Introduction

For most families living in the United States, youth’s attainment of behavioral autonomy and independence is considered a central developmental task during the adolescent years with important implications for parent–youth interactions (Steinberg 1999). Parenting processes leading to youth’s behavioral autonomy and independence likely are unique for Latino families due to Latino cultural values emphasizing family obligation and honor and parental respect and authority (Fuligni and Yoshikawa 2003; Greenfield et al. 2006). Heterogeneity in cultural orientation, however, may differentiate Latino parents in ways that have important implications for parent–adolescent interactions around autonomy and independence (Zimmer-Gembeck and Collins 2003). Some Latino parents may ascribe to traditional Latino values and practices (“separated” or “enculturated” Latinos), others may adopt US cultural orientations (“assimilated” or “US acculturated” Latinos), and still others may endorse various combinations of US and Latino cultures (“integrated” or “bicultural” Latinos; Berry 1990; Guo et al. 2012). Regardless of their parents’ cultural orientations, Latino adolescents growing up in the United States tend to desire amounts and kinds of behavioral autonomy and independence mirroring those of peers being raised in mainstream American families (Martinez 2006; Martinez et al. 2011; Suárez-Orozco and Suárez-Orozco 2001). Given the centrality of autonomy and independence to youth culture in the US, it is valuable to consider how Latino parents from diverse cultural orientations accommodate adolescents’ desires for behavioral autonomy and independence.

Drawing from the notion of parental ethnotheories, or parental belief systems about children and effective parenting (Harkness and Super 1996, 2001, 2006), the present study aims to understand how Latino parents’ cultural orientations are associated with parents’ beliefs about youth behavioral autonomy and independence and, in turn, with related parenting practices. We focus on a sample of Latino parents in Houston, Texas who mostly are of Mexican descent, who represent diverse immigrant generation statuses, and whose adolescent child is in the 10th grade. As in many of the border-state cities in the southwestern US, the parents likely have diverse cultural orientations due to different experiences with, and exposure to, US culture in a city that includes recent and longstanding immigrants. We use latent profile analyses, a person-centered measurement approach useful for modeling heterogeneity among a group (Laursen and Hoff 2006; Magnusson 2003), to capture parents’ cultural orientations. By focusing on youth in middle adolescence, we are able to examine parenting processes at a time in the life course when youth are keenly aware of having opinions and desires for behavioral autonomy and independence distinct from their parents (McElhaney et al. 2009).

Due to cultural values emphasizing parental authority and family interdependence, as opposed to youth’s separation from the family, Latino parents tend to place less emphasis on the importance of adolescent behavioral autonomy (i.e., the ability to make decisions independently) and behavioral independence (i.e., behaving on one’s own; Steinberg 1999) than do parents from many other ethnic backgrounds. For example, when compared to White non-Hispanic parents, Latino parents tend to endorse greater legitimacy of parental authority, have later age expectations for youth to engage in autonomous behaviors, and place less value on youth adopting autonomous behaviors outside the home (DeMent 1998; Fuligni 1998; Fuligni et al. 1999; Phinney et al. 2005). Along these same lines, Mexican–American adolescents often maintain more dependence on their family, as opposed to peers, than both European American and African American youth (Fuligni et al. 1999; Roosa et al. 2005; Tasopoulos-Chan et al. 2009). In terms of parenting practices, Latino parents have been shown to give adolescents responsibilities and rights at older ages, to exert more unilateral decision making, and to provide fewer youth privileges outside the home when compared to European American parents (Bulcroft et al. 1996; Driscoll et al. 2008; Fuligni 1998; Leyendecker and Lamb 1999). In some studies, Latino parents also exert firm control over youth behaviors within the home (Dixon et al. 2008; Finkelstein et al. 2001). For example, among a nationally representative study of high school students, Latino, as compared to European American, parents placed stronger limits on adolescents’ TV watching and had more requirements pertinent to homework and household chores (Blair et al. 1999). Thus, Latino parents’ restrictions have been manifest for a range of youth’s autonomous and independent behaviors.

In addition to having a more restrictive parenting style, Latino parents tend to emphasize traditional gender roles more than parents in mainstream “American” families. Research has shown that Latina girls are more pressured than boys to fulfill household responsibilities (Blair and Cobas 2006; Céspedes and Huey 2008) and that Latino boys receive behavioral privileges at younger ages than do girls (Love and Buriel 2007). Elucidating the reasons for gender differences, Suárez-Orozco and Suárez-Orozco’s (2001) ethnographic work has shown that immigrant parents make a concerted effort to protect their adolescent female daughters from exposure to, and involvement with, American peer groups by restricting girls’ outside-the-home activities. Gender differences in parenting practices may vary, therefore, depending upon the nature of Latino parents’ cultural practices and beliefs.

As with all families, parent–child interactions within Latino families likely differ depending upon the extent to which youth’s autonomous or independent behavior is considered socially acceptable. Studies conducted mostly with samples of European or European American families (with a few exceptions) have shown, for example, that parenting processes vary across domains of youth behaviors such as those pertaining to adolescent peer relations, personal issues, prudential issues (risky to youth’s safety and well-being), morality (“right” from “wrong”), and/or conventional issues (arbitrarily agreed upon behaviors structuring social interactions; Smetana 1995a, b; Tisak and Turiel 1984). Although research unequivocally supports the importance of parental limits around adolescent behaviors in moral and prudential domains (e.g., skipping school, fighting, substance use), less consensus exists for appropriate parental restrictions around “ambiguously personal” behaviors. Ambiguously personal youth behaviors are those that youth perceive as personal (i.e., under youth’s discretion) but that parents might consider as falling under parents’ authority. Ambiguously personal behaviors mostly include activities with friends (e.g., going to girl–boy parties at night) or behaviors crossing both personal and conventional domains (e.g., keeping one’s bedroom clean; Smetana and Daddis 2002).

The ambiguously personal domain of youth autonomy and independence may be especially salient when considering Latino-origin parents raising adolescents in the United States. Many adolescents in Latino families in the US aspire to behave in ways consonant with mainstream American culture. As a result, Latino youth often desire levels and kinds of behavioral autonomy and independence exceeding those considered acceptable in Latino culture and by Latino-origin parents (Sher-Censor et al. 2011). In this regard, it is important to examine linkages between Latino parents’ cultural orientations and parenting processes for ambiguously personal youth autonomous and independent behaviors.

Cultural orientations reflect the complex and multifaceted nature of the acculturation process (Berry 1990; Harwood et al. 2002; Phinney and Flores 2002), whereby, interactions between different cultural groups may change individuals’ “values, ideologies, beliefs and attitudes” and “language, cultural customs, and practices” (Cuellar et al. 1980, p. 209). Cultural orientations, therefore, are expressed in a variety of ways including friendship choices, language use, media preferences, food choice, celebrations, and cultural values, beliefs, and sentiments. Depending upon the specific aspects of US and Latino culture with which individuals identify, biculturalism may take numerous forms. Parents’ identification with English, Spanish, or bilingualism is uniquely distinct from other aspects of acculturation. Unlike other cultural practices, beliefs, and values, English language skills are essential to an individual’s capacity to navigate community and social interactions in the United States. At the same time, English is a difficult skill to acquire absent formal language education and/or daily language immersion. Thus, some parents may make a concerted effort to learn English while maintaining Latino practices, beliefs, and values, while others may resist learning English but otherwise embrace American culture. Consistent with the latter, Guo et al. found that Latino parents in the Miami area reported little preference for English but reported enjoyment of American places and American TV (Guo et al. 2012). Thus, we consider parents’ cultural identifications with language separately from other cultural beliefs, practices and values.

The Current Study

There are three primary objectives in this study of Latino-origin parents and their middle adolescent youth. First, we seek to identify unique types of cultural orientation among parents by using a person-centered approach for analyzing relationships between parents’ language use and cultural behaviors and beliefs (Arends-Tóth and van de Vijver 2006). By modeling associations between US, Latino, and bicultural orientations across indicators of language use and cultural behaviors and beliefs, the person-centered approach will capture patterns of cultural heterogeneity among parents. Our assessment of cultural orientation thus differs from variable-centered measurement approaches, which indicate differences between parents for separate facets of cultural orientation (Laursen and Hoff 2006; Magnusson 2003). We expect to find that patterns of parents’ language use and beliefs and behaviors will include three distinct cultural orientations: Latino, US, and bicultural. Although we do not have a priori expectations regarding cultural orientation profiles beyond these three, we expect to find heterogeneity among bicultural parents. For example, some parents may be predominately English-speaking but ascribe to bicultural or Latino beliefs and behaviors; others may be predominately Spanish-speaking but ascribe to bicultural or US beliefs and behaviors; and others may be bilingual with varied cultural orientations for beliefs and behaviors (Guo et al. 2012).

The study’s second objective is to examine the intervening role of parental belief systems in linking parents’ cultural orientations to parenting. Parental ethnotheories and research on Latino culture suggest that cultural values, beliefs and behaviors reflective of traditional Latino culture will result in parents ascribing to a psychology of parenting placing less emphasis on youth’s behavioral autonomy and independence from parents. Thus, we expect that a US cultural orientation will be associated strongly with parental beliefs and practices emphasizing youth’s behavioral autonomy and independence than will Latino and/or bicultural orientations (although to a lesser extent for bicultural orientations).

The third study objective is to examine gender differences in associations between cultural orientation and parenting. Research suggests that traditional Latino-oriented parents place more limits on the behavioral independence and autonomy of their daughters, as compared to boys, and that such gender differences exceed those occurring in mainstream American families (Suárez-Orozco and Suárez-Orozco 2001). Thus, we expect that a Latino cultural orientation and, to a lesser extent, a bicultural orientation—when compared to a US cultural orientation—will be associated strongly with gender differences in parenting, whereby parents provide less behavioral autonomy and independence to girls than boys (Phinney and Flores 2002).

Analytic models account for factors relevant to cultural orientation and parental autonomy granting. We include youth’s English language use because Latino youth often adopt English more rapidly than do their parents (Telzer 2010), a phenomenon especially likely for this study’s sample of 10th graders who all have lived in the United States since at least the 5th grade (when enrollment in the parent study began, see “Data Source and Sample” below). Youth’s greater facility in English, as compared to their parents, can be disruptive for parent–adolescent relations when youth act as language brokers on behalf of their Spanish-speaking parents (Love and Buriel 2007). We account for youth’s age in years because youth acquire more behavioral autonomy and independence with age (McElhaney et al. 2009) and for youth’s national origin due to the heterogeneity of Latinos from different countries (Hurtado 1995). In terms of parental attributes, we control for parental depression (Simons et al. 1993), linguistic acculturation stress (e.g., pressure to learn English; Gil et al. 1994), and educational attainment due to impacts of each on parenting (Hoff-Ginsberg and Tardif 1995).

Methods

Data Source and Sample

The sample derives from the Variations in Parenting Study (VIPS), a cross-sectional study affiliated with Healthy Passages (HP), a multi-site, community-based study of a representative sample of public school students and their parents or primary caregivers (see Windle et al. 2004, and Schuster et al. 2012 for details). HP youth and parents were interviewed separately at home when youth were in 5th, 7th, and 10th grade. Recruitment, consent, and survey materials were provided in English for youth and in Spanish and English for parents (64 % of parents selected Spanish-language interviews). Among the n = 5,119 dyads interviewed at Time 1, 87 % completed Time 3 surveys.

At the time of the 10th grade survey (2010), HP participants at one site (Houston, TX) were invited to participate in VIPS. The response rate was 86 % (n = 1,347). For the present study, we included Hispanic or Latino-origin VIPS participants (n = 684), a diverse Latino population. Youth self-identified as Mexican (71 %), Central/South American (11 %), and “Other Latino/Hispanic/Spanish” (15 %), and either Puerto Rican, Cuban, or unknown (3 %). Almost a quarter of youth (23 %) were first-generation immigrants (i.e., born outside of the US) who have lived in the US for at least 5 years; 49 % were second-generation immigrants (i.e., born in the US to at least one parent born outside of the US); and, 27 % were third or later generation immigrants (i.e., both youth and their parents were born outside the US). The vast majority of youth (98 %) were 15- to 17-years-old (X = 15.78, SD = .72). Among primary caregivers, most were female (91 % mothers; 3 % stepmothers or grandmothers), with 5 % being fathers and 1 % being stepfathers or grandfathers. VIPS was approved by the University of Texas School of Public Health’s Institutional Review Board.

Measures

Cultural Orientation

Latent profiles of cultural orientation were assessed by four language-use items and 12 items on acculturation behaviors and beliefs, derived from the Pan-acculturation scale (Soriano 2013). Language items included statements such as: “My accent sounds like people from” and “The words I use are from.” Behavior/belief items included: “foods I eat are from;” “traditions I follow are from;” and “the culture that influences that way I think and see things is.” After assigning a value of “1” to response categories, we summed responses for “my own culture” to indicate a score for Latino orientation; responses of “US culture” to indicate a US orientation; and responses of “a mix of my own and US culture” to indicate a bicultural orientation. This resulted in three language-use scales: Spanish speaking; English speaking; and Bilingual (range of scores 0–4); and, three behavior/belief scales: US beliefs and behaviors; Latino beliefs and behaviors; and Bicultural beliefs and behaviors (range of scores 0–12). Variable means and standard deviations are available from the authors.

Young age Expectations for Youth Autonomy/Independence

We measure parental age expectations for youth’s autonomy and independence by parental reports of the age at which it is “OK” for a girl or boy to engage in a behavior (1 = “Before age 12,” 2 = “12–14 years,” 3 = “15–17 years,” 4 = “18 or older,” and 5 = “never”). Items derived from the Teen Timetable Scale (Feldman and Rosenthal 1990), and the responses were reverse coded so that higher scores indicated younger age expectations. The six items (Table 2) pertaining to ambiguously personal forms of youth autonomy indicated the age at which parents think it is okay for a child to choose clothes to buy even if the parent disapproves; go on date; choose his or her hairstyle even if the parent disapproves; do things with friends rather than family; watch as much TV as s/he wants; and, attend girl–boy parties at night. The scale demonstrated adequate internal consistency (α = .72).

Perceived Legitimacy of Parental Authority

We assessed parental perceptions of the legitimacy of parental authority for ambiguously personal behaviors by using the Legitimacy of Parental Authority scale (Smetana 1995b). Parents reported about whether “Is it OK or not OK for parents to make a rule about…” a number of different youth behaviors. Responses ranged from 1 (almost never) to 5 (always). The scale demonstrated adequate internal consistency (α = .75). Items assessed the acceptability of parents having rules for a child’s choice of clothes, cleaning his/her room, eating junk food, and watching TV; a girl wearing heavy makeup; and, a boy wearing an earring.

Parental Supervision

Supervision items inquired about youth’s need to ask parental permission for various behaviors (Stattin and Kerr 2000). The youth-reported measure included five items (α = .88) and the parent-reported measure included four items (α = .77). Items pertained to youth needing to ask a parent prior to staying out late on a weekday evening and before deciding what s/he will do on a Saturday evening when going out with friends; youth having to tell the parent what s/he did and whom if out very late at night; and youth having to tell the parent where they are, who they are with, and what they do with others at night. Parent-reported items mirrored those for youth, with the exception that parents were not asked about youth’s need to inform parents about where and with whom the child would go on a Saturday night. Responses ranged from 1 = no, never to 5 = yes, always.

Youth Autonomy in Decision-Making

Three items from the democratic decision-making scale (Dornbusch et al. 1985; Smetana et al. 2004) were used to assess parent–child decision-making for behaviors such as how late the child stays out, what the child does with friends, and the child being at a friend’s house unsupervised (α = .68). Responses were: 1 = Mother or father decides alone, 2 = I ask my child’s opinion but I have the final say, 3 = We decide together, and 4 = My child decides without discussing with either parent.

Parental Rules

Five parent-reported items inquired about rules regarding how far away from home the child can play; chores; answering the door; using certain things when an adult is not present; and, following after-school routines. Items responses included: 0 = no rule exists, 1 = a rule exists but is not followed, 2 = a rule exists and is sometimes followed, 3 = a rule exists and is always followed (Laird et al. 2009). Higher scores indicated stricter rules (α = .73).

Youth Gender

Adolescents provided a self-report of their gender (0 = female, the reference group; 1 = male).

Youth National Origin

Adolescents reported on their national origin (0 = Mexican, the reference group; 1 = Central/South American; 2 = other).

Youth Age

Adolescents’ age in years was calculated based on youth’s date of birth.

Parental Education

Parents reported the highest education attained by the parent or spouse (1 = 8th grade or less to 7 = college degree or more).

Youth’s English Language Use

We assessed youth’s English language using four language use items on the Pan Acculturation scale (e.g., “My accent sounds like people from;” “The words I use are from”). The summated score was calculated based on responses for “American culture.” (α = .73).

Parental Depressive Symptoms

Parental depressive symptoms were measured by 19 items from the Centers for Epidemiologic Studies of Depression scale (Radloff 1977; α = .92).

Parental Linguistic Acculturation Stress

We assessed parental reports of language discrimination by summating three items on the Riverside Acculturation Stress Inventory (Benet-Martinez 2003 (α = .82).

Analysis

Structural equation modeling (SEM) was conducted using MPlus version 7.0 (Muthén and Muthén 2012). For the minimal amount of missing data (less than 5 %), data were imputed using the full information maximum likelihood method. We used latent profile analyses (LPA) to identify homogeneous subgroups of parents based on observed responses for variables measuring cultural orientation. The optimal number of profiles was determined based on the interpretability and meaning of classes, entropy (measuring how well classes are classified), and fit indices including Akaike information criterion (AIC), Bayesian information criterion (BIC), Lo–Mendell–Rubin likelihood ratio test (LMR; Lo et al. 2001) and Vuong Lo–Mendell–Rubin likelihood ratio test (VLMR; Nyland et al. 2007). We examined indices of fit comparing models with one to six latent profiles; each model was estimated with one additional profile until the model fit the data well (Kass and Wasserman 1995). In terms of the fit indices, we looked for low BIC and AIC, and for the p value from the VLMS to indicate a significant improvement in model fit when compared to the previous model (Golden 2000). After identifying the optimal model, we brought the profile membership variable into SEM as a manifest indicator of cultural orientation.

Intervening and dependent variables were measured as latent constructs. We examined the validity of latent constructs by running Confirmatory Factor Analysis (CFA), which corrects for item measurement error in relationships between survey items and latent constructs. Construct scales were set at one and based on effects-coding methodology so that latent variable estimates were in the metric of observed variables (Little 2013). Measurement and structural models were deemed to fit underlying data adequately when the root mean square error of approximation (RMSEA) was less than .05 and the comparative fit index (CFI) was greater than .90 (Browne and Cudeck 1993; Hu and Bentler 1999). Measurement invariance was examined across both gender and language of survey administration for parents (all youth took the survey in English). In structural models, latent constructs (parents’ beliefs and practices) were regressed on the following covariates: youth’s age, gender, national origin, and English language use, and parent’s education, depressive symptoms, and language acculturation stress. Finally, we examined intervening roles of parental beliefs by testing the joint significance of paths leading to and from constructs measuring parental beliefs (MacKinnon et al. 2002). Following Preacher and Hayes (2008), we examined indirect effects using a bootstrapping methodology—an iterative process for computing the product-of-coefficients estimates for indirect effects based on z scores. Products-of-path estimates leading to, and from, intervening variables were computed a total of k = 5,000 times. Confidence intervals (CIs) were computed for each total and specific indirect effect; CIs not including zero indicated significant indirect effects. Finally, we tested the significance of interaction terms between cultural orientation and youth gender in paths leading to latent parenting constructs.

Results

Latent Profiles of Parents’ Cultural Orientation

Results from latent profile analyses indicated that a five-profile model provided the optimal representation of parents’ cultural orientation. Although BIC and AIC fit statistics improved slightly from the five- to six-class model (Table 1), the theoretical meaning of the five-profile solution was clearer than that of the six-profile solution. Table 2 displays mean scores for each cultural orientation variable across the five latent profiles. As shown, 43 % of parents were characterized as having a “Latino” cultural orientation: within this group, there were high mean scores for Spanish use and Latino behaviors/beliefs. There was considerable heterogeneity among bicultural parents. Bicultural subgroups identified included parents who were “Spanish-speaking/bicultural” (21 %); “bilingual/bicultural” (15 %), and “English-speaking/bicultural” (15 %). A small percentage of parents were “US oriented” (6 %); they had high mean scores for English language use and American behaviors and beliefs.

Table 1 Comparison of models for latent profiles of parents’ cultural orientation (N = 684)
Full size table
Table 2 Standardized mean scores for five-class model of parents’cultural orientation (N = 684)a,b
Full size table

Measurement Model: Autonomy Constructs

The study’s measurement model includes two latent constructs for parental beliefs (young age expectation for youth autonomy/independence; legitimacy of parental authority) and five latent constructs for parenting practices (parental rules; youth-reported autonomy in decision-making; parent-reported autonomy in decision-making; youth-reported parental supervision; parent-reported parental supervision). For latent constructs measured by at least five items, we used item parceling techniques, whereby, responses to two items were summed and averaged. Parceled indicators were developed for measures of young age expectations for youth’s independence; perceived legitimacy of parental authority, parental rules, and youth-reported parental supervision. The advantages of parcels over single items include greater reliability, more communality, a higher ratio of common-to-unique factor variance, reductions in distributional violations, and decreased chance for correlated residuals or dual loadings (Little 2013). Measurement model results are provided in Table 3. As shown, fit statistics indicated a very good fit for the measurement model (χ2 = 603.93, df = 354; CFI = .953; RMSEA = .032, 90 % CI .028, .037), and all parameter estimates were statistically significant at p < .001. Indicator loadings across all constructs ranged from .55 to .85 (see Table 3 for loadings specific to each indicator and construct).

Table 3 Latent variable and unstandardized (standard error) and standardized factor loadings from final measurement modela
Full size table

In order to compare latent means and structural parameters for adolescent males and females and across survey language, we tested the invariance (or equivalence) of latent constructs across gender and parents’ survey language use (Little 2013). Weak and strong criteria were used to establish invariance across groups. Following Chen (2007), we considered invariance to exist when changes in model CFI were less than .01 and when RMSEA values fell within the confidence interval of the prior model. Changes in model fit statistics (results available from the authors) indicated measurement equivalence across both gender and parents’ survey language use when going from configural (unconstrained) to weak (constrained factor loadings) to strong (constrained intercepts) models. See Cheung and Rensvold (2002) for an explanation of measurement invariance tests.

Structural Models: Correlations Among Latent Constructs

We begin with a description of significant associations among latent parenting constructs. In terms of parental beliefs, younger age expectations for youth autonomy/independence were associated with less perceived legitimacy of parental authority (β = −.26, SE = .05, p < .05). In regard to parenting practices, parental reports and youth reports were associated positively for measures of parental supervision (β = .11, SE = .05, p < .05) and youth autonomy in decision-making (β = .26, SE = .05, p < .001). Further, parental reports of supervision were associated with more parental rules (β = .24, SE = .05, p < .001) and with less youth autonomy in decision-making as reported by parents (β = −.28, SE = .05, p < .001) and youth (β = −.16, SE = .05, p < .01). Similarly, youth reports of parent supervision were associated with parents reporting more parental rules (β = .10, SE = .05, p < .05) and with less youth autonomy in decision-making as reported by parents (β = −.17, SE = .05, p < .01) and youth (β = −.54, SE = .04, p < .001). There were a small number of significant associations from background variables to parenting constructs (findings available upon request).

Structural Models: Cultural Orientation to Parenting Beliefs and Practices

Figure 1 illustrates structural model results for the direct and indirect (through parental beliefs) associations linking parents’ cultural orientation to parenting practices. When compared to Latino cultural orientations, bilingual/bicultural, English-speaking/bicultural, and US cultural orientations were associated indirectly, through parenting beliefs, with parenting practices granting adolescents autonomy and independence. When compared to Latino-oriented parents, bilingual/bicultural, English-speaking/bicultural, and US-oriented parents reported less perceived legitimacy of parental authority. Parents who were English-speaking/bicultural or US-oriented also endorsed youth autonomy at a younger age than did Latino-oriented parents. Parental beliefs, in turn, were associated with parenting practices. Specifically, younger age expectations for youth autonomy/independence were associated with less parental supervision (parental and youth reports), fewer parental rules, and with more youth autonomy in decision-making (parental and youth reports). Legitimacy of parental authority also was associated with more parental supervision (parental and youth reports), more parental rules, and less youth autonomy in decision-making (parental and youth reports).

Fig. 1
figure 1

Structural pathways from Latent profiles of parents’ cultural orientations to parental beliefs to parenting practices granting adolescents autonomy and independence, N = 684. Notes Standardized coefficients shown. a Reference group: Latino cultural orientation. Models control for correlations among intervening, among dependent variables, and between both intervening and dependent variables and the following: youth age, national origin, gender, and use of English language; and, parent education, depressive symptoms, and acculturation stress. Model fit statistics: χ2 (354) = 603.93; CFI = .953; RMSEA = .032 (90 % CI .028, .037). *p < .05, ***p < .01, ***p < .001

Full size image

Models using bootstrapping methods with confidence intervals confirmed that parental beliefs functioned as significant intervening variables between cultural orientation and parenting. Results for indirect effects are in Table 4. When compared to a Latino cultural orientation, US and English/bicultural orientations were associated indirectly, through parental beliefs, with fewer parental rules, more youth autonomy in decision-making (parental and youth reports), and less parental supervision (parental and youth reports). Further, when compared to the Latino cultural orientation, the bilingual/bicultural orientation was associated indirectly through parental beliefs with fewer parental rules and less parental supervision (parental and youth reports). The one significant direct effect from cultural orientation to parenting indicated that bilingual/bicultural parents reported more parental supervision than Latino-oriented parents.

Table 4 Total and specific indirect effects from models using bootstrapping methods to test indirect effects, point estimates and 95 % confidence intervals showna,b (n = 684)
Full size table

Results from models including interaction terms between youth’s gender (with male as the reference group) and parents’ cultural orientation (with Latino as the reference group) did not indicate any significant gender differences in associations between cultural orientation and parenting beliefs or practices. Thus, there was no support for the hypothesis that Latino and bicultural (as compared to US) orientations would be associated more strongly with parenting practices placing less emphasis on youth behavioral autonomy and independence for adolescent females, compared to adolescent males. Rather, significant main effects for youth’s gender indicated that boys experienced more autonomy in decision-making (β = .18, SE = .05 parental report, β = .32, SE = .04 youth report, both at p < .001) and less parental supervision (β = −.14, SE = .04 parental report, β = −.23, SE = .04 youth report, both at p < .001) than did girls.

Discussion

Although Latino youth growing up in the United States often prefer “American” ways of behaving and acting, many of their parents do not have the same preferences for their children. In this study, we address the need for a culturally informed understanding of Latino adolescent development by examining how parents think about and attempt to influence their adolescent child’s behavioral autonomy and independence. We focus on youth’s autonomy and independence in the “ambiguously personal” domain of behavior—activities that adolescents perceive as involving youth’s own personal choices and decisions but that parents may view as under parental authority and control. Due to youth’s more rapid immersion in US culture when compared to parents, parent–youth discrepancies in ideas about youth’s behavioral autonomy and independence may be pronounced for Latino families. As a result, the ambiguously personal domain may be especially relevant for Latino parent–youth dyads (Sher-Censor et al. 2011).

Cultural diversity among Latino-origin parents appears to be a salient factor affecting Latino adolescent development. Latino parents’ cultural orientations may influence how parents think about youth’s behavioral independence and autonomy and, in turn, with how much autonomy and independence parents provide to their adolescents. Confidence in the validity of our findings is strengthened by the consistency of results across parental and youth reports of parenting. We begin by discussing the diverse cultural orientations characterizing Latino parents in our study and, subsequently, expand on findings about the links between cultural orientation and parenting processes.

Diversity in Cultural Orientation Among Latino Parents

Concurring with the work of Coatsworth et al., our assessment of parents’ cultural orientations supports the value of a person-centered approach for the assessment of acculturation (Coatsworth et al. 2005). The use of latent profile analyses helped identify cultural orientation patterns for different aspects of parents’ acculturation (language, behavioral/belief systems) as reflecting “American,” Latino, or both cultures. Findings from latent profile analyses reflect a cultural heterogeneity among parents that cannot be captured by single-indicator measures of acculturation (e.g., language use) or by multidimensional measures presuming the same cultural identification (e.g., bicultural, Latino, US) across diverse facets of acculturation.

Consistent with expectations, the beliefs, behaviors, and language use characterizing parents in this study suggest that Latino parents represent US (i.e., “American”), Latino, and bicultural cultural orientations (Berry 1990; Harwood et al. 2002; Phinney et al. 2005). Clearly, the largest proportion of parents (43 %) is “Latino” in cultural orientation; these parents identify with Spanish language and with beliefs and behaviors reflecting their own culture. In contrast, a small proportion of parents (6 %) is characterized by a US cultural orientation; these parents identify with English and with “American” beliefs and behaviors. Perhaps most informative are the study’s findings for diverse kinds of bicultural parents. Bicultural parents include those who are Spanish-speaking, bicultural (21 %), bilingual/bicultural (15 %), and English-speaking bicultural (15 %). Taken together, results from latent profile analyses are consistent with scholarship suggesting that Latino parents’ identification with language (Spanish, English or bilingualism) has an orthogonal relationship with other cultural behaviors (e.g., celebrations, music, food) and cultural beliefs and values (Guo et al. 2012).

Despite variability in parents’ cultural orientation, virtually all parents in this sample (94 %) are characterized by either a Latino or bicultural orientation. We surmise that the retention of Latino cultural practices and values, even if accompanied by the adoption of “American” cultural practices or beliefs, might be unique to Latinos living in areas with a longstanding history of Latino settlement. Cities such as Houston have been home to Latino immigrants for many decades (Coatsworth et al. 2005) and, as such, community institutions may be more tailored to Latino culture than is the case in newer immigrant areas (Martinez et al. 2011). For example, parents may live in neighborhoods and children may attend schools where Latino cultural norms and practices are relatively engrained, facilitating parents’ retention of at least some Latino cultural orientations. It is also possible that parents’ maintenance of cultural beliefs and practices is explained by “selective acculturation theory” (Portes and Rumbaut 2001). According to this theory, Latino families retain cultural values such as those emphasizing family cohesion, support, and mutual obligations in an effort to avoid downward assimilation into high-poverty US contexts, often characterized by family dissolution and chronic unemployment (Portes and Rumbaut 2001). Regardless of the reasons for parents’ bicultural or Latino cultural identification, it is evident that Latino adolescent developmental processes occur in family contexts, which are culturally distinct from those experienced by many other adolescents in the United States.

Cultural Orientations: The Relevance to Parenting Processes

Although the cross-sectional nature of our data prohibits causal inference, structural model results indicate indirect effects from cultural orientation to parental beliefs to parenting practices. When compared to Latino-oriented parents, US-oriented parents and to some extent bicultural parents, appear to provide independence and autonomy to their adolescent by virtue of believing that youth should adopt “ambiguously personal” independent behaviors at younger ages and that parents do not have authority over adolescents’ “ambiguously personal” independent behaviors. Aligning with Harkness and Super’s conceptualization of parental ethnotheories (Harkness and Super 1996, 2001, 2006) and Ogbu’s (1981, p. 420) “native theory of success,” cultural conditions may shape parents’ ideas about developmental tasks important to youth’s success and positive adaptation and, in turn, how parents raise their adolescent children. Parents with US cultural orientations and, to some extent, those with bicultural orientations, might view youth’s success in more individualistic terms than is the case for Latino-oriented parents. Such individualistic beliefs, in turn, might explain why more bicultural and US-oriented parents provided less parental supervision, fewer behavioral limits, and more decision-making autonomy to their adolescent children than did Latino-oriented parents.

Contrary to expectations, we did not find evidence for gender differences in associations between parents’ cultural orientation and parenting processes. Rather, regardless of parents’ cultural orientation, adolescent males received less parental supervision (for both parental and youth reports) and had greater autonomy in decision-making (for parental and youth reports) than did adolescent females. Although the greater behavioral independence and autonomy provided to boys, as compared to girls, aligns with the general literature (Barber et al. 1994; Smetana and Daddis 2002), we were surprised that gender differences were not pronounced for youth whose parent had a traditional Latino cultural orientation. Our measures of parenting, pertinent to ambiguously personal youth behaviors, might not adequately capture the gendered nature of parenting among Latino families. For example, prior research has shown that Latino parents provide girls greater behavioral independence for behaviors occurring inside the home and give boys more behavioral independence outside the home (Bulcroft et al. 1996). Thus, the location in which youth engage in independent or autonomous behaviors may be especially salient to our understanding of cultural variability in parenting influences on Latino boys as compared to Latina girls.

The Importance of Language Cultural Identity

The prominence of linguistic acculturation for cultural orientation patterns and parenting is incontrovertible. Language use not only differentiates Latino- and US-oriented parents but also partitions bicultural parents into those who are English-speaking, bilingual, or Spanish-speaking. When compared to Spanish-speaking, Latino-oriented parents, those who were bilingual/bicultural or English-speaking placed less emphasis on the legitimacy of parental authority; further, English-speaking/bicultural and US-oriented parents reported younger ages at which youth’s behavioral autonomy and independence are acceptable. In contrast, we found no significant differences in parenting beliefs for the Spanish-speaking parents. Our findings build on Guo et al.’ evidence suggesting distinctions in Latino parents’ cultural identification with language versus other cultural practices and beliefs (Guo et al. 2012).

Findings pertinent to linguistic acculturation may be enhanced by considering the social networks of parents in this study. English-speaking and bilingual parents likely have more social ties to individuals characterized by US-oriented behavioral norms than is the case for parents whose social circles are restricted to Latinos (Nawyn et al. 2012). As an example, a parent whose friendships includes parents of other US-born children (in part, by virtue of having English language skills) is more likely to adopt beliefs about appropriate levels of parental authority in ways consonant with US culture than would be possible for a parent whose friendships are limited to other Latinos. Adolescents whose families are integrated into US social networks also may feel justified in demanding behavioral autonomy or independence from parents as a matter of “fitting in” with peers. Regardless of whether effects are parent-driven, child-driven, or both, language use may change social settings for immigrant families in ways that have profound consequences for parents’ beliefs and practices.

Future Directions in Research

Although our work represents an important first step in understanding Latino adolescent development vis-à-vis parenting and parents’ cultural orientations, its shortcomings should be addressed in future research. In terms of study design, for example, we will benefit from longitudinal studies clarifying causal pathways and from mixed-methods research providing an enriched understanding of the cultural meaning and salience of “ambiguously personal” forms of autonomy and independence for Latino youth and their parents. Research also is needed to extend the focus on parenting to examining adolescents’ adjustment. Thus, although studies suggest that very little or very much autonomy and independence is associated with youth experiencing more externally oriented problems and worse mental health (Barber et al. 1994; Smetana 2005; Smetana et al. 2004), links between autonomy/independence and youth adjustment may be culture-specific. For example, parents’ giving their youth more freedom and autonomy for ambiguously personal issues has been related to less delinquency for Mexican-origin, but not European American, adolescents (Mounts 2004). It is also important to examine differences between Latino fathers and Latina mothers (Crean 2008; Crockett et al. 2007), something we were unable to do given that just 5 % of our parent sample included fathers. Prior research on Mexican–American families has shown that fathers exert more unilateral decision-making around some forms of adolescent behavioral autonomy than mothers and grant more autonomy in decision-making to sons than daughters (Perez-Brena et al. 2012). Thus, father-reported data would enable us to examine paternal beliefs about, and experiences with, youth autonomy in the context of cultural orientations and parent–youth gender dyads. Finally, the use of a pan-Latino sample of families limited our capacity to understand differences between Mexicans, Nicaraguans, El Salvadorans, Guatemalans, and other groups and, as a result, there is a need to examine how autonomy processes and parenting evolve among specific Latino groups.

Conclusions

Our research suggests that Latino parents living in a city with a longstanding history of Latino settlement have diverse cultural orientations, which affects how parents think about and behave in terms of their adolescent child’s behavioral autonomy and independence. Culturally informed interventions addressing family functioning for the large and growing Latino population will benefit from attention to nuanced ways in which cultural contexts matter for parenting goals and strategies. In this way, we may be better prepared to ensure the successful development of the large and growing population of Latino youth in the United States.