higher self-esteem and lower self-derogation, moderate levels of familism seem ample to ensure positive academic outcomes.

The results of our study are limited due to a relatively small sample size and cross-sectional research design. Future research should examine prospective associations among familism, parent–adolescent relationships, and adolescent outcomes as youth in immigrant families move from late childhood into adolescence. This is the developmental timeframe when youth begin spending less time with parents and more time with peers. As a result, studies with prospective research designs would more accurately assess how attitudes towards familism might change and how familism influences parent–adolescent relationships across this transition. Although potentially difficult to obtain, larger samples of Armenian American youth could be used to assess issues related to generational status in reference to attitudes regarding familism, as well as to examine the potential for gender differences in youth perceptions of familism and associated outcome variables. The smaller sample size used in our study is likely responsible for the lack of gender differences that were found. Hence, results suggesting minimal gender differences in associations between familism, parent–adolescent relationships, and youth outcomes should be interpreted with caution. Obtaining a larger sample of this ethnic minority group will allow future researchers to elucidate more specifically how gender impacts family dynamics and youth outcomes in Armenian immigrant families.

Our study contributes to the growing literature on adolescent development in immigrant families by demonstrating that adolescent attitudes towards familism relate to parent–adolescent relationships and adolescent outcomes. All elements of familism may not be the same, however, as we found that familism and conformity to parents differed in their associations with self-derogation. This latter finding may represent some of the ambivalence that adolescents in immigrant families feel as they attempt to simultaneously retain values of the family and fit into American society. Examination of individual, family, and school influences pertaining to cultural socialization, such as youth ethnic identity, family ethnic socialization, and peer group influences, might provide additional information on these associations in future research. Additional research on these topics can provide valuable information for immigrant families negotiating the challenges of acculturation into an individualistic society.