The acculturation gap of parent–child relationships in immigrant families: A national study

Abstract

Objective:

We examined the acculturation processes involving intergenerational consonance and dissonance in parent–child relationships in U.S. immigrant families.

Background:

This study is important because we lack national studies that examine the association between acculturation processes and intergenerational relationships among diverse racial/ethnic groups in immigrant families.

Method:

Using national data from Add Health with diverse race/ethnicity, we measured acculturation levels by immigrant generation, age of arrival, and length of time. Intergenerational consonance (the degree to which children and parents share the same values and activities) was measured by family cohesion and sharing meals (specifically dinners) with parents. Intergenerational dissonance (the degree to which parents and children differ in expected norms and parents lose authority over their children) was measured by parent–child conflict and parental control. Ordinary least square, binary logistic, ordered logistic, and Poisson regressions were conducted depending on the nature of the four dependent variables.

Results:

We found robust evidence that adolescents of the second immigrant generation acculturate more rapidly than those of the first generation and their immigrant parents creating a “gap” in intergenerational relationships. Thus, second-generation adolescents experience lower levels of family cohesion, less frequency of sharing weekly dinners with parents, less parental control of adolescents’ activities, and more serious arguments about their behaviors with their parents than their first-generation counterparts.

Conclusion:

This is the new evidence that is based on national data, across multiple measures of intergenerational relationships, and holds for diverse racial and ethnic groups.

Implications:

The findings underscore the importance of developing culturally informed interventions supporting healthy parent–child relationships in immigrant families.

One of the most profound demographic changes that ushered in the new millennium was the dramatic rise in immigration to the United States and the rapidly changing ethnic diversity of the U.S. population (Portes & Rumbaut, 2006). This change echoed a similar phenomenon of over 100 years ago in both old and new ways (Alba & Nee, 2003; Farley, 1996). From 1990 to 2018, the nation’s immigrant population aged 18 years or older increased by about 25 million (from 19.8 million in 1990 to 44.7 million in 2018), adding roughly 900,000 immigrants to the population per year (Migration Policy Institute, 2020); it has been projected that most of the U.S. population growth in the coming decades will continue to be due to immigration, including both the increase from immigrants themselves and the increase from their higher fertility rates (Colby & Ortman, 2015).

Immigrant children (first generation who are foreign born) and U.S.-born children (as second generation) of foreign-born immigrant parents are the fastest growing segment of the U.S. population of young people under age 18 years, accounting for 25.9% of all American children in 2018 (Migration Policy Institute, 2020), including about 54% of all Hispanic children and 17% of all Asian American children (Child Trends Data Bank, 2018). Today, one of four American children under age 18 years are children of immigrants (including both first-generation children with a foreign birth and second-generation children with a native birth to foreign-born parents). Because most children in immigrant families belong to Hispanic or non-White racial and ethnic minorities, reflecting the post-1965 immigration waves from Latin America and the Caribbean and from Asia and the Middle East, future projections indicate that the proportion of children under 18 years who are non-Hispanic White will decline from about 52% in 2014 to about 36% in 2060 (Colby & Ortman, 2015).

The tremendous influx of immigrant families into the United States over the past few decades has transformed the social and economic landscape for family interactions and the development of children under age 18 years. Despite considerable research on the experiences and adaptation of immigrant adults, there has been limited attention to the immigrant experiences of young children and adolescents in studies of immigration (e.g., Jasso & Rosenzweig, 1990; Li & Warner, 2015; Lieberson, 1980; Tienda & Haskins, 2011). This is largely due to a lack of data on immigrant descendants or missing information on nativity (Jensen & Chitose, 1996; Portes, 1996; Rumbaut, 2014; Waters & Gerstein Pineau, 2015). Only within the past several years have studies addressed the adaptation processes and outcomes of children and adolescents in the new immigration, but this research is mainly based on regional surveys that exploit the geographic concentration of immigrant families or on specific immigrant ethnic groups (e.g., Fuligni, 1997; Li & Warner, 2015; Perez, 1994; Portes & Rumbaut, 2001a; Rumbaut, 1994; Suárez-Orozco & Suárez-Orozco, 2001; Tseng & Fuligni, 2000; Van Hook & Balistreri, 2002; Waters, 1996; Zhou, 2001). Research using national data rarely have sufficient sample sizes to identify separate ethnic groups (Rumbaut, 2014). We therefore lack a national and representative view of the well-being of immigrant families and their descendants and have limited understanding of their acculturation processes.

Data from the National Longitudinal Study of Adolescent to Adult Health (Add Health) has made possible exciting new research on the well-being of immigrant descendants and families (e.g., Duong et al., 2016; Harker, 2001; Harris, 1999; Marks et al., 2014; Tillman et al., 2006; Xie & Greenman, 2011. This study oversampled certain Hispanic and Asian ethnic groups (described in the Data section), which increased the number of children from immigrant families who were selected. The study also occurred in the mid-1990s, capturing the increasing representation of young people under 18 years from immigrant families in the United States and has unprecedented diversity in race and ethnicity on a national level.

We use Add Health data to examine processes of acculturation in parent–child relationships across generations among immigrant parents and their adolescents in the United States. We assess acculturation by contrasting parent–child relationships among adolescents of the first generation (foreign-born adolescents with foreign-born parents) with those of the second generation (U.S.-born adolescents with foreign-born parents) and according to the age of arrival for the adolescents and the length of time immigrant families have lived in the United States. In this data set, the large portion of immigrant families were from countries in Latin America and the Caribbean, Asia, and the Middle East where their parents and first-generation adolescents were born.

Both first- and second-generation adolescents from these immigrant families have been exposed in various degrees to cultures and languages of their home country or their parents’ country of origin. Following Portes and Rumbaut (2001b, 2001c) and Zhou (2001), we attempt to capture dissonant and consonant acculturation. Generational dissonant acculturation captures the negative connotations of a generational gap when parents’ values and expected norms of behavior differ substantially from the ones of their adolescent children. In immigrant families, however, generational dissonance takes on unique meaning and can indicate differential rates of acculturation, where children adapt more rapidly into the mainstream American culture than parents. Generational consonance describes the other end of the continuum whereby parents and adolescents do not experience a generation gap but rather share common values, expectations for their future, and similar worldviews. Consonance may represent similar rates of acculturation among adolescents and parents. Therefore, consonance may be expected to be more evident among adolescents of the first generation in which both children and parents are foreign-born.

DIFFERENTIAL ACCULTURATION IN IMMIGRANT FAMILIES

Piaget (1936) was one of the few scholars who first used “assimilation” to conceptualize the adaptation processes. Children and young people are constantly and rapidly assimilating new information and experiences into their existing knowledge about the world and reinterpreting the new experiences. Piaget argued that they not only add new information and experiences into their existing cognitive structure, but also accommodate by changing old ideas and replacing them based on the new information and experiences they gather and gain. Young people often keep a balance (with equilibrium) between assimilation and accommodation as a learning process for adaptation into a new social world (see also Di Paolo et al., 2014).

Immigrant children who were born in the United States but into immigrant families have been exposed to different cultural values and practices of their immigrant parents than the children in nonimmigrant families. Thus, immigrant children are more likely to go through the processes of both assimilation and accommodation in their adaptation into the host society. They constantly add new information gained through the experiences earned in the mainstream, reinterpret and reevaluate the cultural values and practices their parents taught them, and even replace the old ideas with the new information they have received outside the family. Depending on their nativity status, age at arrival, and length of stay, offspring of immigrant families may vary at their pace of adaptation. Some are assimilating but keep a closer connection to their parents’ cultural values, which leads to consonant intergenerational relationships; others keep a balance between assimilating and accommodating with equilibrium; still others are accommodating than assimilating in disequilibrium and deviate rapidly from their family practices, which may lead to dissonant intergenerational relationships.

Furthermore, immigration scholars employ the concept of acculturation to study adaptation of immigrants and their descendants (e.g., Buriel, 1993; Parke & Buriel, 2007). Various definitions of acculturation have been offered, but its general conceptualization refers to the changes that take place among immigrants and their descendants as they come in contact with a new society (Berry, 1997), including a unidirectional shift from native to host culture or a more multidimensional negotiation of old and new traditions (e.g., Buriel & De Ment, 1997).

Buriel and colleagues found variations in socialization practices among Latinos, African American, American Indians, and Asian American families (Buriel & De Ment, 1997; Parke & Buriel, 2007). These scholars argued that children acculturate more rapidly into the main culture than their parents because they are younger, more malleable, and more deeply exposed to the new culture through the native school system (Chud, 1982; Coll & Magnuson, 2004). Adjustment according to immigrant factors, such as birthplace, age of immigration, language proficiency, and ethnic composition of social networks have been used to understand the acculturation processes of immigrants and their descendants (e.g., Feldman et al., 1992, compared first vs. second generation; Fuligni, 1997, compared Asian vs. White; Kao & Tienda, 1995, considered academic achievement 1995; Knight et al., 1993, studied ethnic identity; Suárez-Orozco & Suárez-Orozco, 1995, studied achievement motivation between U.S.-born vs. first generation).

ACCULTURATION AND PARENT–CHILD RELATIONSHIPS

On the one hand, immigrant parents tend to focus on survival as well as economic mobility in the host society and often hold tightly to values, norms, and behaviors acquired in their home country to assess their accomplishments and educate their children. Children, on the other hand, especially those in the second generation who are U.S. born to foreign-born immigrant parents, are more likely to be attracted by the culture of the host society and influenced by their American peers and other forms of mass media and to have a strong desire to fit in (Zhou, 2001). Thus, children in immigrant families may hold different values and expectations for their lives in the United States than their parents, and this is often interpreted as generational dissonance. An alternative interpretation is that generational dissonance simply represents differential acculturation by children and parents to the host society.

Because adolescents of the second generation are U.S. born, speak English fluently, and are socialized in American schools and neighborhoods, generational dissonance tends to be more evident in parent–child relationships among the second-generation adolescents than among first-generation adolescents. In other words, U.S.-born children in immigrant families are more likely to have experienced greater acculturation of American society than their foreign-born parents. On the other hand, generational dissonance is expected to be less among foreign-born adolescents who have spent time in their country of birth and may not have mastered English depending on their age at arrival to and length of stay in the United States and are therefore likely to acculturate to U.S. society at a similar pace as their immigrant parents.

Most research to date that has examined notions of generational dissonance and consonance or intergenerational conflict in immigrant families has focused on language dissonance as the measure of differential acculturation (e.g., Dennis et al., 2010; Portes & Rumbaut, 2001b; Tseng & Fuligni, 2000; Zhou, 2001). The relationship between parents’ proficiency in English and children’s retention of parental language has been used as a framework for predicting consonant and dissonant acculturation (Portes & Rumbaut, 2001b). Portes and Rumbaut (2006) argued, for example, that when parents are fluent in English, consonant acculturation in parent–child relationships is the outcome regardless of children’s language loss because parents can stay in step with the acculturation of children and maintain open channels of communication. Dissonant acculturation occurs, however, when parents speak no or limited English and children have limited bilingualism. Children and parents can communicate in a limited way in the home, but parents cannot keep up with their children’s acculturation of other American ways.

Dissonant acculturation can rupture family ties, result in a loss of parental authority, lead to role reversals where the child becomes a culture broker for his or her parents, and increase levels of intergenerational conflict (Portes & Rumbaut, 2006). Another strand of research has examined parent–child consonance and dissonance (although these terms are often not explicitly used) in educational expectations (Areepattamannil & Lee, 2014; Fuligni, 1997; Hao & Bonstead-Bruns, 1998). In general, research has indicated that high parental expectations for educational achievement enhances children’s actual achievement and that greater parent–child interactions promote consonance in the educational expectations of parents and children.

CURRENT STUDY

We expand measures of intergenerational relationships in this study to reflect parent–child conflict, power dynamics, closeness, interactions, and shared time in the relationship (Portes & Rumbaut, 2001c; Zhou, 2001). As indicators of generational dissonance, we measure parent–child conflict and parental control with respect to the child’s behavior. Generational consonance is indicated by measures of family cohesion and the sharing of weekly dinners together. Our consonance measures are meant to capture closeness and shared values and traditions among children and parents, but note that where such measures are low, this can be interpreted as dissonance and evidence of a generational gap in these indicators (Chen & Harris, 2019; Harker, 2001). For example, when family cohesion is low, this suggests parents and adolescents experience less closeness and warmth in their relationships with parents and are not able to communicate effectively or satisfactorily.

To the extent that generational dissonance is higher and consonance lower among adolescents of the second generation compared with those of the first generation, we anticipate that acculturation processes may underlie these differences whereby U.S.-born adolescents are likely to adopt the values and norms of behavior in the host society more rapidly than their foreign-born counterparts. Parents of both second- and first-generation adolescents are foreign-born, so differential acculturation operating at the child’s level alters parent–child relationships in immigrant families. We therefore consider generation as one proxy for acculturation.

We further test this acculturation hypothesis by examining differences in parent–child relationships by the length of time adolescents have lived in the United States, which can also serve as a proxy for the length of time immigrant parents have been in the United States (and the least amount of time for parents of second-generation adolescents). This allows us to examine the acculturation level represented by differing lengths of time lived in the United States and its association with parent–child relationships. Greater consonance and lower dissonance in intergenerational relationships are expected among adolescents who have lived in the United States for shorter periods than those who lived longer in the United States.

We also examine differentials in parent–child relationships by the age of arrival of the adolescent. Again, this allows us to examine further evidence that differences represent an acculturation process whereby we expect greater consonance and lower dissonance in parent–child relationships among adolescents who arrive in the United States at an older age than those who arrive at a younger age. The longer the stay and younger the age of arrival mean more time and earlier the developmental stage of exposure to American attitudes and norms of behavior by attending American schools, growing up in American neighborhoods, and developing friendships and exposure to peers in those schools and neighborhoods, which may lead to the acculturation gap between the adolescents and their parents, resulting in dissonant parent–child relationships.

Hypotheses

In sum, based on multiple strands of the conceptual framework about the adaptation of immigrants and their descendants, we formulate three hypotheses regarding the relationship between the acculturation levels and parent–child relationships in immigrant families: first, second-generation adolescents will have greater dissonance (in terms of less parental control and higher parent–child conflict) and less consonance (in terms of lower levels of family cohesion and spending less family meal time with parent/s) in their parent–child relationships than first-generation adolescents; second, a longer length of stay in the United States will be associated with greater dissonance and less consonance in parent–child relationships; and third, an earlier age at arrival (controlling for age) will be associated with greater dissonance and less consonance in parent–child relationships.

Covariates of family, school, and neighborhood contexts

We will adjust for contextual measures of the social acculturation environment, including family, school, and neighborhood or community contexts (Portes, 1996; Zhou, 1997) in the multivariate analysis of this study. If the social environments in which immigrant families live reinforce ethnic values and cultural traditions, acculturation by immigrant parents and their children tends to occur at the same pace and leads to consonant intergenerational relationships. For example, children in immigrant families are more closely tied to their ethnic culture, and, in turn, their parents’ culture and cultural norms. Especially when they live in intact families, those families are embedded in tightly knit social networks in ethnic communities, and children are involved in these networks (Bankston, 2004; Zhou, 1997, 2001). Such networks of social relationships involve shared obligations, social supports, and social controls, which operate to monitor children in ethnic communities collectively and promote traditional cultural values of family orientation and respect for authority, which, in turn, promote intergenerational consonance in immigrant families.

When immigrant descendants are less embedded in ethnic communities and networks and less tied to their cultural traditions in the family context, they tend to more readily acculturate the values and norms of the host society which are alien to parents’ cultural values and create an acculturation gap between parents and children. As children acculturate American ways through their exposure to the native culture in nonethnic neighborhoods, schools, and peer relationships, they adopt the developmental behavior typical of American adolescents that opposes the moral authority of parents and other adults and emphasizes peer recognition over family matters. The class background of the parents, social networks in which parents are involved, the type of school adolescents attend, and family structure affect parents’ ability to shield children from neighborhood and school peer groups whose behavior often run counter to parental expectations (Waters, 1996). Thus, adolescent social contexts that contain less exposure to ethnic cultural values and norms and more exposure to mainstream native adolescent values and norms will speed the acculturation processes of immigrant youth compared with their parents, and intergenerational relationships tend to become dissonant.

Family factors are the most salient environmental context for parent–child relationships, reflecting cultural and structural mechanisms. Religiosity and language spoken at home characterize the cultural context and affect the differing acculturation levels between immigrant descendants and their foreign-born family members. Structural familial factors, including parental education, family structure, and number of siblings, result in varying degrees of social mobility in the receiving country, which can influence the adaptation experiences of immigrant parents and their children. These factors may promote strong cultural and ethnic ties through social networks in ethnic institutions, maintaining parents’ native language, and living in a traditional family. Parents’ educational level is another salient factor, with high levels of education indicating social and economic mobility, which is often associated with more rapid assimilation of American ways (and thus greater intergenerational consonance typical of American parent–child relationships in adolescence), and low levels often indicating greater adherence to cultural traditions and ethnic networks.

Additionally, we will adjust for the context of Americanization at school and in communities; such as the region of the school; school type (public or private); school size; percentages of students who are first, second, or third+ generation; percentages of non-Hispanic White students in the school; ethnic and immigrant presence in the neighborhood; the proportion of residents aged 5 or over who speak English not well or not at all (defined by the census as “linguistically isolated”); the proportion of Hispanics in the neighborhood; and urban location.

School-level factors may capture ethnic presence in the school and opportunities for co-ethnic friendships in which adolescents of similar racial/ethnic background are more likely to become friends. Large public schools are likely to contain greater ethnic and immigrant diversity than small private schools, as are certain regions of the country (e.g., West). Greater immigrant presence in a school reinforces ethnic cultural values through co-ethnic friendships and school climate. Conversely, when the student population is largely non-Hispanic White, immigrant youth will be more exposed to American youth culture and American norms governing intergenerational relationships. Neighborhood context factors also can capture ethnic and immigrant presence in the neighborhood, reflecting the potential for co-ethnic social networks of parents, children, and extended kin. Urban neighborhoods with high proportions of Hispanics, foreign-born immigrants and their offspring, and those who are linguistically isolated are more likely to contain immigrant populations with dense ethnic social networks that promote adherence to cultural norms of family interaction and behavior.

METHOD

Participants

Data come from the National Longitudinal Study of Adolescent to Adult Health (Add Health), a nationally representative study of more than 20,000 adolescents in Grades 7 through 12 in the United States in 1995 (Harris, 2013; Harris et al., 2019). Add Health was designed to help explain the causes of adolescent health and health behavior with special emphasis on the effects of multiple contexts of adolescent life. The study used a multistage, stratified, school-based, cluster sampling design. A stratified sample of 80 high schools was selected with probability proportional to size. For each high school, a feeder school was also selected with probability proportional to its student contribution to the high school. The school-based sample therefore has a pair of schools in each of 80 communities. An in-school questionnaire was administered to every student who attended each selected school on a particular day during the period from September 1994 to April 1995 and was completed by more than 90,000 adolescents.

In a second level of sampling, adolescents and parents were selected for in-home interviews. From the school rosters, a random sample of some 200 students from each school pair was selected, irrespective of school size, to produce the core in-home sample of about 12,000 adolescents. Add Health oversampled several subgroups, including ethnic samples (Cuban, Puerto Rican, and Chinese adolescents), physically disabled adolescents, and a genetic sample, for in-home interviews, which were conducted between April and December 1995, yielding Wave I data. The core plus the special samples produced a total sample size of 20,745 adolescents in Wave I in-home survey. A parent, generally the mother, was also interviewed in Wave I. See Harris (2013) for a more detailed description of the Add Health study. All adolescents in Grades 7 through 11 in Wave I were targeted roughly 1 year later for the Wave II in-home interview.

The analytical sample of this research is restricted to first and second-generation adolescents who participated in the Wave I in-home interview and who had valid data on immigrant generation, ethnic group background, and sampling weights. These restrictions result in a sample size of about 4,500 (almost one in four adolescents in Add Health is a first- or second-generation adolescents, higher than the national representation because of over-sampling of various ethnic groups). Missing values on parent–child relationship measures reduce the sample further to 4,101.¹

Measures

Generational consonance and dissonance

The measures of generational consonance include family cohesion and dinners. Family cohesion is measured by an additive index of responses (ranging from 1 = low to 5 = high) of adolescent reports on feelings about how much people in their family understand them, how much they and their family have fun together, and how much their family pays attention to them (alpha = .79). Whether the adolescent eats the evening meal with a parent most days of the week (5–7 days/week) represents a measure of shared time and communication in which parents and children engage on a regular basis. Our data contain two direct measures of generational dissonance as defined in the literature. Parental control is measured by the total count (ranging from 0 = low to 7 = high) of decisions about daily activities that parents do not allow their children to make on their own, including the time adolescents must be home on weekend nights, the people they hang around with, what to wear, how much television to watch, the kind of television programs to watch, the time to go to bed on weeknights, and what to eat. Parent–child conflict is measured by the mean response (ranging from 0 to 1) of adolescent reports on whether they had a serious argument about their behavior with the mother and the father in the past 4 weeks (separate question for each parent). For adolescents who live with only a mother or only a father, we use the one report; for adolescents who live with both parents, we average the report for mothers and fathers. Preliminary analysis reveals that our consonance measures are reasonably correlated, with a correlation of .22 between dinners and family cohesion. These measures are also negatively correlated with intergenerational conflict at about the same level (−.22 correlation between family cohesion and parent–child conflict). Parental control is less correlated with the other measures.

Acculturation levels

Acculturation variables are represented by immigrant generation, length of stay in the United States, and age at arrival to the United States. Immigrant generation is coded as a two-category variable: foreign-born adolescents to foreign-born parents (first generation) and native-born adolescents to foreign-born parents (second generation). Generation is determined by questions about place and country of birth and citizenship status (Harris, 1999). Length of stay is measured in years by subtracting the age of arrival from the age at the Wave I interview. We then categorize years lived in the United States into four dummy variables: <6 years, 6 to 10 years, 11 to 14 years, and ≥15 years. We include second-generation adolescents in this variable by equating years in the United States with their age at Wave I, so all second-generation adolescents fall into either the third or fourth category (i.e., 11–14 and ≥15 years). Age of arrival is also categorized into four development periods: <6 years, 6 to 10 years, 11 to 14 years, and ≥15 years—reflecting preschool, elementary school, pre- and early adolescence, and adolescence stages. Because second-generation adolescents are born in the United States, they fall into the first category of arriving in the United States at age <6 years old.

Covariates of family, school, and neighborhood contexts

Contextual variables for the social acculturation environment include measures of the family, school, and neighborhood or community contexts. First, the family context that captures cultural and structural mechanisms includes five measures: (a) language spoken at home measured by three dummy variables for English, Spanish, and other language to reflect the differing acculturation levels between immigrant descendants and their foreign-born family members; (b) religiosity, which is measured by summing responses on how often the child attends church (responses range from 0 = no religion, 1 = never, to 4 = once a week or more) and on adolescent reports of the importance of religion (range from 0 = not at all to 4 = very important); (c) parental education; (d) family structure; and (e) the number of siblings. Add Health allows for rich detail on family living arrangements, classifying adolescents who live with two biological or adoptive parents, a biological parent (mainly the mother) and a stepparent, single mother, single father, and surrogate or foster parents (including grandparents, aunts and uncles, other adult relatives, or nonrelative adults). Parental education (the higher of the two parents if both are present) is measured as a set of dummy variables: less than high school, high school graduate, some college, college graduate, and missing parental education data.² Number of siblings is a count variable.

Measures for the context of Americanization at school and in communities include the region of the school (West, Midwest, South, or Northeast); school type (public or private); school size; percentages of students who are first, second, or third+ generation; percentages of non-Hispanic White students in the school; ethnic and immigrant presence in the neighborhood, which is measured by the proportion of foreign-born immigrants in the neighborhood; the proportion of residents aged 5 or older who speak English not well or not at all (defined by the census as “linguistically isolated”); the proportion of Hispanics in the neighborhood; and urban location. Neighborhoods are defined by census tracts in the Add Health contextual data.

Control variables

These variables include adolescent’s age, gender, and ethnic group background. Age is measured in single years. Ethnic group background is defined as a nine-category variable: Mexican, Cuban, Central and South American, Puerto Rican, Chinese, Filipino, Other Asian, African and Afro-Caribbean, and Canadian and European. Racial and ethnic backgrounds are self-identified by the respondent.³

Data analysis strategy

We begin with bivariate analysis of our measures of parent–child relationships according to our three acculturation measures, immigrant generation, length of stay, and age at arrival. In addition, we contrast first- and second-generation adolescents on our parent–child relationship measures within nine race and ethnic groups. These analyses allow us to establish whether patterns of intergenerational consonance and dissonance are consistent with our hypotheses about the effects of acculturation. We then move to multivariate analysis to control for the effects of age, gender, and race/ethnicity in assessing the effects of acculturation in a baseline model.

Using four estimation procedures, we conduct four separate multivariate analyses for each of the four dependent variables of parent–child relationships, including family cohesion, dinner with parents most of the days each week, parental control, and intergenerational conflict. Note that consonance is indicated by high levels of family cohesion and more weekly dinners together, and dissonance is indicated by greater parent–child conflict and greater parental control. Note further that low levels on each of these measures represent the reverse relationship (i.e., low levels of family cohesion represent greater dissonance and less consonance). After we estimate a baseline model, we then adjust for family, school, and community contextual factors to assess whether the significant relationships between the primary predictors and outcomes still hold or whether they would be attenuated by contextual measures. This involves entering in the set of cultural family context variables in a second model, the set of structural family context variables in a third model, and the set of school- and neighborhood-level contextual variables in a fourth model, then observing change in the acculturation effects from the baseline model.

Depending on the form of our dependent variable, we employ different estimation procedures. All multivariate analyses use sample survey methods, which account for the special features of the Add Health sampling design, including stratification, clustering, and sampling weights, to correct for biases in standard errors and significance tests if unweighted analyses are used (Chen & Harris, 2020). We use weighted ordinary least square (OLS) regressions for the dependent variable of family cohesion because it approximates continuous underlying distributions. We use weighted binary logistic regression for the estimation of the dependent variable of sharing dinners with parents during most of the days each week because it is a dichotomous variable. The mathematical formulation for the logistic model is:

$$\log (\frac{P}{1-P})=X’\beta$$

Because the results of the estimated coefficients, or betas, are not easily interpretable, we exponentiate the coefficients, e^β, to present odds ratios for more straightforward interpretation of results.

When the dependent variable is limited (unlike the common continuous variable) and measured by a number of events (usually equal or greater than zero), its underlying distribution is a Poisson distribution (Long, 1997). Because the dependent variable, parental control, is measured by the number of activities of the child over which the parents control (with a range between 0 and 7), we use weighted Poisson regression for the multivariate analysis. We choose Poisson regression over negative binomial regression here because the mean and variance of the dependent variable are equal. Thus, there is no need to use negative binomial regression, which is more appropriate when the mean of a dependent variable is not equal to its variance. The function for the Poisson regression can be written as:

$$log\left(numberofactionsforparentalcontrol\right)=X\prime \beta$$

Like binary logistic regression, we do not directly interpret the beta coefficients derived directly from the model. Instead, we use the exponentiated coefficients, e^β, for interpretation.

The fourth dependent variable, intergenerational conflict, takes the values of 0, 0.5, and 1, which resembles the form of an ordinal variable. We therefore use ordinal logistic regression for the estimation (Long, 1997). The function for ordinal logistic regression is expressed as:⁴

$$\log \frac{{\pi }_{ij}}{1-{\pi }_{ij}}=\log (\frac{{\pi }_{\text{greater}\text{conflict}}}{1-{\pi }_{\text{greater}\text{conflict}}})={\theta }_j+\beta ’X_i$$

Ordered logistic regression, known as the proportional odds model, has only one set of coefficient estimates with j intercepts, which distinguishes j sets of comparisons. In this particular case, there are two equations, each having the same coefficient estimates. Like binary logistic regression, we use the exponentiated coefficients, e^β, for interpretation. Exponentiated coefficients are odd ratios of being y = 1 versus y = 0 and y = 0.5, and odds ratios of being y = 1 & y = 0.5 versus y = 0.

RESULTS

Descriptive analyses

Supplemental Table 1 (see supplemental materials) presents our first set of descriptive analysis. On average, immigrant adolescents report moderately high levels of family cohesion (weighted average index is 11.26, where the range is from 2 to 15). Sixty percent (weighted) of immigrant adolescents eat most of their weekly dinners in the presence of one of their parents. On average, immigrant parents exercise complete control over about two of these activities (weighted mean = 2.07). Immigrant adolescents report relatively low levels of parent–child conflict (weighted mean = .28).

Results for the first two dependent variables of intergenerational relationships (shown in columns) by immigrant generations show a pattern of decreasing consonance (through the two measures of family cohesion and dinner with parents most of the days each week) and increasing dissonance across immigrant generations for all ethnic groups. The bivariate results are stronger and more consistent for family cohesion than for dinners. Thus, second-generation adolescents have lower levels of family cohesion and sharing weekly dinners with parents than their first-generation counterparts. Exceptions are noted for immigrant families from Central or South America and Africa and the Afro-Caribbean, where there are no differences by generation in the proportion who share weekly dinners.

Results for the measures that largely tap dissonance in intergenerational relationships (through parental control and intergenerational conflict) are generally consistent with our expectations. In six of the nine ethnic groups, parental control is lower for adolescents of the second-generation adolescents compared with the first-generation adolescents. Less parental control is exercised by parents of the second-generation adolescents for all race and ethnic groups except those from Mexico, Cuba, and Africa and the Afro-Caribbean. The overall mean difference by generation indicates that there is less parental control of adolescents in the second generation compared with the first generation (parental control scores of 2.03 and 2.14, respectively). Similarly, for six of the nine ethnic groups, intergenerational conflict is higher among the second-generation adolescents. The only ethnic group for which both dissonance measures do not operate in the expected direction is among those of Cuban heritage.

Supplemental Table 2 shows the bivariate association between intergenerational relationships (now shown in the rows) and age at arrival and length of stay. Again, results are consistent with the expectations. Earlier ages at arrival and longer stays in the United States are associated with less family cohesion, less sharing of weekly dinners, less parental control of adolescents’ activities, and greater parent–child conflict. The relationship is most consistent across each category on the acculturation measures for family cohesion. The relationship is also most salient by contrasting the lowest versus highest values of these two acculturation measures.

Multivariate analysis

Family cohesion

Table 1 shows the multivariate results of generational differences in family cohesion. Model 1 indicates that there is a significant effect of second generation such that there are lower levels of family cohesion in immigrant families when adolescents are U.S. born (compared with foreign-born adolescents in the first generation). Although we have little substantive interest in the effects of the other controls, our results are consistent with the previous literature, indicating that older children and females report less family cohesion (Acock & Demo, 1994; Maccoby & Martin, 1983). In addition, we see that Filipino and other Asian youth in immigrant families report significantly lower levels of family cohesion than immigrant youth from Europe and Canada.

TABLE 1.

Weighted ordinary least squares regression coefficients for generational differences in family cohesion

	Model 1	Model 2	Model 3	Model 4
Second generation	−0.607 (0.161)**	−0.497 (0.187)**	−0.520 (0.186)**	−0.408 (0.192)*
Age	−0.177 (0.041)**	−0.158 (0.040)**	−0.158 (0.039)**	−0.098 (0.043)*
Female	−0.418 (0.149)**	−0.470 (0.150)**	−0.445 (0.148)**	−0.450 (0.144)**
Ethnicity (Europe/Canada)
Mexico	−0.015 (0.199)	−0.252 (0.244)	−0.165 (0.247)	−0.183 (0.260)
Cuba	−0.009 (0.462)	−0.117 (0.410)	−0.070 (0.368)	−0.751 (0.319)*
Central-South America	−0.042 (0.290)	−0.250 (0.295)	−0.149 (0.303)	−0.324 (0.329)
Puerto Rico	−0.026 (0.263)	−0.168 (0.259)	0.004 (0.271)	−0.038 (0.318)
Chinese	−0.456 (0.330)	−0.408 (0.331)	−0.511 (0.336)	−0.256 (0.314)
Philippine	−0.569 (0.239)*	−0.768 (0.245)**	−0.799 (0.245)**	−0.542 (0.315)
Other Asia	−0.695 (0.321)*	−0.818 (0.323)*	−0.871 (0.322)**	−0.827 (0.325)*
Africa/Afro-Caribbean	0.021 (0.304)	−0.018 (0.304)	0.082 (0.310)	0.076 (0.294)
Language spoken at home (English)
Spanish		0.251 (0.249)	0.253 (0.264)	0.060 (0.241)
Other language		0.293 (0.235)	0.253 (0.229)	0.165 (0.231)
Religiosity		0.138 (0.030)**	0.129 (0.030)**	0.132 (0.029)**
Parents’ education (less than high school)
High school graduate			0.086 (0.181)	0.151 (0.179)
Some college			0.114 (0.245)	0.171(0.238)
College graduate			0.132 (0.225)	0.242 (0.228)
Missing			0.106 (0.420)	0.126 (0.412)
Family structure (two biological or two adopted parents)
One stepparent + one biological parent			−0.694 (0.152)**	−0.744 (0.146)**
Single mom			−0.450 (0.159)**	−0.474 (0.148)**
Single dad			0.015 (0.339)	0.100 (0.339)
Two stepparents or other types			−0.030 (0.224)	−0.007 (0.247)
Number of siblings			−0.051 (0.042)	−0.032 (0.041)
Region of school (south)
West				−0.115 (0.201)
Midwest				−0.487 (0.229)*
Northeast				−0.063 (0.248)
Public school				0.489 (0.378)
School size (small: 1–400)
Medium (400–1,001)				−0.255 (0.374)
Large (1,001–4,000)				−0.739 (0.427)
Percentage of second generation at school				−0.010 (0.023)
Percentage of third generation at school				−0.005 (0.013)
Percentage of White students at school				0.006 (0.005)
Urban				0.325 (0.197)
Proportion of foreign-born within the community				−0.291 (1.248)
Proportion aged individuals ≥5 who speak English “not well” or “not at all” in community				0.384 (2.356)
Proportion of Hispanics in community				0.894 (0.818)
Constant	15.409 (0.723)**	14.188 (0.768)**	14.413 (0.717)**	13.256 (1.900)**
R 2	0.04	0.05	0.07	0.08
N	3,925	3,925	3,925	3,925

Note. Coefficients with standard errors in parentheses.

^*p < .05.

^**p < .01.

^***p < .001.

Model 2, which controls for the family cultural variables of language spoken at home and religiosity shows that the coefficient for religiosity is significant indicating a positive association with family cohesion. The coefficient for language spoken at home across most parent–child outcomes is not significant, probably because our acculturation variables (e.g., generation, age at arrival, and length of stay) serve as proxies for language spoken at home (Perreira et al., 2007). Our main interest, however, is to evaluate whether the size and significance of the coefficient for immigrant generation change when this set of variables is added to the model. We note some attenuation because the size of the coefficient for the generation measure decreases by about 20% but remains significant.

In Model 3 of Table 1, structural variables play no role in attenuating the effects of immigrant generation because its effect remains strong and even increases slightly. Family structure is the only structural factor that influences the differences in family cohesion, such that family cohesion is lower in stepfamilies and single-mother families relative to two biological or adoptive parent families. In Model 4, we add the full set of school- and neighborhood-level variables to assess whether the coefficient for the immigrant generation measure would change. None of the contextual variables are important predictors of family cohesion (except for some regional variation) because their coefficients are not significant. Acting together, however, the full set of school and neighborhood contextual variables explains some of the generation effect, with the coefficient decreasing by about 20% in Model 4. The result also indicates that overall, our family, school, and neighborhood mechanisms tend to explain only a small proportion of the acculturation effects on family cohesion, as the coefficient for the second-generation measure in Model 1 decreases by about 30% in Model 4.

In Supplemental Table 3, we show the results of the same four models of family cohesion for the effects of age at arrival to the United States. The baseline model shows a significant effect of age at arrival. Immigrant adolescents who arrived in the United States when they were less than 6 years old report lower levels of family cohesion by more than 1 point on the additive index compared with adolescents who arrived more recently when they were aged 15 years or older. The effect for arrival in middle childhood, between ages 6 and 10, also indicates lower levels of family cohesion by almost 1 point on the index. Although the less than 6 years category includes second-generation adolescents and its effect is consistent with the generation models in Supplemental Table 3, we still find an important “linear” effect of acculturation as measured by age at arrival at age 6 to 10 relative to more recent arrivals in the first generation. Thus, age at arrival allows us to capture acculturation effects even among foreign-born adolescents. Similar to Table 1, the addition of family cultural variables reduces the acculturation effects of age at arrival slightly in Model 2, but not at all in Model 3 when family structural factors are added and hardly at all in Model 4 with the addition of school and neighborhood factors.

Sharing of weekly dinners

Supplemental Table 4 shows the results for the logistic regression of generational differences in sharing weekly dinners. Focusing on the effect of immigrant generation, we again find that second-generation adolescents are less likely to share most of the weekly dinners with their parents. Interpreting the odds ratio (second number in each column) indicates that the odds of sharing most of the weekly dinners is 32% lower for second-generation adolescents compared with their first-generation counterparts (i.e., 1.0–.68). Interestingly, there are more significant ethnic group effects for dinners, all indicating that adolescents in the various ethnic groups are less likely to share dinners with their parents than European and Canadian youth in immigrant families.

When we enter family cultural factors in Model 2, there is little attenuation of the significant generation effect, and the effect for the acculturation factor increases somewhat when family structural factors are entered in Model 3. The addition of school and neighborhood factors in Model 4 provides little explanatory power as well. Results for these covariates are similar to those in the models with the other consonance measure of family cohesion. That is, religiosity increases and nonintact family structures decreases the sharing of weekly dinners. There are other significant contextual effects indicating that sharing dinners with parents is more common in Hispanic communities and less common when adolescents attend large schools. We present only this one analysis for dinners because neither age at arrival nor length of stay was strongly associated with this outcome.

Parental control

Table 2 presents generational differences in parental control. It is important to note, first, that immigrant parents of almost all the ethnic groups other than Europe and Canada exercise more parental control, indicative of traditional parental behavior in the sending countries. Hispanic parents tend to use more parental control than Asian parents. Controlling for ethnic background, we find the expected effect that immigrant parents exercise less parental control over the second-generation adolescents than their first-generation counterparts. This suggests that U.S.-born children in immigrant families have greater autonomy and freedom to make their own decisions about aspects and activities in their lives, an adaptation experience that can likely be ascribed to greater acculturation of children relative to their parents in the second generation. The effect is not big, reducing the count on parental control by only 8%. When we add family cultural factors in Model 2, this small effect is attenuated and is no longer significant. This effect is likely due, in part, to the greater religiosity of families of first-generation adolescents, which reaffirms parental authority and parental respect compared with families that are less religious. In addition, speaking Spanish at home is associated with greater parental control, which is marginally significant at the .07 level.

TABLE 2.

Coefficients and exponentiated coefficient (e^β in ratios) of weighted Poisson regression for generational differences in parental control

	Model 1	Model 2	Model 3	Model 4
Second generation	−0.086/0.918 (0.041)*	−0.056/0.946 (0.043)	−0.055/0.946 (0.044)	−0.057/0.945 (0.041)
Age	−0.131/0.878 (0.015)**	−0.126/0.881 (0.014)**	−0.129/0.879 (0.014)**	−0.120/0.887 (0.015)**
Female	0.051/1.052 (0.035)	0.039/1.040 (0.036)	0.036/1.037 (0.036)	0.037/1.037 (0.035)
Ethnicity (Europe/Canada)
Mexico	0.310/1.364 (0.067)**	0.255 /1.290 (0.075)**	0.200/1.222 (0.080)*	0.062/1.064 (0.078)
Cuba	0.104/1.109 (0.057)	0.075/1.078 (0.062)	0.089/1.093 (0.065)	0.063/1.065 (0.127)
Central-South America	0.268/1.308 (0.086)**	0.221/1.247 (0.077)**	0.220/1.247 (0.077)**	0.169/1.185 (0.078)*
Puerto Rico	0.172/1.188 (0.076)*	0.138/1.148 (0.074)	0.148/1.159 (0.076)	0.078/1.082 (0.088)
Chinese	−0.002/0.998 (0.111)	−0.008/0.992 (0.107)	−0.004/0.996 (0.103)	−0.008/0.992 (0.102)
Philippine	0.097/1.102 (0.072)	0.047/1.049 (0.070)	0.055/1.057 (0.071)	−0.053/0.948 (0.086)
Other Asia	0.084/1.088 (0.092)	0.042/1.043 (0.092)	0.041/1.041 (0.091)	0.020/1.020 (0.088)
Africa/Afro-Caribbean	0.223/1.249 (0.120)	0.217/1.242 (0.119)	0.235/1.264 (0.115)*	0.173/1.189 (0.103)
Language spoken at home (English)
Spanish		0.068/1.070 (0.047)	0.034/1.034 (0.046)	0.023/1.023 (0.049)
Other language		0.103/1.108 (0.072)	0.067/1.069 (0.070)	0.066/1.068 (0.075)
Religiosity		0.032/1.033 (0.010)**	0.031/1.031 (0.009)**	0.029/1.029 (0.009)**
Parents’ education (less than high school)
High school graduate			−0.076/0.927(0.049)	−0.054/0.947 (0.046)
Some college			−0.103/0.902 (0.070)	−0.092/0.912 (0.071)
College graduate			−0.116/0.890 (0.055)*	−0.088/0.916 (0.053)
Missing			−0.009/0.991 (0.078)	−0.009/0.991 (0.076)
Family structure (two biological or two adopted parents)
One stepparent + one biological parent			−0.059/0.943 (0.055)	−0.068/0.934 (0.057)
Single mom			−0.098/0.906 (0.053)	−0.097/0.907 (0.056)
Single dad			−0.262/0.769 (0.126)*	−0.281/0.755 (0.125)*
Two stepparents or other types			0.024/1.025 (0.132)	0.008/1.008 (0.129)
Number of siblings			0.023/1.023 (0.014)	0.024/1.025 (0.015)
Region of school (South)
West				−0.058/0.943 (0.068)
Midwest				−0.269/0.764 (0.079)**
Northeast				−0.038/0.963 (0.070)
Public school				0.213/1.238 (0.093)*
School size (small: 1–400)
Medium (400–1,001)				−0.188/0.829 (0.076)*
Large (1,001–4,000)				−0.212/0.809 (0.095)*
Percentage second generation at school				0.007/1.007 (0.006)
Percentage third generation at school				0.007/1.007 (0.003)*
Percentage White students at school				−0.001/0.999 (0.001)
Urban				0.054/1.056 (0.051)
Proportion of foreign-born in community				0.643/1.903 (0.459)
Proportion of individuals aged ≥5 who speak English “not well” or “not at all” in community				−1.192/0.303 (0.884)
Proportion of Hispanics in community				0.402/1.494 (0.271)
Constant	2.678 (0.246)**	2.381 (0.253)**	2.513 (0.254)**	1.763 (0.413)**
N	3,925	3,925	3,925	3,925

Note. Coefficients and odds ratios with standard errors in parentheses.

^*p < .05.

^**p < .01.

^***p < .001.

There is no change to the results for the acculturation measure in Models 3 or 4. Several family and contextual factors are significant for levels of parental control, however. For example, greater parental education is associated with less parental control, and single fathers exercise less control over children than parents in two biological families. Private schools and small schools reduce the need for parental control of children because these contexts serve to enforce community norms and expectations. Finally, when a school is largely made up of third-generation adolescents, parental control is weakened by the school climate of mainstream norms for adolescent autonomy from parents. Supplemental Table 5 additionally presents the results for parental control using length of stay as the acculturation variable. Here we find strong support for our acculturation hypothesis. With increasing length of stay in the United States, immigrant parents tend to exercise less parental control over their children’s activities. The addition of family context in Model 2 reduces this effect somewhat, but still indicates that parents exercise greater control over immigrant children in families that have recently migrated to the United States.

Intergenerational conflict

Table 3 shows generational differences in our final parent–child outcome on intergenerational conflict. We again find a significant effect for second-generation adolescents who experience greater conflict and arguments with parents regarding their behavior than the first generation. Second-generation adolescents have 35% higher odds of experiencing high parent–child conflict compared with the first generation. Consistent with prior results, family cultural context reduces this effect so that it is no longer significant in Model 2. Here language spoken at home is important, and an added dimension of acculturation appears to be operating. When language spoken at home is Spanish, intergenerational conflict is less than when English is spoken at home. That this finding is in the same direction as our other acculturation variables lends additional support to our argument that generation status and time in the United States are capturing acculturation. The addition of family structural measures in Model 3 and school and neighborhood contextual measures in Model 4 continue to attenuate the generation effect somewhat but are not nearly as salient as the family cultural factors in attenuating the generation effects in parent–child conflict. Family structure differences play a role at varying levels of parent–child conflict.

TABLE 3.

Coefficients and odds ratios of weighted ordered logistic regression for generational differences in intergenerational conflict

	Model 1	Model 2	Model 3	Model 4
Second generation	0.303/1.354 (0.109)**	0.218/1.243 (0.113)	0.178/1.194 (0.113)	0.135/1.145 (0.120)
Age	0.010/1.010 (0.029)	0.015/1.016 (0.030)	0.028/1.029 (0.029)	0.025/1.025 (0.035)
Female	0.260/1.297 (0.104)*	0.264/1.302 (0.104)*	0.240/1.271 (0.107)*	0.229/1.258 (0.108)*
Ethnicity (Europe/Canada)
Mexico	0.102/1.108 (0.150)	0.247/1.280 (0.199)	0.358/1.430 (0.223)	0.638/1.893 (0.248)*
Cuba	−0.182/0.834 (0.282)	−0.025/0.975 (0.272)	−0.143/0.876 (0.278)	0.337/1.400 (0.245)
Central-South America	0.079/1.083 (0.165)	0.202/1.224 (0.195)	0.117/1.124 (0.215)	0.366/1.442 (0.221)
Puerto Rico	0.530/1.698 (0.259)*	0.581/1.787 (0.254)*	0.457/1.579 (0.261)	0.693/2.000 (0.244)**
Chinese	0.425/1.530 (0.326)	0.540/1.716 (0.338)	0.601/1.824 (0.340)	0.683/1.979 (0.342)*
Philippine	0.273/1.314 (0.196)	0.240/1.271 (0.204)	0.262/1.299 (0.208)	0.399/1.490 (0.247)
Other Asia	0.102/1.107 (0.180)	0.153/1.165 (0.182)	0.210/1.234 (0.184)	0.357/1.429 (0.189)
Africa/Afro-Caribbean	−0.232/0.793 (0.278)	−0.265/0.767 (0.282)	−0.324/0.723 (0.276)	−0.113/0.893 (0.286)
Language spoken at home (English)
Spanish		−0.324/0.724 (0.153)*	−0.316/0.729 (0.163)	−0.244/0.784 (0.149)
Other language		−0.265/0.767 (0.168)	−0.229/0.795 (0.176)	−0.238/0.788 (0.177)
Religiosity		0.024/1.024 (0.020)	0.033/1.034 (0.020)	0.034/1.034 (0.021)
Parents’ education (less than high school)
High school graduate			−0.013/0.987 (0.180)	−0.072/0.930 (0.178)
Some college			0.404/1.498 (0.205)	0.378/1.460 (0.207)
College graduate			−0.022/0.979 (0.184)	−0.078/0.925 (0.182)
Missing			−0.305/0.737 (0.260)	−0.279/0.756 (0.253)
Family structure (two biological or two adopted parents)
One stepparent + one biological parent			0.286/1.331 (0.169)	0.326/1.386 (0.175)
Single mom			0.397/1.487 (0.176)*	0.456/1.578 (0.170)**
Single dad			−1.020/0.361 (0.356)**	−1.005/0.366 (0.364)**
Two stepparents or other types			−0.577/0.562 (0.269)*	−0.591/0.554 (0.263)*
Number of siblings			−0.059/0.943 (0.038)	−0.064/0.938 (0.039)
Region of school (South)
West				0.036/1.037 (0.120)
Midwest				0.369/1.446 (0.187)
Northeast				0.068/1.070 (0.191)
Public school				0.130/1.138 (0.287)
School size (small: 1–400)
Medium (400–1,001)				0.024/1.024 (0.232)
Large (1,001–4,000)				0.005/1.005 (0.241)
Percentage second generation at school				0.018/1.018 (0.018)
Percentage third generation at school				0.001/1.001 (0.008)
Percentage White students at school				0.005/1.005 (0.003)
Urban				−0.258/0.772 (0.140)
Proportion of foreign born in community				0.713/2.041 (0.840)
Proportion of individuals aged ≥5 who speak English “not well” or “not at all” in community				−1.209/0.299 (1.731)
Proportion of Hispanics in community				−0.622/0.537 (0.619)
Intercept 1	1.419** (.508)	1.430** (.537)	1.606** (.572)
Intercept 2	2.256** (.508)	2.270** (.535)	2.462 (.573)
N	3,925	3,925	3,925	3,925

Note. Coefficients and odds ratios with standard errors in parentheses.

^*p < .05.

^**p < .01.

^***p < .001.

In Supplemental Table 6, we examine the effects of age at arrival in relation to intergenerational conflict. The findings reveal an almost linear effect whereby the earlier children in immigrant families arrive in the United States, the greater parent–child conflict is experienced by adolescents. The earlier the age at arrival, the longer the time the child is exposed to American norms and attitudes with peers and in schools and neighborhoods. Again, this effect is reduced somewhat by speaking Spanish at home among those with the latest age of arrival (≥15 years old) relative to those with the earliest age at arrival. This is the only outcome for which entering in family structural context in Model 3 attenuates some of the effects in age at arrival (although all remain significant). Family structure defines a context for more or less conflict whereby more intergenerational conflict occurs in stepfamilies and single-mother families and less in single-father families compared with two biological parent families. School and neighborhood contextual measures play no role in attenuating the acculturation effects of parent–child conflict.

DISCUSSION

This study examined acculturation processes in parent–child relationships among adolescents of immigrant families. Findings indicate that second-generation adolescents, adolescents who arrived in the United States at a younger age, and adolescents who have a longer length of stay experienced greater acculturation than their counterparts. When adolescents experience greater acculturation into U.S. society, parent–child relationships are characterized more by dissonance than by consonance. When adolescents’ acculturation is similar to the acculturation of their parents, consonance in parent–child relationships is more evident.

Our findings are robust for the hypothesis regarding immigrant generation. Across all measures of dissonance and consonance in parent–child relationships, we find empirical support that confirms our hypothesis in that second-generation adolescents experience less consonance and more dissonance. That is, second-generation adolescents report lower levels of family cohesion, less often sharing weekly dinners with parents, less parental control of adolescents’ activities, and greater intergenerational conflict than adolescents of the first immigrant generation. Except for parental control, these effects remain significant even when we adjust for family, school, and community contextual factors with other potential effects for differences in parent–child relationships.

Our findings for age at arrival and length of stay are less robust but do support our hypotheses. We find important acculturation effects of age at arrival for differences in family cohesion and intergenerational conflict, as well as important acculturation effects of length of stay in the United States for differences in parental control. Note that length of stay effects were also evident for family cohesion and intergenerational conflict and age at arrival effects for parental control, but they were not as important as the ones we show.

Importantly, our findings on age at arrival and length of stay provide additional strong support for our hypotheses that dissonance and consonance in parent–child relationships are due to differential acculturation by children and parents because these effects represent increasing degrees of exposure to American society by adolescents of both the second and the first generation. For first-generation adolescents, each additional year of exposure to American society changes their relationships with parents. Because we have hypothesized that foreign-born children’s exposure to American society represents a more rapid acculturation (in Piaget’s [1936] terms, experiencing more accommodation than assimilation) than foreign-born parents’ exposure, we attribute the acculturation effect to children’s adaptation experiences. This is consistent with the common experience of immigrant descendants who must attend school every day where they can only speak English and socialize with other schoolmates, the majority of whom are native-born, whereas parents are not forced into such interactions with other natives in their surroundings and can live a rather isolated life, especially foreign-born mothers who do not work outside the home.

Overall, findings of this study resonate and extend prior research on intergenerational relationships in immigrant families. This research adds new empirical evidence of national representation about immigrant families of multiple Hispanic and Asian racial/ethnic groups by employing enriched acculturation and family measures that capture nuanced and multidimensional pictures for the linkages between the acculturation processes and outcomes of parent–child relationships. Although previous research found socialization gap in educational outcomes (e.g., Areepattamannil & Lee, 2014; Bankston, 2004; Fuligni, 1997) and parent–child conflict (e.g., Basáñez et al., 2014; Dennis et al., 2010; De Santis et al., 1995; Kao & Tienda, 1995; Pasch et al., 2006; Ying & Han, 2007) in family relationships, this work expands the scope of immigrant family studies to a new level that links family cohesion, parental control, and shared family time together with the adaptation through age of arrival and length of stay in which not just second-generation children of immigrant but also those of the first generation with varying acculturation levels would diverge in their adaptation experiences into the receiving country.

Limitations and directions for future research

This study has some limitations. First, although it uses enriched multiple domains (including family cohesion, mealtime together, parental control, and intergenerational conflict) to operationalize parent–child relationships in immigrant families, they may not encompass a complete spectrum of intergenerational mechanisms that apply to every subgroup cultural setting. For example, cultural values and practices in Mexican, Cuban, and Puerto Rican families of Hispanic heritage may vary in nuances that may have slightly different definitions of the meanings of family cohesion, parental control, or conflict. Future studies can develop more culturally nuanced measures to understand each subgroup’s intergenerational relationships so that targeted interventions can be more ethnically suited.

Second, because this study’s focus was on the associations between acculturation levels of the parent and adolescent and family relationships, limited attention was given to the mechanisms through which these associations are linked, although it adjusts for a series of family, school, and community/neighborhood factors setting these immigrant families to assess their attenuating effects for parent–child relationships at the macrostructural/contextual level of the social environment than merely at the individual level. Future research could collect both qualitative and quantitative, in-depth data to understand the processes through which macro-acculturation contexts operate to increase intergenerational dissonance and implement strategies to improve parent–child relationships between the second-generation and more acculturated children of the first generation and their immigrant parents. For example, how can positive aspects of family cohesion and adolescent well-being be maintained when undocumented parents are increasingly at risk of deportation due to the current immigration law? Additional research is also needed to understand the stress specific to immigrant family adaptation when they struggle for survival and social mobility while overcoming the mounting language and cultural barriers and discrimination so that policy-related sources can be adequately and sufficiently allocated to improve the well-being of immigrants and their descendants.

Implications for practice

The findings from this study confirm the importance of identifying and addressing sources of intergenerational dissonance among immigrant families of second-generation adolescents and immigrant descendants who stayed longer or arrived earlier in the United States. Some exemplary programs, such as Bicultural Effectiveness Training (Szapocznik & Kurtines, 1993) and Bridges/Puentes (Gonzales et al., 2007, 2012), are promising interventions to reduce parent–child conflict in Cuban American families by increasing adolescents’ knowledge of and appreciation for their own ethnic culture and heritage (Szapocznik & Kurtines, 1993) and engaging parents and adolescents in culturally specific activities (Gonzales et al., 2007, 2012).

Family therapeutic programs in which family psychologists who are knowledgeable with multicultural settings and skilled in understanding intergenerational conflict or acculturation dissonance in immigrant families can provide culturally appropriate services to improve parent–child relationships in those families, such as the ones implemented for Asian American families (Cheung, 2021; Chung & Shibusawa, 2013; Hong & Ham 2001). As our study indicates, some subgroup differences among Hispanic and Asian immigrant families, future intervention programs can investigate in-depth subgroup cultural differences to address specific needs of immigrant families of each ethnic group and train family therapists with cultural/ethnic awareness to improve the immigrant family relationships (Maldonado-Molina et al., 2006; Muir et al., 2004). Furthermore, instead of focusing on resolving conflicts, family programs can be developed to educate clients through strategies to improve family cohesion, shared family time together, and parental practices of how to exercise appropriate parental control, as indicated by the findings of this study.

Conclusion

Better understanding of acculturation processes among parents and children in immigrant families may require more in-depth qualitative or quantitative research for individual immigrant ethnic groups, a hallmark of immigrant research. Our research shows that parent–child outcomes can vary by ethnic group, and differences across ethnic groups also vary by the outcome. Thus, ethnic group differences may or may not be important depending on the aspect of intergenerational relationships under study. The purpose of our research, however, was to address the theoretical relevance of notions about consonance and dissonance in parent–child relationships in immigrant families, controlling for ethnic group background across an array of parent–child outcomes using national data. Our contribution provides substantial evidence regarding patterns of intergenerational consonance and dissonance that are consistent with hypotheses about differential acculturation by parents and adolescents in immigrant families. This is the first evidence of this kind that is based on national data, across multiple measures of intergenerational relationships, and that holds for a diverse and representative set of ethnic groups, setting the stage for further exploration into acculturation processes among parents and children in immigrant families.