Post-secondary Educational Attainment of Immigrant ... - EBSCOhost

Post-secondary Educational Attainment of Immigrant and Native Youth Ursula Keller, Florida State University Kathryn Harker Tillman, Florida State University We examine immigrant generation differences in college attendance and college type among youth ages 18 through 26 who have graduated from a US. high school. Results indicate thatfirstand second-generation immigrants are significantly more likely to attend college than their third-plus generation counterparts of similar race/ethnicity, socioeconomic and family background characteristics. While parental behaviors and expectations for college attendance do not significantly mediate these generational differences, these factors appear to indirectly affect college-going behavior through their impact on students' verbal ability and academic achievement during high school. Interaction models including race/ethnicity and generation status reveal that the second-generation effects on college attendance are largely driven by Chinese youth, whereas thefirst-generationeffects on college attendance are largely driven by black immigrant students. Introduction The United States' population is beconning increasingly diverse. Today, imnnigration accounts for more than one-third of the nation's total observed annual demographic increase. The foreign-born population is currently hovering at 34 million, which corresponds to about 12 percent of the total U.S. population, the highest percentage observed in more than 80 years (Camarota 2004). Unlike the early 1900s, when the majority of immigrants came from Europe, the majority of immigrants to the United States in 2000 were born in Latin America or in Asia. It is estimated that at least 40 percent of Latinos and 60 percent of Asians in the United States today are foreign-born (Schmidley 2000). First- and second-generation immigrants also make up a considerable proportion of students in the United States. Approximately one-fifth of all school-aged children are immigrants or the children of immigrants (Suarez-Orozco and Suarez-Orozco 2001). School-aged immigrants are the fastest growing and the most ethnically diverse segment of America's Special thanks are due to John R. Reynolds, Isaac W. Eberstein and Rebecca Clark for their helpful comments and suggestions. This research utilizes data from Add Health. Direct correspondence to Ursula Keller, Department of Sociology, 526 Bellamy Building, Florida State University, Tallahassee, FL 32306-2270, USA. E-mail: [email protected].' © The University of North Carolina Press

Social Forces 87(1 )

122 • Socia/Forces 87(1) . September 2008

child population (Schmid 2001). In 1997, the United States had 3 million foreign-born children under 18 years of age, and almost 11 million U.S.born children under 18 were living with at least one foreign-born parent (Alba, Massey and Rumbaut 1999). As the large influx of immigrants to the United States persists at a level approaching 1 million per year, the number of immigrant children is likely to increase as well. Therefore, the future of American society is ultimately related to the adaptation of immigrants and their children, eyen with possible efforts to reduce immigration. The majority of immigrants come to the United States in search of the "American Dream" and/or to provide their children with an enhanced opportunity for security and financial advancement. Educational aspirations are universally high among all adolescents, as most young people anticipate attending college (Kao and Thompson 2003) and ending up in a professional job (Reynolds, Stewart, MacDonald et al. 2006). Research on educational achievement in primary and secondary schools, however, suggests that immigrant children experience an advantage when compared their nativeborn peers of similar race/ethnicity and socioeconomic characteristics (e.g., Kao 1999; Kao and Tienda 1995; Portes and Rumbaut 1996, 2001; Rumbaut 1997; Tillman, Guo and Harris 2006). Furthermore, immigrant parents' behaviors and expectations tend to raise their children's expectations of future college attendance above those of other children (Goyette and Xie 1999). Only recently has research begun to examine whether and how immigrant youth who have been schooled in the United States are able to translate their childhood achievement and expectations into adult socioeconomic outcomes (Feliciano and Rumbaut 2005; Glick and White 2004). This article examines generational differences in post-secondary educational outcomes among young adults who have been raised in the United States. We specifically seek to better understand immigrant educational arrangements in the years immediately following high school graduation. Learning more about generational differences in post-secondary education is important because the success with which individuals navigate the educational system during this time period is crucial to their long-term employment, occupational and economic wellbeing (e.g., Haveman et al. 1994). A more complete understanding of the association between immigrant generation and college enrollment, therefore, could aid in the development of useful programs to target young people at risk for lowered socioeconomic attainment. We use data from the National Longitudinal Study of Adolescent Health to explore immigrant generation and race/ethnic differences in predicting college attendance among young adults ages 18 through 26 who attended secondary school in the United States and who have a high school diploma or GED. In particular, we address four research questions:

College Attendance of Immigrant and Native Youth »123

(1. Net of basic socio-demographic characteristics, is innmigrant generation associated with college attendance during early adulthood? (2. Can the association between immigrant generation and college attendance be explained by generational differences in parental behaviors, academic expectations and/or academic ability and achievement while in high school? (3. Does type of college attendance (two-year vs. four-year institution) vary significantly by immigrant generation? (4. Is the association between immigrant generation and college attendance similar across different racial/ethnic groups, or are the outcomes of youth from some racial/ethnic groups more affected by immigrant generation status? Immigrant Educational Outcomes Although the general public perception, particularly in areas with large foreign-born populations, is that immigrant children undermine the quality of the educational system, research shows that immigrant children generally perform quite well in school, often out-performing native-born peers of similar socio-demographic characteristics (Kao 1999; Kao and Tienda 1995; Portes and Rumbaut 1996, 2001 ; Rumbaut 1997; Tillman, Guo and Harris 2006). Yet, educational and economic progress among immigrant groups is extremely uneven. In 1990, for instance, only 74 percent of Mexican immigrants ages 15 through 17 attended school compared with 95 percent of native and other immigrant youth (Hao and Bonstead-Bruns 1998). Furthermore, Hispanic immigrants are still among the least likely of all youth to graduate from either high school or college (Arbona and Nora 2007). Asian immigrants, conversely, tend to out-perform most other children in the United States (Vernez and Abrahamse 1996; Zhou and Bankston 1994) and are frequently referred to as the "model minority."(Kao 1995) Thus, while immigrant children are successfully navigating the U.S. education system in general, children from all racial/ethnic groups may not follow the same trajectories of adaptation and scholastic success (Kao and Tienda 1995; Portes and Rumbaut 2001). In fact, some research has suggested that the effects of generational status on children's academic outcomes are conditioned by racial/ethnic group, such that Asians are most likely and Hispanics are least likely to benefit from having foreignborn parents. Black immigrants, meanwhile, may benefit most from their own foreign birth (Kao and Tienda 1995). Although much research has been dedicated to the examination of educational achievement and attainment during childhood and adolescence, only recently have researchers begun to examine the educational attainment of immigrant youth after high school or explore differences in post-secondan/ schooling among the various immigrant groups in the United States. This research on immigrant status and college attendance has, however, shown

124 • Social Forces 87(1) . September 2008

some significant generational effects. For example, several studies have found that young adults of first-generation (Song and Glick 2004; Vemez and Abrahamse 1996) and second-generation status (Glick and White 2004) are more likely than their native-born, third-plus generation peers to complete high school and enroll in post-secondary education. There may also be generational differences in the type of postsecondary education that young people obtain. Whether a student attends a two-year or a four-year institution is associated with his or her future socioeconomic well-being, as two-year community colleges and associate's degree programs tend to be more affordable, but are less prestigious and less financially rewarding than are four-year degree programs (Jaeger and Page 1996; Kane and Rouse 1995). Young people of lower socioeconomic backgrounds, racial/ethnic minorities, those with language barriers, and other "non-traditional" students are more likely to enroll in two-year colleges, as opposed to four-year public or private institutions (Dougherty 1994; Santibanez et al. 2007). The sociodemographic characteristics of many immigrant groups, therefore, increase the likelihood that immigrant youth, particularly those of the firstgeneration, will obtain their education in a two-year college environment (Bryant 2001). Indeed, evidence suggests that the foreign-born are more likely than the native-born to either begin their post-secondary education at a two-year institution and go no further or to begin their post-secondary education at a two-year institution and then transfer to a four-year institution (Vemez and Abrahamse 1996). These broad nativity differences, however, may obscure important distinctions between firstgeneration youth who have spent a substantial portion of their lives in the United States and those who immigrated as young adults. Furthermore, the findings may be largely driven by the fact that Hispanic youth are among the most likely to attend two-year colleges (Nora, Rendón Cuadraz 1999; Vernez and Abrahamse 1996). In fact, race/ethnicity may have a stronger direct effect on post-secondary experience than does nativity status (Vernez and Abrahamse 1996), and prior research strongly suggests the need to investigate the interaction between immigrant generation (clearly differentiating between the second- and third-plus generation) and race/ethnicity when examining both college attendance and choices regarding college type (Glick and White 2004; Hagy and Staniec 2002; Kao and Tienda 1995). To date, most studies that have focused on the post-secondary educational outcomes of immigrants have been limited in scope by their data. For example, some of these analyses have lumped together first-generation immigrants who came to the United States as young adults with first-generation youth who were raised and educated in the American school system. In addition, previous studies have not examined

College Attendance of Immigrant and Native Youth • 125

whether there are significant interactions between immigrant generation and Xhe specific racial/ethnic groups to which individuals belong. At best, earlier research was only able to examine the conditioning influence of membership within broad pan-ethnic groups (e.g., "Hispanic" and "Asian"), which may obscure significant subgroup differences in the association between generation status and college attendance outcomes (Bailey and Weininger 2002; Glick and White 2004).^ Furthermore, only one nationally-representative study that we know of (Glick and White 2004) has explicitly examined mechanisms other than basic socio-demographic characteristics that may help to explain generation differences in post-secondary education. While Glick and White (2004) provide evidence for the importance of parental behaviors and expectations in mediating generation status differences, controlling for these factors does not completely "explain away" the greater likelihood of immigrant youth to attend college. Their research is also not able to examine the relative importance of parental behaviors/expectations and past academic performance. Over the course of late adolescence and early adulthood, the effects of parental and familial influences on young people's behavior begin to wane. However, the family environments that encourage academic success during the childhood years may have continuing indirect effects upon college enrollment via the influence of academic achievement, ability and participation in college-preparatory coursework during high school (Charles, Roscigno and Torres 2007; Vernez and Abrahamse 1996). Using the Add Health data, the present study is able to directly examine these issues. Assimilation Theory and Immigrant Education Assimilation theories have long been used to explain the association between immigrant generation and educational outcomes. According to traditional assimilation theory, immigrants, particularly the first-generation, are frequently held back by their newcomer status and are rarely expected to achieve socioeconomic equivalence with the native population. Instead, the social and economic outcomes of immigrants are expected to be lower than those of their native-born peers because of the trauma and stress, as well as the social, economic, and linguistic disadvantages, associated with the immigration process. However, as the foreign-born acculturate across length of residence and increasing generation, these differentials with the native-born population will gradually narrow (Warner and Srole 1945). While this traditional assimilation theory seemed to be well-suited for explaining the trajectories of socioeconomic achievement among immigrants at the turn of the 20"" century, it has been less useful in explaining trends among contemporary immigrants.

126 • Social Forces S7{1) . September 2008

Segmented assimilation theory, on the other hand, suggests that today's immigrants experience divergent pathways of assimilation that may result in either upward or downward mobility for various immigrant groups. Recent research has found that first-generation immigrants are not always at a disadvantage socially and economically. Assimilation may actually lead to a deterioration of outcomes over time and generation spent in the United States (Crosnoe and Lopez-Gonzalez 2005; Harker 2001; Harris 1999; Hirschman 2001; Portes and Rumbaut 2001; Portes and Zhou 1993; Rumbaut 1997; Tillman, Guo and Harris 2006; White and Glick 2000). Hence, one can conclude that immigrant status may occasionally be protective against social and economic hardship. This theory may explain why many immigrant school children, despite being faced with serious disadvantages, have very strong academic outcomes as compared to those of their same-race native-born counterparts. The ways in which assimilation affects the outcomes of immigrants may be conditioned by the human and cultural capital those immigrants bring with them, in addition to the social contexts into which they settle. In terms of human and cultural capital, we investigate socio-demographic and family context measures, including socioeconomic status, family composition, parental behaviors and expectations for academic attainment. Studies indicate that these factors exercise a strong influence on children's educational achievement and plans for the future (Glick and White 2004; Kao and Tienda 1995; Portes and Rumbaut 2001 ). While immigrant families may experience greater risks associated with low socioeconomic status, research finds that immigrant parents generally hold higher expectations for their children and maintain greater parental control than native-born parents and, consequently, place greater demands on their children in terms of school engagement and success (Glick and White 2004; Hao and Bonstead-Bruns 1998). This may lead to higher achievement during childhood and to higher motivation to attend college as a young adult. Higher levels of achievement during childhood, in turn, may place immigrant youth in a position where they are more qualified than their non-immigrant peers to actually pursue a college education and/ or to attend a four-year institution (Vernez and Abrahamse 1996). Yet, the socio-demographic and family context characteristics associated with academic outcomes vary significantly by specific race/ethnic background (Bohon, Johnson and Gorman 2006; Charles, Roscigno and Torres 2007). For example, great diversity exists in socioeconomic status, cultural norms and behaviors, and other resources that set specific immigrant groups apart from one another, as well as from the native-born population. Even within large pan-ethnic immigrant groups such as Hispanics and Asians, some ethnic subgroups may be more likely than others to have the social and financial resources to assist their children's academic

Coilege Attendance of Immigrant and Native Youth «127

performance and finance their post-secondary education, to hold high expectations for academic performance and college attendance, and to maintain control over their children's activities and decisions (Bohon, Johnson, and Gorman 2006; Portes and Rumbaut 2001; Rumbaut 1997; Suarez-Orozco and Suarez-Orozco 2001). It is, therefore, essential to explore the interactions between immigrant generation and race/ethnic group, as well as the direct effects of immigrant generation, on the collegegoing behavior of American youth. Based on findings of previous studies and these theoretical perspectives, we have developed five main predictions regarding immigrant status and educational outcomes. First, we predict a generational disparity in college attendance for young adults, where first- and second-generation individuals are more likely to attend college than their third-generation counterparts of similar socio-demographic backgrounds. Second, we predict that parental behaviors (e.g., parental involvement in school and parental control) and academic expectations, as reported during adolescence, will play an important role in explaining the higher likelihood of college attendance among immigrants. Third, we predict that higher academic achievement and greater college preparation in high school will also play a significant role in the immigrant educational advantage, and may at least partially mediate the effects of parental behaviors during adolescence. In other words, beneficial parenting behaviors during the teen years may continue to affect college enrollment indirectly via the influence of earlier high school outcomes. Fourth, we predict generational differences in the type of college attended, such that first- and second-generation immigrants will be more likely than their third-generation peers to attend a two-year institution, and that these generational differences will be largely mediated by socio-demographic variables. Lastly, we expect that the association of immigrant generation with college attendance and college type will vary significantly by race/ethnicity. Methods Data We use Add Health, a nationally representative study of adolescents in grades 7 through 12 in the United States in 1995. Based on a multistage, stratified, school-based, cluster sampling design, this study was designed to explain the causes of adolescent health and health behavior, primarily focusing on the multiple contexts in which adolescents live. Included in the sample were students from 80 high schools (both public and private) and a corresponding feeder junior high/middle school. Most minority ethnic groups were sampled in proportion to their size within the United States population, however smaller ethnic groups were oversampled (Harris et al. 2003).

128 • Social Forces S7(l) . September 2008

Data for this study were collected in three waves from 1994 through 2002. In Wave I all students were between the ages of 12 and 21 years and by Wave III the students were between the ages of 18 and 26 years. Wave III data contains follow-up interviews with 14,979 original Wave I respondents and pre-test data contain an additional 218 respondents, for a total of 15,197 respondents. We use data from Wave I and Wave III, drawing from the extensive in-home interviews conducted at both times. In addition, we use the parental questionnaire from Wave I and the high school Education Data released with Wave III (Harris et al. 2003). Our sample includes all respondents who have fully completed in-home interviews for both Wave I and Wave III and have graduated from high school or received a GED in the United States.^ Due to these criterion, all respondents in the sample had been living in the United States for a minimum of seven years and had not likely come to this country for the sole purpose of obtaining a higher education. For consistency purposes, the central predictor and most control variables are taken from Wave I. The outcome variables, college attendance and type of college attended, and the linked high school Education Data are taken from Wave III. Our final analytic sample includes 10,163 young adults. Measures The dependent measures in this study are any college attendance and type of college currently attended. We first examine college attendance with a dichotomous variable measuring whether the respondent is: currently attending college or has any college education (1), or not currently attending college and has no college education (0). We then explore current college attendance by type: not currently attending college (0), currently attending two-year college ( 1 ), or currently attending four-year college (2). Immigrant generation is determined by both the youth's and the parents' country of origin as specified in Wave I. We classify respondents as first-, second- or third-plus generation. First-generation immigrants are respondents who were born abroad (and not as a U.S. citizen). Secondgeneration individuals are those who have at least one parent of foreign birth, but who themselves were either born in the United States or in a foreign country as a U.S. citizen. Finally, all respondents who were born in the United States to parents who were also born in the United States are classified as the third-plus generation. The demographic variables of sex, age and race/ethnicity are taken from the Wave I in-home interview. Race/ethnicity is self-identified and is measured with a series of eight dummy variables, including Mexican, Cuban, Central/South American, Chinese, Filipino, Other Asian/Pacific


Islander, Black and White." Region of country, location of residence and middle/high school type are obtained through the Wave I school administrator questionnaire. Region of country is broken down into four dummy variables indicating West, Midwest, South and Northeast, and location of residence is measured with three dummy variables indicating urban, suburban and rural. We also incorporate school type into our analyses; this variable is dichotomized into public and private schools. In addition, we examine several variables that represent the structural and cultural features of the family environment at Wave I. Five dummy variables are created to measure annual family income (measured in thousands): $15,000 or less, $16-34,000, $35-59,000, $60,000 or more, and missing income data. Highest educational attainment achieved by either mother or father (henceforth referred to as "family education") is measured with four dummy variables: less than high school, high school graduate or GED, more than high school, and missing education data. Family structure is also measured with four dummy variables: two-biological parents, stepparents, single-parent, and other family forms (e.g. grandparents, other relatives, group homes). Mother's work status is measured as a dichotomous variable specifying whether a mother had worked for pay outside of the home for more than 35 hours per week in the past 12 months. Finally, primary language spoken in the home is measured with three dummy variables: English, Spanish and an "other" language. Parental control, parental involvement and expectation measures are also incorporated in this analysis. Parental control is an index representing the mean item score of six questions (original responses ranging from 1 to 5) and assesses how much control youth feel they have over their own lives (Cronbach's alpha = .60). Examples of the items include: "Do your parents control how much television you watch on a daily basis?," "Do your parents control the type of clothing you wear?," and "Do your parents control how much time you spend out on a weeknight?" Parental involvement is measured with four dummy variables that indicate whether the respondents had engaged in specific activities with their parents over the past month (spoken with parents regarding schoolwork and grades, spoken with parents about other school related issues, worked on school projects with parents, and had an argument about behavior). Parental expectations are taken from the parental questionnaire, which asked "How disappointed would you be if [name of child] did not graduate from college?" The response categories are in Likert-type scale ranging from being "very disappointed" to "not disappointed." We also assess youth's own expectations for attending college with the question: "How likely do you think you will attend college?" Due to the overwhelming proportion of respondents who answered "very likely," we dichotomize this variable into (1 ) "very likely" and (0) all other responses.

130 . Socio/Forces 87(1) . September 2008

Finally, we use the Add Health's Wave III Picture Vocabulary Test scores and the Wave III High School Education Data to evaluate individual level academic ability and achievement during high school.^ The AHPVT measures knowledge of vocabulary, both quantity and quality. Raw test scores, which ranged from 1 to 87, were standardized by age, with each age group having a mean of 100 and a standard deviation of 15. Cumulative grade point average and highest math and science course taken for credit in high school are acquired through the Education Data, which include transcripts provided directly by each high school. Respondents are coded as having taken above-level courses, average-level courses, belowaverage level courses, or no math and/or science courses. The fact that this information is provided by the schools rather than from self-reports increases our confidence that these measures of high school achievement are accurate. Analysis and Results We use binary logistic regression to estimate the effects of sociodemographic characteristics, parental behaviors (i.e., parental control, parental involvement) and expectations, and high school ability/achievement on the likelihood of college attendance, and multinomial logistic regression to estimate the effects of these factors on type of college attended. The regression analyses account for the multistage, stratified, school-based, cluster sampling design of Add Health by using the robust estimator of variance procedure in STATA (also known as the Huber or White estimator of variance). In addition, we control for differential sampling probabilities among individuals by utilizing the Add Health grand sample weights in all estimation procedures (Chántala and Tabor 1999). For the analysis of both outcome variables, the first model examines generational differences in attendance, controlling for socio-demographic variables. In the second model, we incorporate measures of parental control, parental involvement and college expectations. Lastly, individuallevel student achievement variables are included in the full model. After the first and third of the three main models in the binary logistic regression analysis we examine an additional model with interactions between race/ ethnicity and generation status. Descriptive Analysis Table 1 presents weighted means for any college attendance and type of college currently attended by immigrant generation. The sample mean for any college attendance is .638. Broken down by type of college attendance, the sample means for current attendance at two-year and four-year institutions are .163 and .354, respectively. Hence, about 64


.364 .413 .346

1

.478 .354

.481 .492 .476

ollege

percent of all young adults who have connpleted high school in the United States are likely to experience some college education or more, and about 52 percent are likely to be currently enrolled in a two-year or four-year college. Consistent with current research (Glick and White 2004; Song and Glick 2004; Vernez and Abrahamse 1996), when college attendance is examined by immigrant status, firstand second-generation immigrants CO m have relatively higher levels of college co co CO CD attendance than their third-plus generation co" CD counterparts. This finding holds true of college attendance in general, as well as college attendance at both two-year and CO o four-year institutions. Moreover, even before 13 re a> adjusting for important compositional O differences between generations, a oneO u. way ANOVA test confirms that seconda> generation youth are significantly (p < .01 ) O] CD CJ2 d eu œ CO more likely than third-plus generation youth CO ^ r "o C/5 CD co CO C O to have some college or more. Although a> re 3 the U.S. immigrant population is growing O co ro CX3 co en CJ> CX3 in CD in number, they still suffer from social i and economic inequalities that are highly 's T) associated with an increased risk of poor co WJ a> $60K) Less than $15,000 $16-34,000 $35,000-59,000 Missing

iViodei 3b 1.568 (.59) 1.381 (1.46)

2.145" (2.87) 1.902" (4.36)

Generation 1

Gender (Maie = 1)

iViodei 2

iVIodei 3 1.635 (1.82) 1.533" (2.63)

iVIodei 1

.360" (-7.53) .423" (-8.73) .695" (-3.99) .525" (-6.04)

Continued on the following page.

iViodei 1b

136 • Social Forces 87(1) • September 2008

Table 4 continued Model 1

Model 1b

Model 2

Model 3

Model 3b

1.112 (1.69)

1.11 (1.63)

1.068 (.97)

1.056 (.69)

1.056 (.69)

Region of Country (ref. Midwest) West .985 (-.08) South 1.035 (.23) Northeast 1.274

.993 (-.04) 1.041 (.27) 1.270

1.086 (.49) .980 (-.14) 1.302

1.038 (.22) .907 (-.68) 1.396*

1.043 (.24) .909 (-.67) 1.385

(1.35)

(1.33)

(1.59)

(1.99)

(1.95)

Location of Residence (ref. Rural) Urban .881 (-.83) Suburban 1.127 (1.02)

.878 (-.85) 1.129 (1.02)

.888 (-.80) 1.094 (.71)

1.018 (.12) 1.166 (1.18)

1.017 (.12) 1.164 (1.16)

iVIiddie/High Schooi Type (Pubiic = 1)

.415" (-4.04)

.444** (-4.29)

.584** (-3.27)

.587** (-3.26)

Family Structure (ref. Bioiogicai parents) Stepparents .556" .553** (-6.74) (-6.77) Single parent .925 .932 (-.92) (-.83) Other .400** .400" (-4.39) (-4.38)

.575" (-6.02) .929 (-.87) .426** (-3.93)

.692" (-3.54) 1.138 (1.31) .519** (-3.14)

.690** (-3.55) 1.141 (1.34) .519" (-3.13)

.941 (-1.93)

.941 (-1.70)

.946 (-1.63)

.816 (-.80) 2.701" -2.88

.766 (-1.15) 1.968 -1.82

.743 (-1.46) 1.893 -1.55

.628" (-3.19)

1.084 (.52)

1.077 (.47)

.944 (-.81) .986 (-.13)

.953 (-.58) 1.000 (-.00)

.952 (-.60) 1.002 (.01)

Mother's Work Status (Working = 1)

# of Sibiings at Home

.414" (-4.05)

.936* (-2.26)

.937* (-2.22)

Language Spoken at Home (ref. English) Spanish .783 .803 (-.96) (-.84) Other 2.492** 2.118 -2.78 -1.95 Parental Control

Parental Involvement Talked about school work/ grades Worked on school project

College Attendance of Immigrant and Native Youth «137

Table 4 continued Model 3

Model 3b

1.040 (.45) 1.052 (.76)

1.041 (.47) 1.049 (.72)

Parental College Expectations (ref. Very Disappointed) .817** Somewhat disappointed (-2.63) .488** Not disappointed (-2.28) .727* Missing (-2.28)

.918 (-1.03) .689** (-2.97) .928 (-.51)

.913 (-1.10) .686** (-3.01) .934 (-.46)

Students College Expectations (ref. Not Very Likely]1 2.840** Very likely (15.12)

1.867** (8.35)

1.860** (8.24)

1.019** (6.36)

1.018** (6.30)

.206** (-3.81) .294** (-6.91) .550** (-5.59)

.207** (-3.79) .294** (-6.90) .550** (-5.63)

.565 (-1.64) .439** (-5.54) .752** (-3.93)

.560 (-1.66) .438** (-5.56) .757** (-3.83)

2.693** (15.89)

2.699** (16.00)

Model 1

Model 1b

1.158 (1.89) .893* (-1.96)

Talked about things in school Had an argument about behavior

AHPVT Scores (range 7-122) Highest Math for HS Credit (ref. Advanced) No math credit received Below average Average

Highest Science for HS Credit (ref. Advanced) No science credit received Below average Average

Cumulative Overall GPA in HS Interactions: Mexican - Generation 1 Mexican - Generation 2 Cuban - Generation 1

Continued on the following page.

1Model 2

.935 (-.08) 1.311 (.70) .075* (-2.38)

.979 (-.02) 1.118 (.26) .276 (-1.18)

138 • Soc/fl/Forces 87(1) . September 2008

Table 4 continued Model 1 Cuban - Generation 2

Model 2

Model 3

.146*

Central/South American - Generation 1 Central/South American - Generation 2 Chinese - Generation 1 Chinese - Generation 2 Filipino - Generation 1 Fiiipino - Generation 2 Other Asian - Generation 1 Other Asian - Generation 2 Black - Generation 1 Biack - Generation 2

Pseudo R-squared -2 Pseudo Log Likelihood (N = 10,163)

Model 1b

Model 3b .551

(-¿.14) 1.088 (.12) 1.688 (1.10) 13.781 (1.66) 49.475** (4.03) .450 (-1.08) .435* (-2.21) 1.825 (.63) 1.283 (.40) 27.767* (2.48) 1.825 (1.40)

(-.71) 1.525 (.53) 1.493 (.90) 7.323 (1.44) 32.080** (4.12) .180* (-2.21) .195** (-2.79) 1.142 (.41) 1.045 (.06) 19.655* (2.09) 1.924 (1.22)

14.81

14.97

19.97

33.86

33.97

11330.38

11310.01

10644.66

8797.52

8782.85

Note: Odds Ratios and Z-Scores *p < .05 **p < .01

of the race/ethnic categories are significant predictors of this outcome. This unexpected finding may result from the fact that the sample is selective of those who have graduated from a high school in the United States, disproportionately excluding youth from race/ethnic groups with lower high school graduation rates and youth who immigrated to the United States for the purpose of a post-secondary education. Overall, however, compared to their third-plus generation peers of similar sociodemographic characteristics, first- and second-generation youth who were raised and educated in the United States are more likely to engage in advanced education. In fact, first-generation youth have 114.5 percent higher odds and second-generation youth have 90.2 percent higher odds than their peers of obtaining at least some college education.

College Attendance oflmmigrant and Native Youth . 139

In Model 1 b, we add interactions between race/ethnicity and generation status. The most striking findings appear for Chinese and black youth. Second-generation Chinese and first-generation black students have, respectively, 49,5 and 27,8 higher odds (4,847,5 and 2,676,7 percent) of attending college than U,S,-born white students, after adjusting for socioeconomic status and background characteristics. First- and secondgeneration Cubans, on the other hand, have significantly lower odds of attending college than their white third-generation peers. Overall, we conclude from these interaction models that the positive secondgeneration effects on college attendance shown in Model 1 are largely driven by Chinese youth, whereas black immigrants drive the positive effects of first-generation status. The lack of significance for the majority of the interactions indicates that for most other racial/ethnic groups, generational differences in college attendance are not as important. The next model (Model 2) incorporates parental control, parental involvement, and both parental and self-reported expectations for college attendance (all measured while in high school). Both parental and self-reported expectations have significant direct effects on college attendance, confirming previous research findings on the importance of family beliefs about education for long-term academic outcomes. Level of parental control and having argued with parents about behavioral issues during adolescence are both negatively related to the likelihood of college attendance during early adulthood. None of the other parental involvement measures, however, are significant predictors of college attendance. Thus, many of the parental behaviors that are experienced during adolescence do not appear to exert a continuing influence upon academic attainment during early adulthood. Moreover, contrary to expectations, the addition of these variables does not significantly mediate the association between first- and second-generation status and college attendance. Immigrants remain significantly more likely to attend college than their third-generation peers. When we add interactions between immigrant generation and race/ ethnicity, we observe roughly the same findings as in Model 1b (results available upon request). With the inclusion of individual-level high school achievement and verbal ability measures in Model 3, however, the association between generation status and college attendance, particularly the benefit of being firstgeneration, is partially explained. As expected, AHPVT scores, measures of highest mathematic and science credits received, and cumulative grade point average during high school are all significantly associated with college attendance. For example, students who took no math, below average-level math courses or average-level math courses in high school have 79.5,70,6, and 45 percent lower odds, respectively, of attending college than those students who took above average-level math classes. Moreover, a one-unit

140 . Social Forces 87(1) • September 2008

increase in cumulative GPA is associated with a 169.3 percent increase in the odds of attending college. The fact that the inclusion of these variables explains a significant portion of the generational effects (77.3 percentforfirst-generation status and 85.1 percent for second-generation status) suggests that, above and beyond family/parental environment and expectations, the objectively high academic achievement of immigrant youth during high school is an important part of the reason for their higher than average rates of college attendance. The findings also suggest that the long-term effects of parental control and parental involvement during adolescence are indirect. Given that these factors are reduced to non-significance after the high school achievement variables are added to the model, they appear to influence college-going behavior only through their earlier effects on youths' college preparedness. While still significant, the magnitude of the association between parental expectations during adolescence and college attendance is also reduced with the addition of high school achievement variables. Finally, when the race/ethnicity and generation status interactions are included in Model 3b the lower likelihood of college attendance among Cuban immigrants is reduced to non-significance, indicating that Cuban firstand second-generation youth may be achieving at a lower than average level during their high school years. Although the magnitude of the advantages associated with both first-generation black status and second-generation Chinese status are reduced by controlling for high school achievement, the coefficients for these interactions continue to remain significant. Thus, while these youth may be benefiting from higher than average high school achievement, they remain advantaged over their white native-born peers by additional factors that we have not yet been able to examine. Multinomial Logistic Regression We also use multinomial logistic regression to predict the type of postsecondary educational experience reported by young adults who are currently attending college. We examine whether the respondent currently is not enrolled in college, is attending a two-year college or is attending a four-year college. Those who are not in college serve as the reference category. Table 5 presents the multinomial logistic results in relative risk ratios. Overall the results are very similar to those found in Table 4, suggesting that the processes influencing students' decisions to pursue post-secondary education generally apply to both two-year and four-year degree programs. Both first- and second-generation immigrants have significantly higher odds of attending two-year and four-year colleges (versus having no college attendance) than third-generation native-born youth (see Model 1).


Table 5: Multinomial Logistic Regression (Relative Risk Ratios & Z-Scores) Predicting College Type

Generation 1 Generation 2

Model 3 Model 2 Model 1 Two-year Four-year Two-year Four-year Two-year Four-year (vs. No Collège] (vs. No College) (vs. No College) 2.047* 1.607 2.070* 2.380" 2.387" 2.197" (1.62) (2.44) (2.50) (2.92) (2.92) (2.77) 1.540* 1.800" 1.428* 1.696" 1.580" 1.941" (2.03) (3.18) (3.91) (2.43) (2.69) (4.58)

Parental Control Parental Involvement Talked about School Work/Grades Worked on School Project Talked about Things in School Had an Argument about Behavior Parental College Expectations (ref. Very Disappointed) Somewhat Disappointed Not Disappointed Missing Students' College Expectations (ref. Not Very Likely) Very Likely AHPVT Scores (range 7-122) Highest Math for Credit (HS) (ref. Advanced) No Math Credit Below Average Average Math Continued on the following page.

1.077 (.38)

.666* (-2.37)

1.403 (1.65)

1.243 (-1.07)

.919 (-.80) .973 (-.25) 1.017 (.21) .937 (-.96)

1.017 (.21) 1.097 (.87) 1.204* (2.30) .830** (-2.66)

.932 (-.63) .982 (-.16) .958 (-.50) 1.000 (.01)

1.039 (.36) 1.163 (1.34) 1.040 (.38) 1.054 (.64)

.983 (-.20) .871 (-1.18) .854 (-.85)

.742" (-4.08) .451" (-6.19) .787 (-1.60)

1.022 (.25) 1.039 (.31) .948 (-.28)

.921 (-1.04) .750 (-1.93) 1.159 (.85)

3.430** (15.54)

1.452" (4.11) 1.005 (1.33)

1.926" (7.15) 1.018" (3.96)

.511 (-1.55) .477" (-3.54) 1.049 (.41)

.226 (-1.88) .340** (4.06) .466** (-7.73)

1.763*' (6.53)

142 . Socía/Forces 87(1) • September 2008

Table 5 continued Model 1 Two-year Four-year (vs. No College) Highest Science for Credit (HS) (ref. Advanced) No Science

Model 2 Two-year Four-year (vs. No College)

Below Average Average Science Cumulative Overall GPA in HS Pseudo R-square -2PseudoLog Likelihood (N = 10,163)

12.36 18076.37

15.98 17329.35

Model 3 Two-year Four-year (vs. No College)

.622 .461 (-1.22) (-1.76) .467** .417* (-3.65) (-3.17) .870 .717* (-1.44) (-4.17) 1.617** 4.814* (5.99) (18.39) 28.01 14847.83

Note: All models control for: age, gender, race/ethnicity, family education, family income mothers work status, region of country, urbanicity, school type, family structure, number of co-resident siblings, and language spoken in the home. *p < .05 **p < .01

Parental behaviors and college expectations during adolescence are significantly associated with later college attendance, particularly at a four-year college, but do little to mediate generational differences in type of college attended (see Model 2). The individual-level high school achievement variables, however, reduce to non-significance the greater likelihood of first-generation youth to attend a four-year college (see Model 3). As in the previous analysis, the addition of these high school achievement measures also reduces to non-significance the effects of parental control, parental involvement and parental expectations on the likelihood of attendance at a four-year college. Even after controlling for these factors, however, first-generation immigrants remain more likely than their third-plus peers to attend a two-year college. In addition, the full model does not fully explain the second-generation immigrant advantage in college attendance for either a two-year or four-year institution. It is also interesting to note that when we change the reference category from "no college" to "two-year college" attendance, neither firstnor second-generation immigrants have significantly lower odds than their third-plus counterparts of attending a four-year institution as opposed to a two-year institution. Surprisingly, this finding holds even in the absence

College Attendance of Immigrant and Native Youth • 143

of controls for socio-dennographic characteristics (results available upon request). Thus, while first- and second-generation youth are more likely than native youth to attend either a two-year or a four-year institution than to not attend college, there are no significant generational differences in the type of institution attended annong those who do go to college. As with the logistic analyses, we explored interactions between race/ ethnicity and generation status. In contrast to the findings of those analyses, we find no conclusive pattern of significant interactions when predicting the type of college attended by youth. We speculate that small cell sizes in this analysis may lead to a lack of power necessary to find significance. Consequently, these results have not been included in Table 5 (results available upon request). Conclusion Using data from the Add Health survey, we sought to determine if immigrant youth who have been raised and educated in the United States experience net advantages or disadvantages in seeking post-secondary education after obtaining a high school diploma or GED. We specifically wished to better understand the extent to which post-secondary educational differences by generation status can be explained by socio-demographic characteristics, parental behaviors and college expectations during adolescence, and individual ability and achievement while in high school. This study is unique in that the large, nationally representative dataset employed permits the examination of smaller immigrant groups, allowing specific race/ethnic differences to be explored in detail. The Add Health also includes detailed information on the family environments of youth, as well as links to externally validated sources of information on academic achievement during high school (i.e., transcripts released from each individual high school). As a result, we are able to examine the relative importance of both parental behaviors/expectations and past academic performance in explaining generation differences in post-secondary outcomes. The school-based design of the study also ensures that all respondents in the sample had been enrolled in and graduated from a U.S. high school (or obtained a GED after having been enrolled in a U.S. high school). As such, all youth in the sample had been living in the United States for a minimum of seven years, were technically "qualified" to apply to and enter an institution of higher education, and had not likely come to the United States for the sole purpose of obtaining a higher education. The findings show that there are several important ways in which first-generation and second-generation youth vary from their third-plus generation counterparts. Confirming our first prediction, we find that even though immigrants disproportionately experience social, economic and

144 . Sod«/forces 87(1) . September 2008

political inequities in the United States, these youth are no less likely to attend college than their native-born peers. In fact, they are significantly more likely to do so than their native-born counterparts of similar sociodemographic characteristics. Furthermore, contrary to our expectations, we find no significant generation differences in the type of institution that college-bound youth attend (two-year vs, four-year). Thus, among youth who have successfully completed their high school education in the United States, first- and second-generation immigrants exhibit a clear post-secondary educational advantage that is not obtained through disproportionate enrollment in more affordable, but less prestigious and less financially rewarding two-year community colleges and associate's degree programs. Although parental behaviors and academic expectations have significant direct effects on the likelihood of college attendance among young adults, controlling for these factors does little to explain the immigrant advantage. Generational differences in parental behaviors and expectations have been shown to provide immigrant children with an advantage over native-born youth in terms of academic achievement during primary and secondary school (Glick and White 2004; Hao and Bonstead-Bruns 1998), Yet, immigrants' greater likelihood of making the transition into college appears to be more directly affected by that previous academic achievement than by the continuing influence of parental behaviors and expectations. Overall, immigrant youth are academically outperforming their thirdgeneration peers in high school, and controlling for this achievement fully "explains away" the first-generation, and partially explains the secondgeneration, advantage in college attendance. High school achievement measures also mediate the direct effects of parental behaviors during the teen years, indicating that parental behaviors largely affect postsecondary educational outcomes indirectly, via their influence on children's preparedness for college. Thus, as past research shows, immigrant parents' behaviors toward and expectations for their children may push immigrant youth to succeed throughout their secondary schooling, often at levels higher than their native-born peers of similar socio-demographic backgrounds (e,g. Portes and Zhou 1993; Kao 1999; Kao and Tienda 1995; Portes and Rumbaut 1996, 2001 ; Rumbaut 1997; Tiilman, Guo and Harris 2006), Upon high school graduation, immigrants who have been raised and educated in the United States are simply better prepared to take advantage of post-secondary educational opportunities than are their thirdgeneration peers. As a result, immigrant youth are entering college at levels higher than would be expected given their background characteristics. The fact that their higher than average levels of preparation make them more competitive college applicants may be particularly important for enrollment of first-generation youth because foreign-born non-citizens


may face greater limitations and more barriers to acquiring scholarships, federal grants, and other financial assistance for college expenses than do citizens. Further research into how immigrants navigate the hurdles associated with financing a college education is clearly warranted. Although generation status is significantly associated with the likelihood of college attendance, the multivariate results confirm our prediction that this association is conditioned by race/ethnicity. In fact, the interaction models provide evidence that immigrant generation is not a significant predictor of college attendance among youth of most racial/ethnic backgrounds. The beneficial effects of immigrant status are most pronounced among second-generation Chinese and first-generation black youth. This finding is not entirely surprising given that past research on immigrant schoolchildren has shown Asians to benefit more than other children from having foreign-born parents and blacks to benefit more from their own foreign birth (Kao and Tienda 1995). On the whole, while youth in some racial/ ethnic groups may benefit more than others from their immigrant status, it is important to note that first- and second-generation immigrant status are generally not negatively related to college attendance for any youth. These findings offer some general support for the theory of segmented assimilation and suggest that, despite having been raised and educated alongside U.S.-born peers, immigrant youth from various racial/ethnic backgrounds may experience divergent pathways of assimilation that result in different levels of educational attainment. While our ability to account for group differences in the social and community contexts of immigrant youth was limited by our data, our analyses did examine some important aspects of human and cultural capital that vary across specific immigrant groups, including family socioeconomic status, family composition, parenting behaviors and expectations for academic attainment. Yet, even after controlling for these measures of human and cultural capital, as well as the respondents' verbal ability and high school achievement, second-generation Chinese and first-generation black immigrants remain significantly more likely to attend college than their third-generation White peers. What other factors may be driving this heightened college enrollment? We believe that future analyses (and future data collection efforts) should more closely examine the social and community networks of these youth. Some research suggests that Asian immigrants are more likely than others to be enmeshed in ethnically homogenous communities and organizations, such as religious groups and after-school language/ heritage schools (Garcia Coll and Magnuson 1997; Portes and Zhou 1993). Participation in and access to these kinds of organizations may reinforce ethnic identities and cultural norms that place high value on educational attainment (above and beyond the messages received from parents alone), increase access to adults who can serve in a mentoring capacity, and

146 . Soda/Forces 87(1) . September 2008

increase access to informal social connections that may help youth to obtain information about higher education and to procure employment or otherforms of financial assistance. The Chinese, with higher than average socioeconomic status (Portes and Rumbaut 2001), may be among the most likely of Asian groups to have access to these kinds of useful social connections. Moreover, compared to their first-generation counterparts, the second-generation may be more able to translate social capital into increased access to college education because they do not face barriers associated with non-citizen status or language difficulties. Likewise, the college-going behavior of black immigrants may be affected by their location in ethnically homogenous social and community contexts. Many immigrants from the Caribbean, the location of origin for the majority of first- and second-generation blacks, feel a strong desire to maintain their unique ethnic heritage and to distance themselves culturally and socially from native-born blacks (Waters 1996). As with the Chinese, black immigrants as a group also tend to be selective of professionals and those with above average levels of education (Portes and Rumbaut 2001 ). For black immigrants, however the advantages conferred by engagement in ethnically homogenous communities that reinforce educational aspirations and access to social resources may erode quickly over generation. Unlike Asian and Hispanic immigrants, black immigrants are often assumed to be a part of the native-born population, particularly if they do not speak with a foreign accent or dress in a distinctively ethnic manner. As a result, secondgeneration blacks may face more pressure than other immigrant youth to take on the racial/ethnic identity of their U.S.-born counterparts (Portes and Rumbaut 2001; Waters 1996). In doing so, they may also more quickly distance themselves from the kinds of social and community networks that have the potential to enhance post-secondary educational outcomes. In order to test these suppositions, future data collection efforts need to gather more detailed information regarding the ethnic identification and social/community contexts of young immigrants. While the results of this study are compelling, it is important to note its limitations. First, we are only able to focus on youth who have graduated from a U.S. high school or who have obtained a GED in the United States. There are several important implications of this; our study omits individuals who have dropped out of school (without obtaining a GED) as well as immigrant youth who never enrolled. As a result this sample is a select group of youth who attended high school in the United States and who are "eligible" to attend college. Unfortunately, the Add Health data set also does not permit us to separate out those respondents who have an actual high school diploma from those who have a GED. We recognize that these two forms of high school education are different and consequently may be associated with different post-secondary educational outcomes. We are


also unable to explore the respondents' legal status. Non-citizen immigrants and children of illegal immigrants may experience more roadblocks in their quest for higher education, and so may be less likely to attend college than their naturalized and/or legal counterparts. Future research should better examine these various groups to see if they differ substantially. Overall, this research adds to the growing literature on immigrant educational outcomes. Past research has found that immigrant schoolchildren generally perform quite well when compared to their nativeborn counterparts. The findings presented here provide evidence that the immigrant advantage extends to the likelihood of attending college. However, this advantage may only apply to immigrants of specific racial/ ethnic groups, particularly Chinese and black immigrants. Immigrants of other racial/ethnic backgrounds appear to be less able to translate their childhood achievements into better than average educational outcomes during their early adult years. A better understanding of why some particular immigrant groups are disproportionately likely to enter the American postsecondary educational system will allow us to help all children in American schools achieve to their academic potential. Notes 1.

Add Health is a project designed by J. Richard Udry, Peter S. Bearman and Kathleen Mullan Harris, and funded by a Grant P01 HD31921 from the National Institute of Child Health and Human Development, with cooperative funding from 17 other agencies. Acknowledgement is due to Ronald R. Rindfuss and Barbara Entwisle for assistance in the original design. The Education Data component of the Add Health was also supported by a grant from the National Institute of Child Health and Human Development under grant ROÍ HD40428-02 to the Population Research Center, University of Texas at Austin; Chandra Müller (PI) and the National Science Foundation grant number REC-0126167, Chandra Müller (PI). Persons interested in obtaining data files from Add Health should contact Add Health, Carolina Population Center, 123 Franklin Street, Chapel Hill, NC 27516-2524, USA (www.cpc.unc.edu/addhealth/contract.html).

2. Although Goyette and Xie (1999) suggest that the college expectations of Asian immigrant youth are similar across ethnic subgroups, past research on the actual achievement of immigrant adolescents suggests that significant differences by racial/ethnic subgroup exist among both Asians and Hispanics (Feliciano and Rumbaut 2005; Portes and Rumbaut 2001). We expect that racial/ethnic subgroup might also condition the association between immigrant generation and the likelihood of attending college. 3. The data do not allow us to separate out respondents who have a GED from those who have an actual high school diploma.

148 • Social Forces 87(1) • September 2008

4.

Puerto Ricans are excluded from this study because individuals born in Puerto Rico hold U.S. citizenship, making their migration experience quite different from that of other immigrants. Removal of Puerto Ricans from the analytic sample does not significantly alter the main findings. Given the small number of immigrants from European nations and Canada, all immigrant youth from these national backgrounds who indicated that they were "White" or "Caucasian" are combined together under the racial/ethnic category of "White." Likewise, all immigrant youth from African and Caribbean backgrounds who indicated "Black" or "African American" as their race are combined together under the racial/ethnic category of "Black." The vast majority of black immigrants in the Add Health report a Caribbean background. While not ideal, this strategy of combining together respondents from a variety of national backgrounds has been commonly employed in the study of immigrant youth (e.g. Bailey and Weininger 2002; Glick and White 2003; Kao and Tienda 1995), and is only done here for the smallest of groups. Furthermore, all analyses that examine the interaction between generation status and race/ethnicity use the "third-generation, white" group as the reference category Doing so minimizes the potential for comparisons that confound nativity and race/ ethnicity (e.g. comparisons between first-generation black immigrants, who are largely from the Caribbean, and third-generation blacks, whose ancestors likely hailed fronn countries within Africa).

5. We also examined measures of school-related behavior problems during middle school and high school, including whether or not the respondent had ever faced suspension or expulsion from school. While immigrant youth report lower levels of suspension and expulsion than native youth, these behavioral factors are never significant in the muitivariate analyses. As such, they are not shown here.

References Alba, Richard D., Douglas Massey and Ruben G. Rumbaut. 1999. The Immigration Experience for Famines and Children .\Nash'\r\gtor\, D.C: American Sociological Association. Arbona, Consuelo, and Amaury Nora. 2007. "The Influence of Academic and Environmental Factors on Hispanic College Degree Attainment." Review of Higher Education 30(3): 247-69. Astone, Nan Marie, and Sara S. McLanahan. 1991. "Family-Structure, Parental Practices and High-School Completion." American Socioiogicai Review 56(3):309-20. Bailey, Thomas R., and Elliot B. Weininger 2002. "Performance, Graduation, and Transfer of Immigrants and Natives in City University of New York Community Colleges." Educationai Evaluation and Policy Analysis 24(4):359-77. Bohon, Stephanie A., Monica K. Johnson and Bridget K. Gorman. 2006. "College Aspirations and Expectations Among Latino Adolescents in the United States." Social Problems 53(2):207-25.


Bryant, Alyssa N. 2001. "Community College Students: Recent Findings and Trends." Community College Review 29(3):77-93, Camarota, Steven A, 2004. Economy Siowed, But Immigration Didn't - The Foreign-Born Popuiation, 2000-2004. Found at: www.cis,org/articles/2004/ back1204,html, Washington, DC: Center for Immigration Studies. Chántala, Kim, and Joyce W. Tabor. 1999. Strategies to Perform a Design-Based Analysis Using the Add Health Data, Found at: www.cpc.une,edu/projects/ addhealth/news.html. Charles, Camille Z,, Vincent J. Roscigno and Kimberly C. Torres. 2007. "Racial Inequality and College Attendance: The Mediating Role of Parental Investments." Social Science Research 36(1 ): 329-52. Crosnoe, Robert, and Lorena Lopez-Gonzalez. 2005. "Immigration from Mexico, School Composition, and Adolescent Fur\ci\or\'mg." Socioiogicai Perspectives 48(1): 1-24. Dougherty, Kevin J. 1994, The Contradictory Coiiege: The Contacting Origins, impacts, and Futures of the Community College. State University of New York Press. Feliciano, Cynthia, and Rubén G. Rumbaut. 2005. "Gendered Paths: Educational and Occupational Expectations and Outcomes among Adult Children of Immigrants." Ethnic and Raciai Studies 28(6): 1087-1118. Garcia Coll, Cynthia, and Katherine Magnuson, 1997. "The Psychological Experience of Immigration: A Developmental Perspective." Pp. 91-131, immigration and the Famiiy: Research and Poiicy on US immigrants A. Booth, A.C. Crouterand N. Landale, editors. Lawrence Eribaum. Glick, Jennifer E., and Michael J. White, 2003, "The Academic Trajectories of Immigrant Youths: Analysis Within and Across Cohorts." Demography 40(4):759-83. . 2004. "Post-secondary School Participation of Immigrant and Native Youth: The Role of Familial Resources and Educational Expectations." Sociai Science Research 33(2):272-99. Goyette, Kimberly, and Yu Xie. 1999, "Educational Expectations of Asian American Youths: Determinants and Ethnic Differences." Socioiogy of Education 72[]}:22-36. Hagy, Alison P, and J. Farley Ordovensky Staniec. 2002. "Immigrant Status, Race, and Institutional Choice in Higher Education." Economics of Education Review ' 21(4):381-92,

150 • Social Forces 87{l) • September 2008

Hao, Lingxin, and Melissa Bonstead-Bruns. 1998. "Parent-Child Differences in Educational Expectations and the Academic Achievement of Immigrant and Native Students." Socioiogy of Education 71(3):175-98. Harker, Kathryn 2001. "Immigrant Generation, Assimilation, and Adolescent Psychological Well-being." Sociai Forces 79(3):969-1004. Harris, Kathleen M. 1999. "The Health Status and Risk Behavior of Adolescents in Immigrant Families." Pp. 286-347. Chiidren of immigrants: Heaith, Adjustment and Pubiic Assistance. D.J. Hernandez, editor. National Academy of Sciences Press. Harris, Kathleen M., Francesca Florey, Joyce W. Tabor et al. 2003. The National Longitudinal Study of Adolescent Health: Research Design. Found at: www. cpc.unc.edu/projects/addhealth/design.html. Haveman, Robert, Barbara Wolfe, Brent Kreideret al. 1994. "Market Work, Wages, and Men's Health." Journai of Heaith Economics 13(2): 163-82. Hirschman, Charles. 2001. "The Educational Enrollment of Immigrant Youth: A Test of the Segmented-Assimilation Hypothesis." Demography 38(3):317-36. Jaeger, David A., and Marianne E. Page. 1996. "Degrees Matter: New Evidence on Sheepskin Effects in the Returns to Education." Review of Economics and Statistics 78(4):733-40. Kane, Thomas J., and Cecilia E. Rouse. 1995. "Labor-Market Returns to 2-Year and 4-Year College."/4mer/'ca/7 Economic Review 85(3):600-14. Kao, Grace 1995. "Asian-Americans as Model Minorities - a Look at Their AcademicPerformance." American Journai of Education 103(2): 121 -59. . 1999. "Psychological Well-being and Educational Achievement among Immigrant Youth." Pp. 410-77. Chiidren of immigrants: Heaith, Adjustment, and Puhiic Assistance. D.J. Hernandez, editor National Academy Press. Kao, Grace, and Jennifer S. Thompson. 2003. "Racial and Ethnic Stratification in Educational Achievement and Attainment." Annuai Review of Socioiogy 29:417-42. Kao, Grace, and Marta Tienda .1995. "Optimism and Achievement-the Educational Performance of Immigrant Youth." Soc/a/Sc/ence üíya/íeA/)/76(1):1-19. McLanahan, Sara S., and Gary Sandefur. 1994. Growing Up with a Singie Parent: What Hurts, WhatHeips. Harvard University Press. Nora, Amaury, Laura I. Rendón and G. Cuadraz. 1999. "Access, Choice, and Outcomes: A Profile of Hispanic Students in Higher Education." Education of Hispanics in the United States: Poiitics, Poiicies, and Outcomes. A. Tashakkori and S.H. Ochoa, editors. AMS Press.


Portes, Alejandro, and Rubén G. Rumbaut. 1996. immigrant America: A Portrait. University of California Press. , 2001. Legacies: Tiie Story oftiie immigrant Second Generation. University of California Press. Portes, Alejandro, and Min Zhou. 1993. "The New Second-Generation: Segmented Assimilation and Its Variants." Annais oftiie American Academy of Poiiticai and Sociai Science 530(November):74-96. Reynolds, John, Michael Stewart, Ryan MacDonald et al. 2006. "Have Adolescents Become Too Ambitious? High School Seniors' Educational and Occupational Plans, 1976 to 2000." Sociai Probiems 53(2): 186-206. Rumbaut, Rubén G. 1997. "Assimilation and Its Discontents: Between Rhetoric and Reality." internationai Migration Review 31(4):923-60. Santibanez, Lucrecia, Gabriella Gonzalez, Peter A. Morrison et al. 2007. "Methods for Gauging the Target Populations that Community Colleges Serve." Popuiation Researcii and Poiicy Review 26(1 ):51-67. Schmid, Carol L. 2001. "Educational Achievement, Language-Minority Students, and the New Second Génération." Socioiogy of Education 74(Extra lssue):71-87. Schmidley, Dianne A. 2000. "Profile of the Foreign-born Population in the United States: 2000, Series P23-206." Current Popuiation Reports, 2000. Washington, DC: U.S. Census Bureau. Song, Chunyan Y, and Jennifer E. Glick. 2004. "College Attendance and Choice of College Majors among Asian-American Students." Sociai Science Quarteriy 85(5):1401-21. Suarez-Orozco, Carola, and Marcelo M. Suarez-Orozco. 2001. Ctiiidren of immigrants. Harvard University Press. Tillman, Kathryn H., Guang Guo and Kathleen M. Harris. 2006. "Grade Retention among Immigrant Children." Sociai Science Researcii 35(1 ): 129-56. Vernez, Georges, and Allan Abrahamse. 1996. i-iow immigrants Fare in U.S. Education. RAND Corporation. Warner, W. Lloyd, and Leo Srole. 1945. Tiie Sociai Systems of American Etiinic Groups. Yale University Press. Waters, Mary C. 1996. "Ethnic and Racial Identities of Second-Generation Black Immigrants in New York City." Pp. 171-96. Tiie New Second Generation. A. Portes, editor. Russell Sage. White, Michael J., and Jennifer E. Glick. 2000. "Generation Status, Social Capital, and the Routes Out of High School." Socioiogicai Forum 15(4):671-91.

152 . Social Forces 87(1) • September 2008

Zhou, Min, and Carl L. Bankston .1994. "Social Capital and the Adaptation of the 2nd Generation - the Case of Vietnamese Youth in New-Orleans." International Migration Review 28(4):821-45.