Immigrant Generations and Delinquency

Abstract

Recent research has demonstrated an “immigrant paradox”: a parallel between assimilation and increased engagement in adolescent delinquency. Although evidence has suggested that social control and differential association contribute to the criminological understanding of the immigrant paradox, not many studies have simultaneously incorporated measures of social control and differential association to explore delinquency across immigrant generations. Based on a sample of nationally representative adolescents (N = 13,121) from the National Longitudinal Study of Adolescent Health (Add Health), this study investigates whether and how social control (i.e., family control and school control) and differential association (i.e., delinquent friends association) explain the relationship between immigrant generation and delinquency. Results indicate that delinquent friendships play a more important role and account for much of the difference in delinquency across generations. Moreover, within each generation, delinquent friendships have a more consistent effect on delinquency. However, family control and school control are stronger predictors of delinquency for the third-plus generation than for earlier generations.

Keywords

immigrant youth delinquency family school association with delinquent friends

Introduction

Following the rapid growth in the number of immigrants in the United States in the past few decades, understanding how well immigrants and their children adapt to the host society has generated tremendous research interest. The relationship between immigration and crime is a focal concern of this research. Conventional theories of crime and delinquency have typically argued that immigrants, because of the barriers to assimilation into American life and the prevalence of crime in resettlement communities, tend to be part of a crime-prone population (Miller, 1958; Shaw & McKay, 1942/1969; T.Waters, 1999; Wolfgang & Ferracuti, 1967). However, quantitative evidence in research literature has revealed that immigrants engage in crime and delinquency at a lower rate than their native-born counterparts (Lee, Martinez, & Rosenfeld, 2001; Martinez, Lee, & Nielsen, 2004; Martinez, Stowell, & Lee, 2010; Reid, Williams, Harald, Adelman, & Jaret, 2005; Sampson & Bean, 2006; Stowell, Messner, McGeever, & Raffalovich, 2009). Recent studies have also indicated that crime and delinquency rates are relatively low among first-generation immigrants but gradually increase with each succeeding generation (Bersani, 2013; Martinez & Lee, 2000; Morenoff & Astor, 2006; Reid et al., 2005; Sampson, 2008). This counterintuitive parallel between assimilation and inclining rate of criminal outcomes is known as the “immigrant paradox” in crime (Lee & Martinez, 2006).

The increasing rate of delinquent behavior across immigrant generations has led researchers to identify the characteristics that place later-generation immigrant youth at a greater risk for delinquency. This body of research suggests that the characteristics of social institutions, such as family control, school control, and association with delinquent friends, shape the patterns of involvement in delinquency across immigrant generations (Bui, 2009; Dipietro & McGloin, 2012; Myers et al., 2009; Peguero, Bondy, & Hong, in press; Peguero & Jiang, 2014). Specifically, the research attributed generational variations in delinquency to a deterioration of parental and school control and greater exposure to delinquent friends. However, much attention in these previous studies has been on the predictors in a single context (i.e., family control, school control, or delinquent friends). No research has examined the independent influence of these characteristics through an analysis of delinquency across immigrant generations. It may be true that family control, school control, and delinquent friends combine to affect generational change in development and behaviors. For instance, parents are known to provide supervision and suggestions on their children’s choice of friends (Dillon et al., 2008). Thus, youth of earlier immigrant generations who have been under strong parental control may be less likely to associate with delinquent friends, and thus be less likely to engage in delinquency than those of later generations. As a result, variables from a single context that account for generational differences in delinquency may prove insignificant in a more comprehensive examination.

Furthermore, despite the increasing attention on family control, school control, and friends’ delinquency in predicting immigrant delinquency, there is no consensus as to whether or not there are generational differences. Some researchers have applied the same measures to predict delinquency across immigrant generation under the assumption that the risk factors for conducting delinquent behaviors within all generations are similar (e.g., Bui, 2009). However, a few studies have shown that school control and association with delinquent friends differentially influence youth across immigrant generation on delinquency and other developmental outcomes (e.g., Dipietro & McGloin, 2012; Jiang & Peterson, 2012; Peguero, 2013). The inconsistent evidence suggests that immigrants experience familial and educational processes and develop friendships that are unique or specific to their immigrant status. Thus, variations in delinquent behaviors within each generation may not be explained by the same predictors and mechanisms. If immigrant generation conditions how family control, school control, and delinquent friends influence youth delinquent outcomes, immigrant-specific theories or measurements must be developed to incorporate the unique factors that contribute to delinquency among immigrants. However, few studies have systematically and simultaneously modeled the effects of these predictors from different social contexts or investigated the relative impact of these predictors on delinquent behaviors for each immigrant generation.

In response to these research limitations, this study examines the independent effect of family control, school control, and association with delinquent friends on adolescent delinquency across immigrant generations. Our study is focused on investigating three outcomes: violent delinquency, nonviolent delinquency, and substance use. Much of the research on the relationship between conventional control (e.g., family control and school control), association with delinquent friends, and delinquency has its root in Hirschi’s social control theory and Sutherland’s differential association theory. Thus, we draw insights from these two leading criminological theories as frameworks to explore the importance of family control, school control, and association with delinquent friends in explaining generational differences in delinquent outcomes. First, we assess, within a single model, the extent to which generational differences in family control, school control, and association with delinquent friends contribute to generational differences in delinquent outcomes. Second, we assess whether the effect of family control, school control, and association with delinquent friends differ within each immigrant generation. Clarifying the unique contributions of these social contexts (family, school, and friends) to delinquency across immigrant generation will allow scholars and policy makers to formulate informed theoretical accounts of the immigration–crime nexus and make more efficient delinquency reduction policies for today’s immigrant youth. We now review the theories that explain how family, school, and delinquent friends may influence the risk of delinquent behaviors.

Social Control Theory, Family Control, and School Control

Social control theory posits that individuals are inherently inclined to be deviant, and conformity, rather than deviance, must be explained (Akers & Sellers, 2004; Hirschi, 1969). Assuming the intrinsic nature of delinquency, social control theory suggests that a strong bond to conventional social institutions offers some protection against deviant inclinations. Social bonds become weak or broken when an individual fails to (1) possess emotional attachment to significant others; (2) develop a stake in conformity to long-term educational, occupational, or other conventional goals; (3) engage in conventional activities as opposed to delinquent activities; and (4) believe in the moral validity of the law (Hirschi, 1969).

Social control theory has been used to explain how the connection between individuals and conventional social institutions produces variations in the rate of deviance. In the context of juvenile delinquency, bonds to family and school are especially salient for understanding youth delinquent behaviors (Hirschi, 1969; Nye, 1957). Family control operates mainly through children’s attachment to parents, through their commitment to parental aspirations, and through parents’ monitoring of their children’s behaviors. Children who feel attached to and connected to their parents are more likely to value their relationship and to distance themselves from delinquent behaviors that would disappoint their parents (Agnew, 2001; Hirschi, 1969). Similarly, children who are committed to the prosocial values and goals taught by their parents are likely to avoid activities that jeopardize goal achievement (Wright & Cullen, 2001). In addition, parents who are vigilant about keeping track of their children’s whereabouts, companions, and activities can control their children’s behaviors because children believe their parents know what they are doing and expect them to act appropriately (Dorius, Bahr, Hoffmann, & Harmon, 2004). There is an abundance of empirical research indicating that these aspects of family control have a consistent association with delinquency (Booth, Farrell, & Varano, 2008; De Kemp, Scholte, Overbeek, & Engels, 2006; Demuth & Brown, 2004; Deutsch, Crockett, Wolff, & Russell, 2012).

The school provides another central venue for social control. The school is where children build strong social bonds by forming attachments to prosocial figures. School also provides many conventional activities that keep children occupied, so that they have little time to engage in delinquent or other unstructured activities. Bondings to the school, including connection to teachers and prosocial peers, commitment to educational goals, participation in school-related activities, and acceptance of established norms for school behavior, function as protective factors against delinquency (Hawkins & Weis, 1985; Jenkins, 1997). As suggested by social control theory, children who maintain a sense of belonging to school, who have high educational aspirations, who engage in a variety of school activities, and who believe in school rules are more likely to refrain from delinquent behaviors. Although the magnitude in the effect of school bonds has varied across studies, consistent with social control theory, research has provided considerable empirical evidence on the importance of school bonds in understanding delinquency (Cernkovich & Giordano, 1992; Gottfredson, Gottfredson, Payne, & Gottfredson, 2005; Stewart, 2003; Wright & Cullen, 2001).

Differential Association Theory and Association With Delinquent Friends

Differential association theory assumes that delinquent behavior, like most other types of behavior, is learned through socialization. Central to differential association theory is the idea that delinquency is learned within intimate personal groups whose members acquire social definitions that favor the violation of legal code (Sutherland, 1947). When the frequency, duration, priority, and intensity of exposure to delinquency definitions increase to an extent that definitions favorable to violation of law outweigh the definitions favorable to prosocial behaviors, individuals are more likely to adopt these definitions and use them to justify delinquent behaviors as acceptable and appropriate (Sutherland, 1947).

Although differential association theory and Akers’s (1985) later extension to it suggest that various mechanisms are involved in the learning process, the primary way is in direct interactions with significant others, especially delinquent friends. The relationship between friends’ delinquency and individuals’ own delinquency is one of the most consistent and robust criminological findings. Research has provided consistent evidence that association with friends who are involved in risk-taking behaviors is a strong predictor of status offenses, self-reported delinquency, violence, and substance abuse (Akers, Krohn, Lanza-Kaduce, & Radosevich, 1979; Dishion, Eddy, Haas, Li, & Spraclden, 1997; Elliott, Huizinga, & Ageton, 1985; Haynie, 2001; Jensen, 1972; Matsueda & Heimer, 1987; Nofziger & Lee, 2006; Nye & Short, 1957; Vitaro, Brendgen, & Tremblay, 2000; Warr & Stafford, 1991).

Social Control, Delinquent Friends, and Immigrant Generation

According to the statistics reported by the Center for Immigration Studies, new immigration plus births to immigrants added more than 22 million people to the U.S. population in the last decade, equal to 80% of total population growth (Camarota, 2012). A significant segment of these immigrants and their children are comprised by Hispanic and Asian populations from different cultural and socioeconomic background. Given the prevalence and diversity of today’s immigrants, understanding determinants of delinquent behaviors among immigrant youth in the United States is a central concern in criminological research. When explaining immigrant paradox in delinquency, the widely studied social contexts for immigrant youth’s assimilation have been the family, the school, and the friends. Many researchers have suggested that there are important differences across immigrant generations in their socialization processes in the family, the school, and the friends’ group, which explain why the risk of delinquency increases with each subsequent immigrant generation (e.g., Bui, 2009; Le & Stockdale, 2008). These ideas led to a prediction of an indirect effect of immigrant generation on delinquency. In particular, immigrant generation is expected to affect delinquency because it affects the strength of social bonds in family and school and exposure to delinquent friends. Following the prediction, although immigrant youth were found to endorse a great expectation of mutual support and remain bounded to their family (Berry, 2006; Driscoll, Russell, & Crockett, 2008; Kao, 2004; Waters, 1999), this strong family bond may be attenuated, as immigrant youth adopt norms and values of host society that clash with those held by their less acculturated parents, thereby increasing the risk of problematic behaviors in adolescence (Portes & Zhou, 1993). Such assimilation process, along with the deterioration of residential communities, and the hostilities encountered by many immigrant youth in U.S. schools may also lead to weakened attachment to teachers and schools and heightened exposure to delinquent friends’ influence across immigrant generations (Myers et al., 2009).

To date, consistent with these perspectives, research evidence showed that family bond, school bond, and delinquent friends explain the link between immigrant generation and problematic behaviors among youth (Dipietro & McGloin, 2012; Myers et al., 2009; Peguero & Jiang, 2014; Peguero et al., 2014). However, the findings are not uniform. Some research focused on a single context and showed variations in the predictors of behavioral outcomes within each immigrant generation, suggesting unique pathways to delinquency and other behavioral outcomes experienced by youth across immigrant generations (Chao, 2001; Driscoll et al., 2008). For example, the literature suggests that parenting practice and the relationships between parents and children may change generationally in response to a gradual deterioration of the culture of origin and the need to adjust to the parenting culture of mainstream society (Driscoll et al., 2008). Chao (2001) reported that immigrant parents, especially first generation, tend to emphasize obedience to authority, while parents of later generations are more likely to adapt the dominant culture that encourages independence and open communication between parents and children.

Similarly, Chang and Le’s (2005) study shows that the lack of socioeconomic resources combined with language barriers limits immigrant parents’ control over their children’s whereabouts, companions, and studies, contributing to their low involvement in children’s education. Moreover, differences in parenting practice may not necessarily translate into differences in behavioral outcomes for immigrant youth. Despite low parental involvement in education, first-generation immigrant youth usually demonstrate higher levels of academic performance than later generations (Bankston & Zhou, 2002; Pong, Hao, & Gardner, 2005). The literatures support the argument that family control variables derived from social control theory, such as parental monitoring, or parental attachment, may not be similarly efficacious in predicting delinquency across immigrant generations.

Recent research on the immigrant-delinquency paradox has also identified the differential responses of youth across generations to school control and association with delinquent friends. In their study, Jiang and Peterson (2012) and Peguero (2013) found that participating in extracurricular activities, which represents a bonding element of social control, is indeed associated with an increase in the risk of violence and victimization for first- and second-generation immigrant youth. They suggested that the detrimental effect of extracurricular activities may emerge as a response to isolation and segregation from domestic friends and school authorities experienced by immigrant youth. The feeling of segregation and a strong desire to adjust in U.S. society may also amplify immigrant youth’s sense of pressure to conform to the behaviors of their friends, heightening their vulnerability to the influence of delinquent friends (Chang & Le, 2005; Dipietro & McGloin, 2012). Dipietro and McGloin’s (2012) study of friend-based risks for delinquency concluded that first- and second-generation immigrant youth are more susceptible to the criminogenic influence of delinquent friends than third-plus generation youth. These varied findings indicate that mechanisms of social control and differential association may not be uniformly experienced by youth from different immigrant generations. Thus, it is necessary to account for the possibility of generation-specific effects of social control and differential association on delinquent outcomes.

The findings of the previous studies underscore the importance of examining the generation-specific effect of family and school control and delinquent friends on delinquency across immigrant generation. In this context, the first goal of this study is to assess the relative strength of the three dimensions of social control and differential association—family control, school control, and friends’ delinquency—in explaining the relationship between immigrant generation and delinquency. Although the previous studies have examined these factors, most of them examined them in a single context. An examination of both independent effects and the combined effects of these factors on explaining generation–delinquency relationship is important because these processes are presumably related. For example, some studies suggest that the parent–child relationship should be a focal area of inquiry for understanding immigrant youth behavior in the school and friend contexts. Strong parental control and obligation to family commonly found in immigrant families may constrain opportunities available for immigrant youth to interact with deviant friends (Portes & Rumbaut, 2001; Portes & Zhou, 1993; Zhou, 1997). Lacking other resources, parental expectation in education may also be crucial in cultivating an atmosphere conducive to immigrant youth engagement in school and commitment to education (Kao & Rutherford, 2007; Portes & Rumbaut, 2001; Turney & Kao, 2009). At the same time, the effect of exposure to delinquent friends is salient in explaining generational difference in delinquency in that association with delinquent friends may attenuate family and school control (Bowman, Prelow, & Weaver, 2007; Deutsch et al., 2012). Thus, this study contributes to the literature by simultaneously investigating the extent to which exposure to parental control, school control, and association with delinquent friends explains the relationship between immigrant generation and delinquency.

The second goal of this study is to examine the importance of school control, family control, and associations with delinquent friends for adolescent delinquent behavior in each immigrant generation. As mentioned, immigrant youth may experience school, family, and friends differently; consequently, these factors may have different effects on immigrant youth compared with youth from the majority culture. For example, because immigrant families have different parenting practices, family control variables that are found to be protective for mainstream youth may not adequately explain the delinquent behavior of immigrant youth. Similarly, the school may be less important as a control agent for earlier immigrant generations because they are not engaged with, or attached to, school to the same extent as native-born students. This perspective underscores the need to investigate unique pathways to delinquency experienced by youth across immigrant generations.

In sum, many empirical studies incorporated the measure of both social control and differential association toward explaining delinquent behaviors (e.g., Alarid, Burton, & Cullen, 2000; Le, Monfared, & Stockdale, 2005; Whaley, Hayes-Smith, & Hayes-Smith, 2010). The present study draws from the conceptual frameworks of these two theories to better understand the processes underlying adolescent violent and nonviolent delinquency and substance use across immigrant generations. By employing these two perspectives to guide the present study, we compare the strength of family control, school control, and association with delinquent friends in explaining differences in self-reported violent and nonviolent delinquency and substance use for adolescents across immigrant generations. We also investigate whether and to what extent the exposure to family control, school control, and friends’ delinquency differentially influence delinquent outcomes of adolescents from different generations. By assessing the effect of these social institutions, we hope to arrive at a more precise understanding of the significance of control theory and differential association theory as explanations of delinquent and criminal behavior for demographically diverse youth.

Sample and Data

This study uses data from the first two waves of National Longitudinal Study of Adolescent Health (Add Health). Add Health is a longitudinal, nationally representative, school-based survey of adolescents in Grades 7 through 12 in a sample of 134 high schools throughout the United States. In the initial in-school survey (conducted in 1994–1995), questionnaires were administered to all students in Grades 7 through 12 who were present on the day of survey (n = 90,118). The in-school questionnaires covered a wide range of topics including adolescents’ demographic characteristics, health status, parental education levels, friend networks, academic grades, and school activities. Then, a smaller sample of approximately 200 students from each of the 80 pairs of schools (n = 20,745) was randomly selected to participate in subsequent in-home interviews, along with their mother or other female heads of the households. The in-home interview covered additional topics, including adolescents’ and their parents’ country of birth; language used at home; their social, economic, physical, and psychological well-being; and health risk behaviors.

In 1996, a second round of interviews was instrumented among all adolescents who were interviewed at Wave I in-home survey. These respondents were asked questions very similar to those asked at Wave I interview. Eighty-eight percent (n = 14,738) of the students eligible to be reinterviewed completed the second-wave survey.

Add Health data are suitable for our study, given that it contain a rich variety of social contextual variables in school and family and measures of adolescent delinquency and other deviant behaviors (e.g., delinquent behaviors, fighting and violence, substance use, suicide, and premarital pregnancy). A particular strength of the data is that the respondent record includes adolescents’ and their parents’ country of birth, which allows us to identify the different generations of immigrants. Drawing on these data components, the sample of our study includes all respondents who have completed both in-school survey for first-wave data and in-home interview for first-wave and second-wave data. After the cases with missing data for weights, immigrant generations, and delinquency were excluded, the final sample for this study consists of 13,121 adolescents nested within 120 schools.

Measures

Dependent Variables

Three measures of delinquency from the second-wave in-home survey were used. The first, violence, is an additive scale consisting of whether a respondent has engaged in any of five violent activities. Based on self-reported responses from the in-home survey, adolescents were asked during the past year whether they had participated in the following violent behaviors: pulled a knife or gun on someone, shot or stabbed someone, had a serious physical fight, use a weapon in a fight, or hurt someone badly enough to need a bandage or care from a doctor or nurse (α = .72).

The second measure, nonviolent delinquency, is an additive scale consisting of 13 items of nonviolent delinquent behaviors. These delinquency items tap how often in the past 12 months did the adolescent paint graffiti, deliberately damage property, lie to parents, run away from home, shoplift, steal something worth less than US$50, steal something worth US$50 or more, burglarize, borrow a car without the owner’s permission, sell drug, use or threaten to use a weapon, be loud in a public place, and participate in a group fight. Each item in the delinquency scale has four response categories (0 = never committed the act, 1 = committed 1 or 2 times, 2 = committed 3 or 4 times, and 3 = committed 5 or more times). The delinquency scale ranges from 0 to 39 (α = .80).

The third measure, substance use, is a binary measure indicating whether or not the adolescent reported engaging in any of the three forms of substance use during the past year: smoking cigarettes, using marijuana, and drinking alcohol. These items were combined to form the binary dependent variable indicating using any of the three forms of substances (1 = use substances) versus never using substance (0 = don’t use substances). It should be noted that the three offense groups are not mutually exclusive. For example, 27% of the respondents who reported using drugs also engaged in violent behaviors. Twenty-seven percent of the respondents who engaged in nonviolent delinquent behaviors also reported violent involvement.

Independent Variables

Three binary measures of immigrant generations (first generation [n = 1,262], second generation [n = 2,033], and third-plus generation [n = 9,826]) were determined by adolescents’ responses to questions on whether or not they and their parents were born in the United States. from the Wave I in-school survey. First generation indicates that the respondent was born outside of the United States (1 = not U.S. born; 0 = others). Second generation includes respondents who were U.S.-born but have at least one parent who was born outside of the United States (1 = U.S.-born with non-U.S.-born parent(s); 0 = others). U.S.-born respondents with parents who were also born in the United States are considered as belonging to the third-plus generation (1 = U.S.-born with parents who were also born in the United States; 0 = others). Because the third-plus generation is much farther removed from the possible impact of immigration on childhood and adolescent development, this group is considered as part of the domestic population. For the convenience of the current study, this group is named as third-plus generation immigrants.

Following previous Add Health studies, the measure of family attachment is an additive scale based on 4 items assessing adolescents’ feelings about family relationships (Haynie, 2001). Adolescents were asked about how people in the family were understanding, how people in the family had fun together, how the family paid attention to the respondent, and whether the respondent wanted to leave home. Each item in the attachment scale has five response categories ranging from 1 = not at all to 5 = very much. The additive scale was computed based on these items, ranging from 5 to 25 (α = .75). Following Soller and Haynie (2013), another family measure, parental educational expectations, indicates the degree to which respondents feel their mother and/or father would be disappointed if they did not graduate from college. The variable has a range from 1 = not disappointed to 5 = very disappointed. Our final family measure, parental control, is measured by a scale consisting of 7 dichotomous items which ask if the adolescent’s parents controlled their weekend curfew, friends, clothing, TV time, TV programs, weekday bedtime, and diet. Responses for each of these 7 items were summed to form the control index, with a higher score indicating stronger control (α = .63,¹ range = 0–7).

The measure of school attachment is the average value of the adolescent’s responses to 3 items assessing whether adolescents felt close to people at school, felt like part of school, and felt happy to be at school. Responses for each item range from 1 = not at all to 5 = very much (α = .78). School commitment is measured by a variable that combines 2 items assessing how much the adolescent wanted to attend college and how likely they would attend college. Responses for each item range from 1 = not at all to 5 = very much. Measurement of school trouble is an additive scale of 4 items asking whether adolescents had trouble (0 = no; 1 = yes) in getting along with teachers, trouble in paying attention, trouble in getting homework done, and trouble with other students (α = .65; Bui, 2009).

School involvement is measured based on items of participating in school extracurricular activates (Peguero, 2013). Adolescents were asked to indicate whether during the school year they were/would be participating in any of the 33 clubs, organizations, or teams. Extracurricular participation distinguishes adolescents who reported that they were/would be participating in one or more of these activities (1 = participate in activities) from adolescents not participating in any extracurricular activities (0 = don’t participate in activities).

Based on previous work using the Add Health data (Beaver & Wright, 2005; Bellair, Roscigno, & McNulty, 2003; Daigle, Cullen, & Wright, 2007), the measure of friends’ delinquency is based on 3 items asking adolescents about how many of their three best friends smoke, use alcohol, and use marijuana. Responses for each items include 0 = no friend, 1 = one friend, 2 = two friends, and 3 = three friends (α = .76). Many researchers employed the measures of friends’ drug use as a reasonable proxy to estimate delinquent friends associations because it is strongly correlated with other forms of delinquent behaviors (Bellair et al., 2003; Deutsch et al., 2012).

Control Variables

To identify the independent effects of independent variables on adolescent delinquency, it is important to control for the influence of other variables found in prior research to be related to adolescent deviance. In this study, we control for the youth’s demographic characteristics, their parents’ level of education, family structure, and community characteristics. The demographic factors include gender which distinguishes between males (1 = male) and females (0 = female) and age measured in years. The demographic factors also include five measures that indicate the adolescent’s racial/ethnic affiliation: Whites (n = 6,831), Blacks (n = 2,663), Hispanics (n = 2,265), Asians (n = 998), and others (n = 364). Whites are considered as reference group. Family structure indicates whether an adolescent lives in a household with two married parents (1 = two-parent family, 0 = other family arrangements). Parental education is operationalized as the mean value of education (on a 7-point scale) for both parents; if only one parent’s education information is available, the scale value for that parent is used. Response options are 0 = no formal education, 1 = completed 8th grade or less, 2 = completed more than 8th grade but did not complete high school, 3 = finished high school or received a General Educational Development (GED), 4 = completed a trade or vocational school education after high school, 5 = some college, 6 = is a college graduate, and 7 = has a graduate or professional degree. Residential stability is measured by an item that captures the proportion of occupied housing units moved into in the last 5 years. Community disadvantage is measured by an index that combines 5 items, including the proportion of female-headed households, the proportion of persons living below the poverty level, the proportion of persons 18 and over with no high school diploma, the proportion of households receiving public assistance, and the proportion of unemployed residents (α = .87).

Analytic Strategy

Our analyses begin with descriptive statistics for dependent variables and independent variables across three immigrant generations. This enables us to identify major differences among these groups in delinquency as well as the factors of social control and differential associations. Next, we use survey-corrected logistic regression to further investigate generational differences in violence and substance use and how the differences are explained by the measures of family control, school control, and association with delinquent friends. Given that nonviolent delinquency scale is highly skewed and contains disproportionate zero (about 25%), we use negative binomial regression models to estimate the effect of the independent variables on nonviolent delinquency. For each delinquency outcome, we first regress the dependent variable on immigrant generations and control variables in a baseline model. We then build on this model by adding the three measures of family control, the four measures of school control, and the two measures of association with delinquent friends.

Because Add Health sample is based on a complex sampling survey design, appropriate weighting was used to account for the design effect. In addition, Add Health uses clustered sampling design, where adolescents in the survey were sampled within schools and so are not independent observations. Thus, we perform all statistical analyses by using survey-corrected regression to correct for multiple stages of cluster sample design and unequal probability of selection. In order to maintain a large sample size, we use multiple imputation (MI) technique to deal with missing data (see Little & Rubin, 1987; Schafer & Olsen, 1998, for more about MI). Assuming that data are missing at random, MI provides unbiased estimates of the missing values using other variables in the model, thereby strengthening its precision (Schafer & Graham, 2002). In our sample, parental education, family attachment, parental commitment, parental control, school attachment, school involvement, school trouble, friends’ deviance, and friends’ influence contain missing items. We impute replacement values for each missing value to produce a data set with complete information for 13,121 cases.

Results

Table 1 presents weighted means for all variables by immigrant generations. Asterisks signify means that are significantly different from those of the third-plus generation. Based on Table 1, the sample is largely composed of youth who are third-plus generation (74.89%). First- and second-generation youth comprise 9.62% and 15.49% of the sample, respectively. Male and female respondents are almost equally distributed across immigrant generations, whose mean ages are around 16 years. Focusing on descriptive statistics indicates a disparity in delinquency between the first generation and the second generation as compared with the third-plus generation. Not surprisingly, the first generation has the lowest involvement in all forms of delinquency. Means of violence (0.26), nonviolent delinquency (2.32), and substances use (0.41) reported by the first generation are significantly lower than those reported by the third-plus generation (0.37, 2.76, and 0.58). The second generation, however, has mean scores of delinquency comparable to the third-plus generation.

Table 1.

Weighted Means of All Variables by Immigrant Generation.

	First Generation	Second Generation	Third Plus
Variables	(n = 1,262)	(n = 2,033)	(n = 9,826)
Nonviolent delinquency	2.32*	2.39*	2.76
Violent delinquency	0.26***	0.40	0.37
Substance use	0.41***	0.55	0.58
Gender	0.47	0.49	0.50
Age	16.19*	15.68	15.72
Whites	0.13***	0.31***	0.74
Blacks	0.04***	0.08***	0.17
Hispanics	0.48***	0.42***	0.05
Asians	0.31***	0.14***	0.01
Others	0.04**	0.05***	0.03
Two-parents family	0.78	0.74	0.73
Parental education	3.62**	3.84*	4.16
Residential stability	0.08	0.08	0.08
Prior delinquency	5.85***	8.15	8.20
Community disadvantage	0.15	0.14	0.13
Family attachment	18.27	18.05*	18.23
Parental commitment	4.18***	4.00	3.95
Parental control	2.25**	2.04*	1.90
School attachment	3.84	3.70*	3.77
School commitment	8.61	8.50	8.56
School trouble	2.88	2.92	2.97
School involvement	0.77***	0.83*	0.87
Friends’ delinquency	1.61***	2.27*	2.54

Note. N = 13,121.

*p < .05. **p < .01. ***p < .001.

Descriptive statistics also show significant differences in the measures of family control, school control, and association with delinquent friends across generations. When looking at the measures of family control, the first generation reported higher commitment to their parents and stronger parental control than the third-plus generation. However, the first generation has a lower level of participation in extracurricular activities. In terms of association with delinquent friends, the first generation reported a lower mean of friends’ delinquency than the third-plus generation. For the second generation, although demonstrating a higher mean of parental control (2.04) than the third-plus generation (1.90), the parental control is lower than that of the first generation (2.25), and the mean of family attachment (18.05) is lower than that of the third-plus generation (18.23). In addition, the second generation has weaker school attachment and a lower level of participation in extracurricular activities than the third-plus generation. The second generation also has a higher mean of friends’ delinquency than the first generation, showing a less pronounced difference from the third-plus generation.

Do these differences in family control, school control, and association with delinquent friends explain delinquency across immigrant generations? In our next step, we employ regression analyses predicting three forms of delinquency separately. Tables 2 and 3 present negative binomial regression results using nonviolent delinquency and violence as the dependent variable. Table 4 provides logistic regression results using substance use as the dependent variables. Model 1 of Table 2 through Table 4 is a baseline model that includes all control variables (not shown) and immigrant generation. The results demonstrate that the gaps in delinquency between the first and the third-plus generations persist after we control for adolescents’ sociodemographic characteristics. The first generation is significantly less likely than the third-plus generation to become involved in nonviolent delinquency, violence, and substance use. For example, being a first-generation immigrant reduces the risk of substance use by 38% (((exp(−0.47)) − 1) × 100) compared with the third-plus generation. However, the second generation does not show any difference in forms of delinquency from the third-plus generation.

Table 2.

Survey-Adjusted Negative Binomial Regression of Nonviolent Delinquency Involvement on the Variables of Social Control and Differential Associations.

Variables	Model 1	Model 2	Model 3	Model 4
Gender	0.02 (0.03)	0.03 (0.03)	0.02 (0.03)	0.03 (0.03)
Age	−0.06*** (0.01)	−0.07*** (0.01)	−0.07*** (0.01)	−0.09*** (0.01)
Family structure	−0.09** (0.03)	−0.07* (0.03)	−0.04 (0.03)	−0.02 (0.03)
Parental education	0.02 (0.01)	0.02 (0.01)	0.03** (0.01)	0.03** (0.01)
Blacks	0.06 (0.06)	0.09 (0.06)	0.08 (0.06)	0.10 (0.06)
Hispanics	0.28*** (0.06)	0.30*** (0.06)	0.30*** (0.06)	0.30*** (0.06)
Asians	0.12 (0.08)	0.09 (0.08)	0.10 (0.08)	0.10 (0.08)
Others	0.19* (0.09)	0.16 (0.09)	0.16 (0.09)	0.13 (0.09)
Prior delinquency	0.04*** (0.00)	0.04*** (0.00)	0.04*** (0.00)	0.03*** (0.00)
Residential stability	−1.49** (0.48)	−1.47** (0.47)	−1.36** (0.46)	−1.33** (0.45)
Community disadvantage	1.26 (0.82)	1.25 (0.81)	1.09 (0.77)	1.03 (0.79)
First generation	−0.14* (0.07)	−0.14* (0.07)	−0.15* (0.07)	−0.10 (0.07)
Second generation	0.10 (0.05)	0.09 (0.05)	0.08 (0.05)	0.11 (0.06)
Family attachment		−0.05*** (0.01)	−0.05*** (0.01)	−0.04*** (0.01)
Parental control		−0.02* (0.01)	−0.03** (0.01)	−0.02* (0.01)
Parental educational expectations		−0.01*** (0.01)	0.00 (0.02)	0.00 (0.02)
School attachment			−0.05* (0.02)	−0.04 (0.02)
School commitment			−0.00 (0.01)	0.00 (0.01)
School involvement			−0.12* (0.05)	−0.10 (0.05)
School trouble			0.09*** (0.02)	0.08*** (0.02)
Friends’ delinquency				0.05*** (0.01)

Note. N = 13,121. Standard errors in parentheses.

*p < .05. **p < .01. ***p < .001.

Table 3.

Survey-Adjusted Logistic Regression of Violent Delinquency on the Variables of Social Control and Differential Associations.

Variables	Model 1	Model 2	Model 3	Model 4
Gender	0.53*** (0.06)	0.54*** (0.06)	0.52*** (0.06)	0.54*** (0.06)
Age	−0.04* (0.02)	−0.05** (0.02)	−0.05** (0.02)	−0.08*** (0.02)
Family structure	−0.17** (0.05)	−0.15** (0.05)	−0.13* (0.05)	−0.09 (0.05)
Parental education	−0.05* (0.02)	−0.04* (0.02)	−0.03 (0.02)	−0.02 (0.02)
Blacks	0.42*** (0.08)	0.44*** (0.08)	0.43*** (0.08)	0.47*** (0.08)
Hispanics	0.44*** (0.08)	0.47*** (0.08)	0.46*** (0.08)	0.44*** (0.08)
Asians	−0.00 (0.14)	−0.02 (0.13)	−0.01 (0.14)	−0.02 (0.14)
Others	0.44** (0.13)	0.40** (0.13)	0.40** (0.13)	0.39** (0.13)
Prior delinquency	0.05*** (0.00)	0.05*** (0.00)	0.04*** (0.00)	0.03*** (0.00)
Residential stability	0.32 (0.78)	0.37 (0.77)	0.52 (0.76)	0.50 (0.74)
Community disadvantage	0.97 (1.25)	0.88 (1.25)	0.72 (1.21)	0.73 (1.23)
First generation	−0.38** (0.13)	−0.36** (0.13)	−0.36** (0.13)	−0.28* (0.13)
Second generation	−0.04 (0.10)	−0.05 (0.09)	−0.05 (0.09)	0.00 (0.09)
Family attachment		−0.04*** (0.01)	−0.03*** (0.01)	−0.03*** (0.01)
Parental control		−0.02 (0.02)	−0.02 (0.02)	−0.02 (0.02)
Parental educational expectations		−0.05** (0.02)	−0.03 (0.02)	−0.02 (0.02)
School attachment			−0.07* (0.03)	−0.05 (0.03)
School commitment			−0.03 (0.02)	−0.02 (0.01)
School involvement			−0.12 (0.09)	−0.08 (0.09)
School trouble			0.07* (0.03)	0.06* (0.03)
Friends’ delinquency				0.09*** (0.01)

Note. N = 13,121. Standard errors in parentheses.

*p < .05. **p < .01. ***p < .001.

Table 4.

Survey-Adjusted Logistic Regression of Substance Use on the Variables of Social Control and Differential Associations.

Variables	Model 1	Model 2	Model 3	Model 4
Gender	−0.16** (0.05)	−0.14** (0.05)	−0.15** (0.05)	−0.11* (0.05)
Age	0.18*** (0.02)	0.15*** (0.02)	0.16*** (0.02)	0.09*** (0.02)
Family structure	−0.25*** (0.05)	−0.21*** (0.05)	−0.19*** (0.05)	−0.14* (0.06)
Parental education	−0.01 (0.02)	−0.01 (0.02)	−0.01 (0.02)	−0.00 (0.02)
Blacks	−0.60*** (0.09)	−0.57*** (0.08)	−0.60*** (0.08)	−0.55*** (0.08)
Hispanics	−0.18 (0.10)	−0.15 (0.10)	−0.14 (0.10)	−0.14 (0.10)
Asians	−0.45*** (0.12)	−0.48*** (0.13)	−0.48*** (0.13)	−0.50*** (0.13)
Others	0.09 (0.13)	0.07 (0.13)	0.06 (0.13)	0.03 (0.14)
Prior delinquency	0.08*** (0.00)	0.08*** (0.00)	0.07*** (0.00)	0.05*** (0.00)
Residential stability	−3.16*** (0.59)	−2.98*** (0.59)	−2.82*** (0.60)	−2.88*** (0.58)
Community disadvantage	3.13*** (0.86)	2.98*** (0.86)	2.79** (0.88)	2.59** (0.87)
First generation	−0.47*** (0.13)	−0.45*** (0.13)	−0.43*** (0.12)	−0.31* (0.12)
Second generation	−0.05 (0.08)	−0.06 (0.08)	−0.06 (0.08)	−0.01 (0.09)
Family attachment		−0.07*** (0.01)	−0.05*** (0.01)	−0.05*** (0.01)
Parental control		−0.07*** (0.02)	−0.07*** (0.02)	−0.06*** (0.02)
Parental educational expectations		0.01 (0.03)	0.02 (0.02)	0.02 (0.02)
School attachment			−0.11** (0.04)	−0.07* (0.04)
School commitment			0.01 (0.02)	0.03 (0.02)
School involvement			−0.02 (0.08)	0.05 (0.08)
School trouble			0.10*** (0.02)	0.09*** (0.02)
Friends’ delinquency				0.20*** (0.01)

Note. N = 13,121. Standard errors in parentheses.

*p < .05. **p < .01. ***p < .001.

In Model 2 for each delinquency outcome, we add the measures of family control. Looking at Model 2, we find few surprises in the relationships between family control and adolescent delinquency. Consistent with the social control perspective, family attachment and parental control have significant effect in preventing the adolescents from involvement in almost all forms of delinquency. Adolescents living in families in which they receive more attention and closer parental supervision experience a lower level of nonviolent delinquency and substance use. Adding family control to the model reduces the statistical effect of immigrant generations on violence and substance use. Yet, the reduction in coefficients for the first generation is modest. The coefficients for the first generation still provide a sizable protection from delinquency.

To test the effect of school control, we incorporate the measures of school control in Model 3 for each delinquency outcome. On the one hand, the results in Model 3 show that, as school attachment increases, adolescents become less likely to engage in all forms of delinquency. On the other hand, adolescents who have a more trouble in school are more likely to be delinquent. Adding school control measures further attenuates the effect of immigrant generation on delinquency, but this attenuating effect is also modest. Indeed, variables of family control and school control together attenuate about 5% (violence) to about 9% (substance use) of the coefficients for the first generation when we are looking across delinquency outcomes. The difference between the first generation and the third-plus generation in all forms of delinquency remains significant, suggesting that social control variables do not fully explain the generational gap in delinquency.

The final model (Model 4) adds the measures of friends’ deviance and friends’ influence to assess if the effect of generations on delinquency works through its impact on association with delinquent friends. Consistent with the differential association perspective, having more friends who are delinquent is associated with a higher likelihood of involvement in all delinquent behaviors. Results also show that friends’ associations attenuate the generational effect on delinquency. Compared with Model 3, introducing the measure of association with delinquent friends in Model 4 brings down the coefficient of the effect of the first generation on nonviolent delinquency by 33%, the effect on violence by 22%, and the effect on substance use by 28%. Although the coefficients for violence and substance use are still significant, the coefficients of the first generation either cease to be significant (nonviolent delinquency) or remain marginally so (violence and substance use). It appears that friends’ associations explain most of the difference in delinquency between the first and the third-plus generation.

Do the measures of social control and differential association differentially affect delinquent outcomes for adolescents from different generations? Next, regression analyses for the three delinquency outcomes by immigrant generation are conducted to reach a more complete understanding of independent effect social control and differential associations on delinquency within each generation. The findings are presented in Table 5. The results of separate regression equations for three generational groups reveal important generational differences with respect to the effect of family control, school control, and association with delinquent friends.

Table 5.

Survey-Adjusted Negative Binomial Regression of Delinquency Involvement on the Independent Variables of Social Control and Differential Associations by Immigrant Generational Status.

	Nonviolent Delinquency			Violent Delinquency			Substance Use
Variables	First Generation	Second Generation	Third Plus	First Generation	Second Generation	Third Plus	First Generation	Second Generation	Third Plus
Gender	0.05 (0.10)	0.13 (0.09)	0.02 (0.03)	0.42 (0.26)	0.50** (0.15)	0.55*** (0.06)	−0.11 (0.17)	−0.04 (0.15)	−0.12* (0.06)
Age	−0.07 (0.04)	−0.11*** (0.02)	−0.08*** (0.01)	−0.11 (0.09)	−0.11** (0.04)	−0.08*** (0.02)	0.18** (0.05)	0.00 (0.05)	0.10*** (0.02)
Family structure	−0.22* (0.11)	−0.15 (0.09)	0.01 (0.03)	−0.02 (0.20)	−0.21 (0.13)	−0.06 (0.06)	−0.35* (0.18)	−0.12 (0.12)	−0.12 (0.06)
Parental education	0.03 (0.04)	0.02 (0.02)	0.03* (0.01)	0.06 (0.08)	−0.04 (0.04)	−0.03 (0.02)	0.05 (0.06)	0.03 (0.04)	−0.01 (0.02)
Blacks	0.43 (0.25)	0.04 (0.15)	0.10 (0.06)	0.83* (0.40)	0.25 (0.25)	0.47*** (0.09)	−1.61** (0.49)	−0.69* (0.27)	−0.52*** (0.08)
Hispanics	0.10 (0.18)	0.44*** (0.12)	0.28*** (0.07)	0.50 (0.31)	0.40* (0.17)	0.47*** (0.11)	−0.56 (0.32)	−0.14 (0.18)	−0.06 (0.13)
Asians	0.07 (0.18)	0.19 (0.12)	0.11 (0.14)	0.12 (0.33)	−0.08 (0.19)	−0.31 (0.21)	−0.61 (0.33)	−0.65** (0.21)	−0.66** (0.24)
Others	0.25 (0.35)	0.32 (0.17)	0.09 (0.12)	0.24 (0.51)	0.13 (0.36)	0.42** (0.14)	0.04 (0.48)	0.02 (0.32)	−0.01 (0.15)
Prior delinquency	0.04*** (0.01)	0.03*** (0.01)	0.03*** (0.00)	0.05* (0.02)	0.02 (0.01)	0.03*** (0.00)	0.05 (0.03)	0.04* (0.01)	0.05*** (0.01)
Residential stability	−0.52 (0.94)	−0.67 (0.78)	−1.50** (0.53)	3.42 (2.84)	−1.70 (1.36)	0.81 (0.88)	−2.57 (1.62)	−2.18 (1.67)	−2.98*** (0.63)
Community disadvantage	−2.15 (1.73)	0.21 (1.30)	1.33 (0.89)	−7.24 (5.28)	5.00* (2.08)	0.34 (1.43)	−0.22 (3.25)	3.72 (2.59)	2.57** (0.97)
Family attachment	−0.04* (0.02)	−0.04* (0.02)	−0.05*** (0.01)	−0.01 (0.04)	−0.03 (0.02)	−0.03** (0.01)	−0.01 (0.03)	0.07** (0.03)	0.05*** (0.01)
Parental control	0.01 (0.03)	−0.06* (0.02)	−0.02 (0.01)	−0.03 (0.05)	−0.05 (0.04)	−0.01 (0.02)	−0.05 (0.04)	−0.06 (0.05)	−0.06** (0.02)
Parental educational expectations	−0.07 (0.05)	−0.03 (0.03)	0.02 (0.02)	−0.10 (0.12)	−0.10 (0.05)	−0.01 (0.02)	−0.01 (0.09)	−0.10 (0.05)	0.04 (0.02)
School attachment	−0.01 (0.07)	−0.07 (0.06)	−0.03 (0.02)	0.18 (0.15)	−0.05 (0.09)	−0.06 (0.03)	−0.07 (0.14)	−0.17 (0.09)	−0.06 (0.04)
School commitment	0.02 (0.03)	0.03 (0.02)	−0.00 (0.01)	−0.06 (0.06)	−0.05 (0.04)	−0.02 (0.02)	−0.01 (0.06)	0.08* (0.04)	0.02 (0.02)
School involvement	0.01 (0.14)	−0.02 (0.12)	−0.12 (0.06)	0.37 (0.28)	−0.22 (0.24)	−0.09 (0.09)	0.12 (0.25)	0.04 (0.20)	0.05 (0.09)
School trouble	0.09 (0.05)	0.09* (0.04)	0.08*** (0.02)	0.20* (0.10)	0.02 (0.06)	0.06 (0.03)	0.05 (0.09)	0.09 (0.07)	0.09** (0.03)
Friends’ delinquency	0.08** (0.03)	0.03* (0.02)	0.06*** (0.01)	0.11 (0.05)	0.06* (0.03)	0.09*** (0.01)	0.15** (0.05)	0.21*** (0.04)	0.20*** (0.02)

Note. Standard errors in parentheses.

*p < .05. **p < .01. ***p < .001.

Specifically, these findings indicate that for the first generation, despite exhibiting stronger social control than later generations, the measures of social control in family and school have little effect on their delinquency outcomes. Although some measures of social control significantly predict delinquency for the second generation, the protective effect of social control is not consistent across different type of delinquency. For the third generation, strong family attachment significantly reduces nonviolent delinquency and substance use, while trouble in school increases the likelihood of these two types of delinquency.

When looking at the effects of differential association on delinquency, findings indicate that friends’ delinquency is the most consistent predictor of delinquency across generations. All three generational groups who reported higher levels of friends’ delinquency are more likely to have higher levels of involvement in all types of delinquency. The only exception is the nonsignificant effect of friends’ delinquency on violent delinquency among the first generation.

We then examine a series of equality of coefficients tests to determine whether the coefficients relating association with delinquent friends to delinquency differ significantly across generations. The results do not show any significant difference in the coefficients across immigrant generations. Taken together, these findings reveal that when social control and differential association are examined simultaneously, differential association is a stronger and more consistent predictor of delinquency than social control for each generation.

Conclusion and Discussion

Although stereotypes of criminality among immigrants remain strong in public opinion, empirical research in the past few decades has consistently found that immigrants are typically underrepresented in crime and delinquency. Such a linkage between immigrants and crime has motivated considerable research interest in examining the fit of conventional theories of crime as explanations of the immigrant paradox. While previous studies have shown that measures derived from traditional criminological theories such as social control and differential association are crucial in the linkage between immigrant generation and crime, rarely have researchers evaluated the impact of these measures at the same time.

The primary goal of this study is to assess simultaneously the effect of three dimensions of social control and differential association—family control, school control, and association with delinquent friends—on adolescent delinquency across immigrant generations. We are particularly interested in assessing the relative impacts of these three dimensions in explaining the relation between immigrant generations and delinquent outcomes. Our analyses yield several salient results.

Our first research question is to what extent the differential exposures to protective (i.e., family control and school control) and risk factors (i.e., delinquent friends) contribute to differential involvement in delinquency across immigrant generations. Not unexpectedly, the results show a significant lower rate of violence, nonviolence, and substance use among the first generation relative to the third-plus generation but a comparable rate of these forms of delinquency between the second and third-plus generations. These results provide support for the immigrant paradox literature which found that today’s immigrants are relatively healthier and less crime-prone when they arrived to the United States, but these advantages become smaller overtime. In addition, the results show that both social control theory and differential association theory appear to have a general effect in explaining generational gap in delinquency. Stricter family and school control and less exposure to deviant friends observed among the first generation combine to account for the lower level of delinquency for the first generation.

Further exploration of these independent factors, however, indicates that the measures derived from social control theory, including family and school control, have weaker explanatory power. The inclusion of the measures of association with delinquent friends influence substantially reduces the effect of immigrant generation on delinquent outcomes. In total, these findings suggest that the important predictors of delinquency as measured in this study (family control, school control, and differential associations) are shaped by immigrant generation, yet association with delinquent friends is crucial in determining generational difference in delinquency.

Our second research question focuses on generation-specific analyses to determine whether the pathways to delinquency were different for adolescents across immigrant generations. Although previous research suggested that social control, especially family control, is expected to exert a strong protective effect on delinquency, risky behaviors, and school performances for immigrant youth (Kao & Tienda, 1995; Pong et al., 2005), our study indicates that the pathways to delinquency diverge across generations in an opposite way. In particular, certain measures of control theory, such as family attachment and school attachment, decrease the risk for different types of delinquency among the third-plus generation, but these control measures do not exhibit a consistent relationship to delinquency among the first and second generations. This result is contrary to Hirschi’s (1969) prediction of social control for general population. However, our results support the general effect of differential association across immigrant generations. The exposure to friends’ delinquency significantly increases the likelihood of delinquency for all generational groups regardless of delinquency type.

While less assimilated immigrant youth may be more influenced by their parents and less by their friends (Myers at al., 2009), our study finds the opposite to be true. Although earlier immigrant generations demonstrate advantages along dimensions of social control, we find that the ability of social control to explain the effect of immigrant generation on delinquency is, at least, very limited. Such results add evidence to the research in the generality of social control theory which claims that social bond may operate in a different manner to explain offending across gender, age, race, and cultural boundaries (Alarid et al., 2000; Anderson, 1999; Daigle et al., 2007; Hwang & Akers, 2003).

Based on our results, earlier-generation immigrant youth are not subject to the types of social control (i.e., family attachment and school attachment) native adolescents experience. Thus, it is reasonable to ask whether the classical perspectives such as social control are applied merely to U.S. youth who are not from immigrant families. Indeed, the relationship between commonly examined predictors and delinquency is so well established among samples of native youth that many researchers argue for immigrant-specific theories of crime and delinquency (e.g., Mears, 2001). It is premature, however, to conclude that control in family and school suggested by social control theory could not expand our understanding of delinquency and crime among current immigrant youth. To illuminate criminal behaviors from different populations, scholars claimed that it is necessary to operationalize and test variables that incorporate the unique experiences of a particular population that are contributive to their criminality (Steffensmeier & Allan, 1996). For example, perhaps the weak explanatory power of family control is attributable to the fact that family bonding experienced by immigrant youth is qualitatively distinct from that generally observed and measured in nonimmigrant families. Research into immigrant assimilation has shown that certain family influence that is unique to immigrant status, such as authoritarian parenting style, hurts White students but is beneficial to the family control and the subsequent school performances among certain immigrant minorities (Chao, 2001; Vartanian, Karen, Buck, & Cadge, 2007). Other immigrant-specific factors found in immigrant family include consonant acculturation, keeping track of school work, instilling guilt over parental sacrifice, collective culture, and fulfilling obligations (Chang & Le, 2005; Choi, He, & Harachi, 2008; Kao, 1995, 2004; Portes & Rumbaut, 2001; Portes & Zhou, 1993). Future research needs to examine how the factors that are uniquely drawn from immigrant families influence family bonding and delinquency.

The findings of the present study appear to be more consistent with differential association theory than with social control theory. This is consistent with previous research which concluded that immigrant status may offer some protection against exposure to delinquent friends (Le et al., 2005; Le & Stockdale, 2008; Myers et al., 2009). It may stand to reason that language and cultural barriers contribute to a homogenous friend network for immigrant youth (Myers et al., 2009). As a result, earlier immigrant generations that are less assimilated are less likely to be influenced by delinquent friends, thereby contributing to their lower likelihood of delinquency. Although Dipietro and McGloin (2012) found that earlier immigrant generations are more susceptible to the effect of delinquent friends, our study does not find similar evidence. Including the measures of family control and school control might offset the risk effect of delinquent friends for first- and second-generation adolescents. Even though the friends-relevant risk remains consistent, some studies have suggested that immigrant youth tend to be more family oriented and are less vulnerable to friends’ influence (Myers et al., 2009). In sum, our study confirms the conclusion that differential association theory offers a great explanatory power for the extent of delinquency across different populations (Alarid et al., 2000; Kreager, 2007; Le et al., 2005; Neff & Waite, 2007; Whaley et al., 2010).

Although our results provide valuable information about the function of family control, school control, and association with delinquent friends to explain the immigrant paradox in crime, the limitations of this study are worthy of discussion. Our assessment of differential association uses only the measures of number of delinquent friends. The process through which delinquent friends influence the adaptive outcomes of immigrant youth has not been fully captured by these two measures. In addition to our measures, attitudinal measures such as friends’ definitions, one’s own definitions favorable toward violating law, peer pressure to engage in delinquent behaviors, and the structure and composition of friends’ network have been found to explain the link between delinquent friends and risk behaviors (Haynie, 2001; Haynie & Osgood, 2005; Warr, 2002; Whaley et al., 2010). The influence of all these factors on immigrant youth needs future exploration.

A second limitation is that this study is concerned with the differences in delinquency across immigrant generations, but categories, such as country of origin and race and ethnicity, obscure the diversity of groups of which they are comprised. One important reason for the inconsistent findings in the research examining generalizability of classical criminological theories lies in the sociodemographic characteristics of respondents. According to Min Zhou (1997, p. 65), immigrant youth are “extraordinarily diverse in national origins, socioeconomic circumstances, and settlement patterns.” Research showed that immigrants’ characteristics other than immigrant generation have strong influence on their assimilation outcomes (e.g., Decker, Gemert, & Pyrooz, 2009; Levels, Dronkers, & Kraaykamp, 2006). For example, studies found substantial heterogeneity exists among immigrants depending on their country of origins (e.g., Chiswick & DebBurman, 2004; Zhou & Bankston, 2001). Although Asian immigrants have developed a reputation as model minorities, Zhou and Bankston (2001) found that many Vietnam immigrants live in neighborhoods that are poor and socially isolated. As a result, Vietnamese youth gang have emerged and become notoriously threatening. It is logical to anticipate that moderating factors, such as race and ethnicity, country of origin, and gender, may complicate the relationship between social control, differential associations, and adolescent violence. In addition, the aggravation of third-generation youth and higher generation youth may conflate the experiences and outcomes of those who have foreign-born grandparents with those who have native-born grandparents. Indeed, some studies have suggested that the demographic characteristics and behavioral outcomes of the third-generation immigrants are distinguishable from other generational cohorts (e.g., Alba, Logan, Lutz, & Stults, 2002; Livingston & Kahn, 2002). These research evidence underscores the need for better data to help separate third-generation youth from the consequent generation youth.

These noted shortcomings notwithstanding, the present study examines the generality of classical criminological theories across family, school, and friends among immigrant youth. Our study indicates that some classical perspective of crime, such as the social control perspective, may not have comparable effect among today’s immigrant youth. This leads to the question of whether the common measures of classical theories are adequate to explain delinquent behaviors among immigrant youth. This study’s findings, along with research findings regarding other important assimilation outcomes such as health, echo the continued call for researchers to further explore, identify, and explain the mechanisms and the complexities associated with the immigrant paradox. Subsequent research that aims at explaining immigrant paradox should look at other factors affecting behavioral or health outcomes in later life, such as income, access to health care, and neighborhood quality. It is also necessary to continue assessing the generalizability of criminological theories for a rapidly growing population of immigrant youth in the United States in order to understand and address delinquency for this marginalized population.

Footnotes

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.

Note

References

Agnew

(2001). Building on the foundation of general strain theory: Specifying the types of strain most likely to lead to crime and delinquency. Journal of Research in Crime and Delinquency, 38, 319–361.

Akers

R. L.

(1985). Deviant behavior: A social learning approach (3rd ed.). Belmont, CA: Wadsworth.

Akers

R. L.

Krohn

M. D.

Lanza-Kaduce

Radosevich

(1979). Social learning and deviant behavior: A specific test of a general theory. American Sociological Review, 44, 636–655.

Akers

R. L.

Sellers

C. S.

(2004). Criminological theories: Introduction, evaluation, and application (4th ed.). Los Angeles, CA: Roxbury.

Alarid

L. F.

Burton

V. S.

Cullen

F. T.

(2000). Gender and crime among felony offenders: Assessing the generality of social control and differential association theories. Journal of Research in Crime and Delinquency, 37, 171–199.

Alba

Logan

Lutz

Stults

(2002). Only English by the third generation: Loss and preservation of the mother tongue among the grandchildren of contemporary immigrants. Demography, 39, 467–484.

Anderson

(1999). Code of the street: Decency, violence, and the moral life of the inner city. New York, NY: Norton.

Bankston

C. L.

III Zhou

(2002). Social capital and immigrant school performance: The case for a reconceptualization. Research in Sociology of Education (Schooling and Social Capital in Diverse Cultures), 13, 13–39.

Beaver

K. M.

Wright

G. P.

(2005). Evaluating the effects of birth complications on low self-control in a sample of twins. International Journal of Offender Therapy and Comparative Criminology, 49, 450–471.

10.

Bellair

Roscigno

V. J.

McNulty

T. L.

(2003). Linking local labor market opportunity to violent adolescent delinquency. Journal of Research in Crime and Delinquency, 40, 6–33.

11.

Berry

J. W.

(2006). Acculturative stress. In Wong

P. T. P.

Wong

L. C. J.

(Eds.), Handbook of multicultural perspectives on stress and coping (pp. 287–298). Dallas, TX: Springer.

12.

Bersani

B. E.

(2013). An examination of first and second generation immigrant offending trajectories. Justice Quarterly, 31, 315–343.

13.

Booth

J. A.

Farrell

Varano

S. P.

(2008). Social control, serious delinquency, and risky behavior: A gendered analysis. Crime & Delinquency, 54, 423–456.

14.

Bowman

M. A.

Prelow

H. M.

Weaver

S. R.

(2007). Parenting behaviors, association with deviant peers, and delinquency in African American adolescents: A mediated-moderation model. Journal of Youth and Adolescence, 36, 517–527.

15.

Bui

H. N.

(2009). Parent-child conflicts, school troubles, and delinquency among immigrants. Crime and Delinquency, 55, 1412–1441.

16.

Camarota

(2012). Projecting immigration’s impact on the size and age structure of the 21st century American population. Washington, DC: Center for Immigration Studies.

17.

Carmines

Zeller

(1979). Reliability and validity assessment. Beverly Hills, CA: Sage.

18.

Cernkovich

S. A.

Giordano

P. C.

(1992). School bonding, race, and delinquency. Criminology, 30, 261–291.

19.

Chang

T. N.

(2005). The influence of parents, peer delinquency and school attitudes on academic achievement in Chinese, Cambodian, Laotian or Mien, and Vietnamese youth. Crime and Delinquency, 51, 238–264.

20.

Chao

R. K.

(2001). Extending research on the consequences of parenting style for Chines Americans and European Americans. Child Development, 72, 1832–1843.

21.

Chiswick

B. R.

DebBurman

(2004). Educational attainment: Analysis by immigrant generation. Economics of Education Review, 23, 361–379.

22.

Choi

Harachi

T. W.

(2008). Intergenerational cultural dissonance, parent-child conflict and bonding, and youth problem behaviors among Vietnamese and Cambodian immigrant families. Journal of Youth and Adolescence, 37, 85–96.

23.

Daigle

L. E.

Cullen

F. T.

Wright

J. P.

(2007). Gender differences in the predictors of juvenile delinquency: Assessing the generality-specificity debate. Youth Violence and Juvenile Justice, 5, 254–286.

24.

Decker

S. H.

Gemert

Pyrooz

(2009). Gangs, migration and crime: The changing landscape in Europe and the United States. Journal of International Migration and Immigration, 10, 393–408.

25.

De Kemp

R. A. T.

Scholte

R. H. J.

Overbeek

Engels

R. C. M. E.

(2006). Early adolescent delinquency: The role of parents and best friends. Criminal Justice and Behavior, 33, 488–510.

26.

Demuth

Brown

S. L.

(2004). Family structure, family processes, and adolescent delinquency: The significance of parental absence versus parental gender. Journal of Research in Crime and Delinquency, 41, 58–81.

27.

Deutsch

A. R.

Crockett

L. J.

Wolff

J. M.

Russell

S. T.

(2012). Parent and peer pathways to adolescent delinquency: Variations by ethnicity and neighborhood context. Journal of Youth and Adolescence, 41, 1078–1094.

28.

Dillon

Pantin

Robbins

Szapocznik

(2008). Exploring the role of parental monitoring of peers on the relationship between family functioning and delinquency in the lives of African American and Hispanic adolescents. Crime and Delinquency, 54, 65–94.

29.

Dipietro

S. M.

McGloin

J. M.

(2012). Differential susceptibility? Immigrant youth and peer influence. Criminology, 50, 711–742.

30.

Dishion

T. J.

Eddy

J. M.

Haas

Spraclden

(1997). Friendship and violent behavior during adolescence. Social Development, 6, 207–223.

31.

Dorius

C. J.

Bahr

S. J.

Hoffmann

J. P.

Harmon

E. L.

(2004). Parenting practices as moderators of the relationship between peers and adolescent marijuana use. Journal of Marriage and Family, 66, 163–178.

32.

Driscoll

A. K.

Russell

S. T.

Crockett

L. J.

(2008). Parenting styles and youth well-being across immigrant generations. Journal of Family Issues, 29, 185–209.

33.

Elliott

D. S.

Huizinga

Ageton

S. S.

(1985). Explaining delinquency and drug use. Beverly Hills, CA: Sage.

34.

Gottfredson

G. D.

Gottfredson

D. C.

Payne

A. A.

Gottfredson

N. C.

(2005). School climate predictors of school disorder: Results from a national study of delinquency prevention and schools. Journals of Research in Crime and Delinquency, 42, 412–444.

35.

Hawkins

J. D.

Weis

J. G.

(1985). The social development model: An integrated approach to delinquency prevention. Journal of Primary Prevention, 6, 73–97.

36.

Haynie

D. L.

(2001). Delinquent peers revisited: A network approach for understanding adolescent delinquency. American Journal of Sociology, 106, 1013–1057.

37.

Haynie

D. L.

Osgood

D. W.

(2005). Reconsidering peers and delinquency: How do peers matter? Social Forces, 84, 1109–1130.

38.

Hirschi

(1969). Causes of delinquency. Berkeley: University of California Press.

39.

Hwang

Akers

R. L.

(2003). Substance use by Korean adolescents: A cross-cultural test of social learning, social bonding, and self-control theories. In Akers

R. L.

Jensen

G. F.

(Eds.), Social learning theory and the explanation of crime: A guide for the new century. Advances in criminological theory (Vol. 11, pp. 39–64). New Brunswick, NJ: Transaction Publishers.

40.

Jenkins

P. H.

(1997). School delinquency and the school social bond. Journal of Research in Crime and Delinquency, 34, 337–367.

41.

Jensen

G. F.

(1972). Parents, peers, and delinquent action: A test of the differential association perspective. American Journal of Sociology, 78, 562–575.

42.

Jiang

Peterson

(2012). Beyond participation: The association between school extracurricular activities and involvement in violence across generations of immigration. Journal of Youth and Adolescence, 41, 362–378.

43.

Kao

(1995). Asian Americans as model minorities? A look at their academic performance. American Journal of Education, 103, 121–159.

44.

Kao

(2004). Parental influences on the educational outcomes of immigrant youth. International Migration Review, 38, 427–449.

45.

Kao

Rutherford

L. T.

(2007). Does social capital still matter? Immigrant minority disadvantage in social capital and its effects on academic achievement. Sociological Perspectives, 50, 27–52.

46.

Kao

Tienda

(1995). Optimism and achievement: The educational performance of immigrant youth. Social Science Quarterly, 76, 1–19.

47.

Kreager

D. A.

(2007). Unnecessary roughness? School sports, peer networks, and male adolescent violence. American Sociological Review, 72, 705–724.

48.

T. N.

Monfared

Stockdale

G. D.

(2005). The relationship of school, parent, and peer contextual factors with self-reported delinquency for Chinese, Cambodian, Lao/Mien, and Vietnamese youth. Crime and Delinquency, 51, 192–219.

49.

T. N.

Stockdale

(2008). Acculturative dissonance, ethnic identity, and youth violence. Cultural Diversity and Ethnic Minority Psychology, 14, 1–9.

50.

Lee

M. T.

Martinez

Jr Rosenfeld

(2001). Does immigration increase homicide? Negative evidence from three border cities. The Sociological Quarterly, 42, 559–580.

51.

Lee

M. T.

Martinez

(2006). Immigration and Asian homicide patterns in urban and suburban San Diego. In Martinez

Valenzuela

(Eds.), Immigration and crime: Race, ethnicity, and violence (pp. 90–116). New York: New York University Press.

52.

Levels

Dronkers

J. D.

Kraaykamp

(2006). Educational achievement of immigrant children in western countries: Origin, destination, and community effects on mathematical performance. American Sociological Review, 73, 835–853.

53.

Little

R. J.

Rubin

D. B.

(1987). Statistical analysis with missing data. New York, NY: John Wiley.

54.

Livingston

Kahn

J. R.

(2002). An American dream unfulfilled: The limited mobility of Mexican Americans. Social Science Quarterly, 83, 1003–1012.

55.

Martinez

Jr Lee

M. T.

(2000). On immigration and crime. In LaFree

Sr Bursik

R. J.

Short

J. F.

Taylor

R. B.

(Eds.), Criminal justice 2000: The nature of crime: Continuity and change (Vol. 1, pp. 485–524). Washington, DC: U.S. Department of Justice.

56.

Martinez

Jr Lee

M. T.

Nielsen

A. L.

(2004). Local context and determinants of drug violence in Miami and San Diego: Does ethnicity and immigration matter? International Migration Review, 38, 131–157.

57.

Martinez

Jr Stowell

Lee

M. T.

(2010). Immigration and crime in an era of transformation: A longitudinal analysis of homicides in San Diego neighborhoods, 1980–2000. Criminology, 48, 797–829.

58.

Matsueda

R. L.

Heimer

(1987). Race, family structure and delinquency: A test of differential association and social control theories. American Sociological Review, 52, 826–841.

59.

Mears

D. P.

(2001). The immigration-crime nexus: Toward an analytic framework for assessing and guiding theory, research and policy. Sociological Perspectives, 44, 1–19.

60.

Miller

W. B.

(1958). Lower class culture as a generating milieu of gang delinquency. Journal of Social Issues, 14, 5–19.

61.

Morenoff

J. D.

Astor

(2006). Immigrant assimilation and crime: Generational differences in youth violence in Chicago. In Martinez

Jr Valenzuela

Jr (Eds.), Immigration and crime: Race, ethnicity, and violence (pp. 36–63). New York: New York University Press.

62.

Myers

Chou

C. P.

Sussman

Baezconde-Garbanati

Pachon

Valente

T. W.

(2009). Acculturation and substance use: Social influence as a mediator among Hispanic alternative high school youth. Journal of Health and Social Behavior, 50, 164–179.

63.

Neff

Waite

(2007). Male versus female substance abuse patterns among incarcerated juvenile offenders: Comparing strain and social learning variables. Justice Quarterly, 24, 105–132.

64.

Nettler

(1978). Explaining crime. New York: McGraw-Hill.

65.

Nofziger

Lee

(2006). Differential associations and daily smoking of adolescents: The importance of same sex models. Youth & Society, 37, 453–478.

66.

Nye

F. I.

Short

J. F.

Jr (1957). Scaling delinquent behavior. American Sociological Review, 22, 326–331.

67.

Peguero

A. A.

(2013). An adolescent victimization immigrant paradox: School-based routines, lifestyles, and victimization across immigration generations. Journal of Youth and Adolescence, 42, 1759–1773.

68.

Peguero

A. A.

Bondy

J. M.

Hong

J. S.

(in press). Social bonds across immigrant generations: Bonding to school and examining the relevance of assimilation. Youth & Society. doi:10.1177/0044118X14560335

69.

Peguero

A. A.

Jiang

(2014). Social control across immigrant generations: Adolescent violence at school and examining the immigrant paradox. Journal of Criminal Justice, 42, 276–287.

70.

Pong

Hao

Gardner

(2005). The roles of parenting styles and social capital in the school performance of immigrant Asian and Hispanic adolescents. Social Science Quarterly, 86, 928–950.

71.

Portes

Rumbaut

(2001). Legacies: The story of the immigrant second generation. Berkeley: University of California Press.

72.

Portes

Zhou

(1993). The new second generation: Segmented assimilation and its variants. Annals of the American Academy of Political and Social Sciences, 530, 74–96.

73.

Reid

Williams

Harald

Adelman

R. M.

Jaret

(2005). The immigration-crime relationship: Evidence across US metropolitan areas. Social Science Research, 34, 757–780.

74.

Sampson

R. J.

(2008). Rethinking crime and immigration. Contexts, 7, 28–33.

75.

Sampson

R. J.

Bean

(2006). Cultural mechanisms and killing fields: A revised theory of community-level racial inequality. In Peterson

Krivo

Hagan

(Eds.), The many colors of crime: Inequalities of race, ethnicity and crime in America (pp. 8–29). New York: New York University Press.

76.

Schafer

J. L.

Graham

J. W.

(2002). Missing data: Our view of the state of the art. Psychological Methods, 7, 147–177.

77.

Schafer

J. L.

Olsen

M. K.

(1998). Multiple imputation for multivariate missing-data problems: A data analyst’s perspective. Multivariate Behavioral Research, 33, 545–571.

78.

Shaw

C. R.

McKay

H. D.

(1942/1969). Juvenile delinquency and Urban areas (Rev. ed.). Chicago, IL: University of Chicago Press.

79.

Soller

Haynie

(2013). Structuring the future: Anticipated life events, peer networks, and adolescent sexual behavior. Sociological Inquiry, 83, 1–33.

80.

Steffensmeier

Allan

(1996). Gender and crime: Toward a gendered theory of female offending. Annual Review of Sociology, 22, 459–487.

81.

Stewart

E. A.

(2003). School social bonds, school climate, and school misbehavior: A multilevel analysis. Justice Quarterly, 20, 575–604.

82.

Stowell

J. I.

Messner

S. F.

McGeever

K. F.

Raffalovich

L. E.

(2009). Immigration and the recent violent crime drop in the U.S.: A pooled cross-sectional time-series analysis of metropolitan areas. Criminology, 47, 889–928.

83.

Sutherland

E. H.

(1947). Principles of criminology (4th ed.). Philadelphia, PA: Lippincott.

84.

Turney

Kao

(2009). Pre-kindergarten childcare and behavioral outcomes among children of immigrants. Early Childcare Research Quarterly, 24, 432–444.

85.

Vartanian

T. P.

Karen

Buck

P. W.

Cadge

(2007). Early factors leading to college graduation for Asians and non-Asians in the United States. The Sociological Quarterly, 48, 165–197.

86.

Vitaro

Brendgen

Tremblay

R. E.

(2000). Influence of deviant friends on delinquency: Searching for moderator variables. Journal of Abnormal Child Psychology, 28, 313–325.

87.

Warr

(2002). Companions in crime: The social aspects of criminal conduct. New York, NY: Cambridge University.

88.

Warr

Stafford

(1991). The influence of delinquent peers: What they think or what they do? Criminology, 29, 851–866.

89.

Waters

(1999). Crime and immigrant youth. Thousand Oaks, CA: Sage.

90.

Waters

(1999). Black identities: West Indian immigrant dreams and American realities. New York, NY: Russell Sage Foundation.

91.

Whaley

R. B.

Hayes-Smith

(2010). Gendered pathways? Gender, mediating factors, and the gap in boys’ and girls’ substance use. Crime & Delinquency, 59, 651–669.

92.

Wolfgang

M. E.

Ferracuti

(1967). The subculture of violence: Towards an integrated theory in criminology. Beverly Hills, CA: Sage.

93.

Wright

J. P.

Cullen

F. T.

(2001). Parental efficacy and delinquent behavior: Do control and support matter? Criminology, 39, 677–706.

94.

Zhou

(1997). Growing up American: The challenge confronting immigrant children and children of immigrants. Annual Review of Sociology, 23, 63–95.

95.

Zhou

Bankston

C. L.

III (2001). Family pressure and the educational experience of the daughters of Vietnamese refugees. International Migration, 39, 133–151.