Educational Research and Review Vol. 1 (8), pp.

286-318, November 2006

Available online at http://www.academicjournals.org/ERR
© 2006 Academic Journals

Full Length Research Paper

Academic Achievements of Children in Immigrant

Families
Wen-Jui Han
Columbia University School of Social Work, 1255 Amsterdam Avenue, New York, NY 10027. Email:
wh41@columbia.edu.
Accepted 27 October, 2006

Utilizing data on approximately 16,000 children from the Early Childhood Longitudinal Survey-Kindergarten
Cohort and a rich set of mediating factors on 16 immigrant groups, this paper examined the associations
between children’s immigrant generation status and their academic performance. The changes in academic
achievements during kindergarten and first-grade were also examined to explore the varying learning paces
exhibited by children from different countries of origin. Results indicate that, compared to third and later
generation non-Hispanic white children, children of Latin American regions tended to have lower reading
and math scores, while children of Asian regions tended to have higher reading and math scores. In
addition, although children of immigrants may have either higher (e.g., children from East Asia) or lower
scores (e.g., children from Mexico) by first-grade compared to third and later generation non-Hispanic white
children, the former generally learned skills at faster paces, thus widening (e.g., for children from East Asia)
or narrowing (e.g., for children from Mexico) academic achievement gaps. Child and family characteristics
accounted for a large share of the differences in children’s academic achievements. Home, school, and
neighborhood environments may also matter but to a lesser extent. Research implications are discussed.

Keywords: academic achievements; immigrants; immigrant generation status; neighborhood characteristics; school
environments.

INTRODUCTION

Academic Achievements of Children in Immigrant importance of early childhood experiences to later

Families cognitive and social development (for review see
Shonkoff and Phillips, 2000), there is a noticeable void in
The United States is a nation shaped by immigration. In research on preschool and school-aged children of
the 1930s, the 14.2 million foreign-born individuals had immigrants (Board on Children and Families, 1995;
migrated mainly from Northern or Western Europe and Booth, Crouter, and Landale, 1997; Nord and Griffin,
made up 12 percent of the total population, while in 2003 1999), as well as a lack of longitudinal research to help
the 33.5 million foreign-born individuals had migrated us understand a variety of time-dependent aspects of
mainly from Latin America or Asia and represented their development.
11.7% of the total population. Now nearly 17 percent of This paper examines the developmental experiences of
children under age 18, or 11.5 million children, are living young children of immigrants in the context of several
with a foreign-born householder, and the percentage is individual, family, home environment, and school and
almost double for children under 6 years old (U.S. neighborhood characteristics that theories and empirical
Census Bureau, 2004). The unique cultural traditions of studies have suggested are important to children’s
the new immigrant groups present challenges to development. Specifically, using a longitudinal dataset
understanding their children's developmental trajectories. with a large, contemporary sample of children from the
Despite a large body of research demonstrating the Early Childhood Longitudinal Survey-Kindergarten Cohort
 Wei-Jui Han 287

(ECLS-K), the academic achievements of native-born (e.g., residential, economic, and social and psychological
(i.e., third and later generations) and foreign-born (i.e., segregation) are considered important components of
first- or second-generation) children entering kindergarten school and neighborhood's impact on learning
in the fall of 1998 are examined. environments. Borrowing from all of this research, child
This approach allows us to explore the likely development in immigrant families is hypothesized to be
mechanisms by which immigrant generation status related to (at least) 1) family background, 2) parental
(hereafter, generation status) may be associated with expectations, aspirations, and educational practices, and
child development. 3) school and neighborhood resources (Chao, 2001;
Conchas, 2001; García Coll, et al., 1996; Fuligni, 1997;
Fuligni, Tseng and Lam, 1999; Kao and Tienda, 1995;
Child Development Theoretical Framework Louie, 2001; Rumbaut, 1994, 1995; Suárez-Orozco and
Suárez-Orozco, 2001).
Ecological models developed by Bronfenbrenner (1979, In regard to the first hypothesis, theory and previous
1986) have substantially benefited the child development empirical evidence suggest that family socioeconomic
field over the past 30 years. Specifically, this model background may partially explain the academic success
emphasizes that the family's interaction with other groups of many European and Asian immigrants and the
and institutions will influence how children adapt to non- academic struggles experienced by many Latin American
familial environments (e.g. school), and has identified a immigrants. This is most likely linked to the fact that,
variety of risk and protective factors for children’s compared to the native-born population, European and
optimum development, such as child, parent, family, and Asian immigrants have similar or even higher parental
environmental characteristics (for reviews, Belsky, 2001; educational achievement and household incomes, while
Bornstein et al., 2001; Lamb, 1998; Johnson, et al., 2003; Latin American immigrants tend to have lower levels of
1
Shonkoff and Phillips, 2000; Weinraub and Jaeger, both (U.S. Census Bureau, 2004). However, even when
1990). Protective and/or risk factors attributable to the studies have controlled for family socioeconomic status, a
children themselves may involve age, gender, health, or significant association between generation status and
temperament; factors attributable to parents may involve academic achievement persists (Fuligni, 1997; Kao and
demographic characteristics (e.g., age, education, marital Tienda, 1995; Rumbaut, 1997). This suggests that family
status, employment) and the quality of the parent-child socioeconomic status alone would not be sufficient to
relationship (e.g., maternal depression, home explain the variations in academic achievements between
environment); and factors attributable to the family and foreign-born and native-born children.
the external environment may involve resources available Ethnographic and qualitative studies help explain such
inside or outside the home (e.g., family income, the variations. There is some evidence to show that children
presence of two parents, and the type and quality of early from Central America, Vietnam, India, and East Asia may
child care). be raised in family environments that strongly support
While Bronfenbrenner’s theory is generally valuable in academic achievement (Caplan et al., 1991; Chao, 1994,
understanding child development, issues important to 2001; Gibson, 1991; Gibson and Bhachu, 1991; Fuligni,
children’s development in immigrant families such as 1997; Louie, 2001). For example, personal accounts from
culture (Ogbu, 1978, 1981, 1988), discrimination, racism, a recent study describe a Latin American father who sat
and segregation are more fully addressed by the with his children while they were doing homework despite
integrative model developed by García Coll and her not understanding the material, which conveyed his
colleagues (1996, 2004). Drawing upon social dedication to education to his children and helped shape
stratification and ecological theory, this model assumes their commitment to academic performance (Pérez
that, in addition to children’s (e.g., age, temperament, Carre n, Drake and Barton, 2005). Serious attitudes such
biological factors) and families’ (e.g., structure and roles, as this are a manifestation of high academic expectations
values and goals) characteristics, children’s daily and aspirations for their children, and significantly
experiences and surrounding environments contribute to influence adolescents’ own attitudes and behavior.
their behavioral, emotional, and cognitive development Consistent with this, previous studies have shown the
and are closely tied to a social position significantly great effort and time devoted by adolescent children of
influenced by discriminatory and oppressive forces. The immigrants to doing homework with the desire to achieve
model further assumes that neighborhood and school
environments are in turn affected to either promote or 1
For example, in 2004, for the population aged 25 and over, the
inhibit the development of minority children and families. percentages of foreign-born immigrants from Europe and Asia that held
a bachelor’s degree or above were 36% and 50%, respectively,
Social position (e.g., race/ethnicity, social class, and compared to 26% of the native-born population. Only about 11% of the
gender), racism (e.g., prejudice, discrimination, foreign-born population from Latin America had achieved the same
institutionalized or symbolic oppression), and segregation education (with only 4% of immigrants from Mexico having achieved
such education) (U.S. Census Bureau, 2004).
288 Educ. Res. Rev.

academic success (Caplan et al., 1991; Gibson, 1991; communication to parents about children’s learning
Gibson and Bhachu, 1991; Fuligni, 1997; Louie, 2001; process and curriculum, teacher’s efforts in helping
Rosenthal and Feldman, 1991). Thus, the second students’ learning process), parental involvement, and
hypothesis incorporates the family’s values, beliefs, and school safety. These attributes have been identified
goals to account for their intergenerational transfers to largely through their associations with student
their children. Although previous studies have found achievement test scores. In addition, studies have shown
many factors related to home environment and parental that a safe and orderly school environment is linked to
educational practices, 9 variables seem to be the most the affirmation of healthy social behavior that is
important (Smolensky and Gootman, 2003): maternal characteristic of resilient children (Lee, Winfield and
depression, family routines, and the parents’ educational Wilson, 1991; Masten, 1994).
expectations, the importance they place on having skills Regarding the influence of neighborhoods, it is known
before attending kindergarten, their participation in school that the majority of new immigrants to the U.S. settle and
events, the difficulty they face in attending school events, live in inner-city areas, where the urban problems of
the learning materials they provide at home, provision of poverty, unemployment, crime, and social disorganization
extracurricular activities that may promote academic have historically been most intense (Sampson and
performance and/or physical/artistic skills, and use of Groves, 1989; Wilson, 1987) and which exacerbate the
physical discipline. For example, previous studies negative effects of the low socioeconomic status
indicate that children benefit more cognitively if they have observed in some immigrant families (e.g, Latin
less depressed mothers (NICHD early Child Care American) (Pessar, 1995; Portes and MacLeod, 1999).
Research Network [NICHD ECCRN], 1999; Peterson and Research has consistently found associations between
Albers, 2001), a high quality home environment (enriched stressful environmental conditions, such as poverty or
by the availability of and frequent interaction with books) unemployment, and negative parental psychological
(Bradley, 1995; Bradley et al., 1989), or attend center- functioning and parenting behavior, all of which adversely
based care (NICHD ECCRN, 1999, 2000, 2002a). affect child cognitive and socio-emotional development
Children benefit more socioemotionally if they participate (Conger et al., 1992; Elder et al., 1992; McLoyd, 1990;
in well-organized, positive extracurricular activities (such McLoyd and Wilson, 1991).
as sports, lessons, and clubs) (Mahoney, 2000; McNeal, Taken together, developmental theories and the
1995; Moore and Halle, 1997). integrative model put forward by García Coll and her
A third hypothesis concerns the impact of school and colleagues (1996, 2004) identify a rich set of factors
neighborhood resources. Previous studies have shown related to children's learning experiences and possible
that schools serving primarily children of color or living in links between generation status and child development.
poverty, for example, are likely to have fewer resources, All of these theoretical perspectives emphasize the
weaker academic focus, lower teacher expectations, and importance of examining child development in an
constricted curriculum (Griffith, 2000; Matute-Bianchi, ecological context, given that children’s learning is
1986; Ogbu, 1991; Ogbu and Simons, 1998; Valencia, heavily influenced by culturally guided family practices
2000; Valenzuela, 1999), which may adversely affect and interactions. At the same time, children’s surrounding
children’s learning experiences and academic environments (e.g., relatives, neighborhood, and ethnic
performance (Masten, 1994) and is essentially a form of community) shape their daily learning experiences.
segregation affecting children’s learning (García Coll, et However, given that previous research on child
al., 1996, 2004). Previous studies have also shown that development has mainly focused on middle-class white
differential treatment of students by race or ethnicity – children and research on immigrants has mainly focused
such as viewing Mexican children as less industrious on adolescents, we do not know whether the conclusions
than Asian American children – has hindered the from previous studies apply to young children from
achievement of some groups of children (Conchas, 2001; different cultural backgrounds (Hernandez, 1999; Siantz,
Moody, 2001; Suárez-Orozco and Suárez-Orozco, 1995, 1997).
2001). A large body of educational literature has Taking advantage of the large-scale, longitudinally-
identified factors important to the promotion of children’s designed ECLS-K data set, this study carefully
learning (Bernard, 1991; Borman and Overman, 2004; categorizes immigrant groups based on their country of
Crosnoe, 2005; Griffith, 2000, 2003; Herdenson and origin, reasons for migrating to the US, and cultural
Milstein, 1996; Huff and Trump, 1996; Lee and Burkham, background to examine whether generation status is
2002; McNeal, 1997; Moody, 2001). Among them, 7 associated with children’s academic achievements.
factors that may tap contextual (dis)advantages are Additionally, child/parent/family characteristics, home
teachers’ and school administrators’ qualifications, school environment and parental educational practices (e.g.,
student composition (e.g., minority representation), learning activities at home, participation in extracurricular
students’ academic performance, school’s efforts in activities and school events), and school (e.g., student
providing an optimal learning experience (e.g., school’s composition and average academic performance, parent-
 Wei-Jui Han 289

tal involvement, school safety) and neighborhood (e.g., trators on parental involvement and classroom and
residential neighborhood quality) environments are school characteristics; and from observational ratings of
considered possible mediating factors for any such school environments by study supervisors. More
associations. Three hypotheses are derived from the information on the ECLS-K can be found in the NCES
above research and models. First, if the child and family (2002) codebooks, in research reports published by
characteristics are important to the links between Denton and West (2002) and Lee and Burkham (2002),
generation status and children’s academic achievements, and in a research article by Magnuson et al (2004).
then we should see a reduction in the magnitude of the The study sample consists of approximately 16,000
estimate of generation status after controlling for child children for whom information was available on country of
and family characteristics (i.e., child and family origin, immigrant status, and at least one outcome
characteristics may mediate the association between variable at the spring of first-grade. Over 90% of the
generation status and children’s academic 4,000 excluded cases were not used because of missing
achievements). Second, if home environment is crucial to generation status or country of origin data. The raw data
the links between generation status and children’s suggest that the children with missing information tended
academic achievements, then we should see a reduction to be shorter and lighter, have mothers who are younger,
in the generation status magnitude after controlling for less educated, and less likely to be married at child’s
home environment. Finally, if school and neighborhood birth, and to have lower family socioeconomic status and
environments are critical to the links between generation move more frequently. The regression estimates may be
status and children’s academic achievements, then we thus biased downward due to these attributes.
should see a reduction in the generation status
magnitude after controlling for the school and
neighborhood backgrounds. MEASURES

Immigrant Generation Status and Country of Origin

DATA The parent respondent was asked in the spring of first grade to
report whether s/he was born in the U.S., and in the spring of
The ECLS-K, collected by the U.S. Department of kindergarten whether the child was born in the U.S.2 These two
Education's National Center for Educational Statistics, questions were used to identify a family’s immigrant status and
consists of a nationally representative cohort of 21,260 whether or not the child was a first- (child not born in U.S.) or
second-generation (child born in U.S. with at least one parent born
children who entered kindergarten in the fall of 1998 and outside of U.S.) immigrant. If the parent reported s/he or the child
who will be followed longitudinally until twelfth grade. was not born in the U.S., the parent was also asked to report the
These children were drawn randomly from a nationally country from which s/he came. A total of 16 regions were identified
representative sample of about 1,000 U.S. public and in this study based on country of origin, cultural background, and
private schools that offer kindergarten. In addition, the reasons for migrating to the U.S. (e.g., Vietnam, Thailand,
Cambodia, and Laos were categorized together primarily because
ECLS-K includes an over-sampling of Asian/Pacific
they are countries of refugee origin resulting from the Vietnam
Islander children, which allows for more detailed analyses War): North America (e.g., Canada), Europe (e.g., Denmark,
than other national data sets that lack sufficient numbers Greece, France, Hungary, including Russia), Puerto Rico (U.S.
of children of Asian origin. Given that slightly more than commonwealths such as Virgin Islands [n=20], Guam [n=3], and
1% of children in the ECLS-K did not complete a direct American Samoa [n=3] were not included due to small sample sizes
assessment of academic measures due to limited English and their different cultural backgrounds from Puerto Rico)3, the
Caribbean (e.g., Bahamas, Jamaica, Haiti; including mainly English-
proficiency, the study sample may not truly be nationally speaking or French-speaking countries), Central America (e.g.,
representative and may particularly affect the Belize, Costa Rica, El Salvador), South America (e.g., Argentina,
representation of children of Hispanic or Asian origin as Brazil, Colombia, Peru), Dominican Republic, Mexico, Cuba, East
detailed below. Asia (e.g., China, Japan, Korea),
The present study utilizes the data available as of this
writing, including the fall and spring of kindergarten and
the spring of first grade in which the full sample of 2
Because the interview only asked the nativity of one parent, it is likely
children were interviewed (a random sample of that not all children of immigrants would be identified in the ECLS-K
approximately 27% of the children were also interviewed (e.g., if we only had information on the mother for a native-born child
in the fall of first grade, mainly about their experiences with a native-born mother and a foreign-born father). Thus, estimates
presented here may be biased downward.
during the summer between kindergarten and first grade). 3
It is important to note that although children from Puerto Rico were
Direct assessments of children’s academic achievements also identified as first- or second-generation if they themselves or their
(i.e., reading and math skills) are examined, as well as parent(s) were not born in the U.S. mainland, these children are U.S.
information gathered from parents on family citizens. However, this paper acknowledges the importance of the
geographical and cultural differences between children from Puerto
characteristics and parental involvement in home learning Rico and those born in the U.S. and thus separates them in the
and school activities; from teachers and school adminis- analyses.
290 Educ. Res. Rev.

Table 1. Percentage Distribution of Country Origin by Immigrant Generational Status.

First Second
Generation Generation
North America (e.g., Canada) (n=46) 3.13 1.30
Europe (including Russia) (n=282) 13.78 9.34
Caribbean (e.g., Bahamas, Jamaica; including mainly English-speaking or 0.84 2.72
French-speaking countries) (n=91)
Puerto Rico (n=74) 3.13 2.55
Central America (e.g., Belize, Costa Rica, El Salvador) (n=165) 2.30 6.66
South America (e.g., Argentina, Brazil, Colombia, Peru) (n=168) 5.01 6.22
Dominican Republic (n=60) 1.46 3.33
Mexico (n=897) 25.68 33.46
Cuba (n=46) 1.25 1.73
East Asia (e.g., China, Japan, Korea) (n=212) 10.65 6.96
Vietnam/Thailand/Cambodia/Laos (n=147) 1.46 6.05
Other South East Asia (e.g., Indonesia, Malaysia, Philippines) (n=262) 7.31 9.81
India (n=109) 4.38 3.80
South-Central/West Asia (e.g., Armenia, Iraq) (n=78) 0.63 3.24
Africa (e.g., Ethiopia, Chad, Sudan, South Africa, Ghana) (n=46) 1.25 1.73
Oceania (Solomon Islands, Marshall Islands; excluding Australia) (n=110) 17.74 1.08
N 479 2313
(2.78%) (13.44%)

Vietnam/Thailand/Cambodia/Laos, other Southeast Asia (e.g., were more likely to be first-generation).

Indonesia, Malaysia, Philippines), India, South-Central/Western
Asia (e.g., Armenia, Iraq), Africa (e.g., Ethiopia, Chad, Sudan,
South Africa, Ghana)4, and Oceania (e.g., Solomon Islands, Academic Achievements
Marshall Islands; Australia was excluded due to its significant
cultural difference from other Oceania countries and because there Direct assessments of children’s competence in reading
was only 1 case from Australia available in the sample). Because (language and literacy) and mathematics were collected during the
previous studies have found that second-generation immigrant fall and spring of kindergarten and the spring of first grade via one-
adolescents generally perform better academically than their first- on-one testing sessions. These assessments were created
generation counterparts (e.g., Kao and Tienda, 1995), immigrant especially for the ECLS-K study with some items adapted from
generation status (2 generations) and country of origin (16 regions) existing instruments such as the Peabody Individual Achievement
were combined to create 32 dummy variables. Details on the Test-Revised and the Woodcock-Johnson Psycho-Educational
distribution of generation status by country of origin are provided in Battery-Revised. A brief language screening was administered to
Table 1. Approximately 16% of the ECLS-K sample is identified as 15% of children who were identified by teachers or school records
either a first- (3%) or second-generation (13%) child of immigrants. as having a non-English language background. Approximately 51%
About 40% of first-generation children originated from Latin of these children (7% of the overall sample) scored below the cut-
American regions (with more than half of those from Mexico), off point and received a reduced version of the assessments in
another quarter from Asian regions, and then followed by Oceania. order to be included in the analyses.5 Among low-scorers who did
Approximately 50% of second-generation children had parents who not complete the assessment and thus were not included in the
came originally from Latin American regions (again more than half analyses (n=317 in this sample), 75% were originally from Mexico,
from Mexico), followed by another third originating from Asian followed by another 5% who were third and later generation
regions. Hispanic. All in all, approximately 90% of the cases were of
For children of third and later generations (both child and parent Hispanic and 9% of Asian origin, and the raw data suggest that
born in the U.S.), race/ethnicity was identified with five groups: non- these children had different attributes from their counterparts (e.g.,
Hispanic white, non-Hispanic black, Hispanic, Asian, and other more children under 18 and more adults over 18 at home, poorer
(including multiracial). Table 2 provides the distribution of these and lower socioeconomic status, younger and less-educated
groups, with non-Hispanic white occupying more than half of the mothers, parents less likely to work full-time, child less likely to
total sample. It is important to note that a sample as young as this attend center-based care before kindergarten). Given these
is more likely to have second-generation children compared to different family backgrounds, it is possible the coefficients in the
samples used in previous studies of adolescent immigrants (who regression analyses might be underestimated for children of
Hispanic origin, and to some extent for children of Asian origin.
4 However, it is not clear whether the coefficients would be
It may be preferable to separate white and black immigrants from
underestimated after controlling for the three sets of mediators as
Africa due to differences in culture and societal treatment in their home
countries and the U.S. However, given that only 1 out of 6 first-
described above given children may respond differently (or have
generation African children was white (13 out of 40 for the second
generation), it would not be statistically possible to make such
5
distinctions. Nonetheless, the impacts of not making this separation are It should be noted that by taking reduced versions of the tests, the test
discussed below whenever possible. scores might not reflect the “true” ability of the child.
 Wei-Jui Han 291

Table 2. Selected Sample Characteristics and Mean Academic Skills by Immigrant Generational Status and Race/Ethnicity of Children in Third+ Generation

First Second Third+ Generation Full sample

Generation Generation (n=14411)
(n=479) (n=2313) Non-Hispanic Non-Hispanic Hispanic Asian Other
White Black (n=1590) (n=489) (n=655)
(n=9369) (n=2308)
Reading skills
Fall kindergarten 52.26 (10.49) 50.95 (11.43) 52.23 (9.54) 47.39 (9.12) 47.80 (9.91) 50.89 (9.61) 46.61 (10.15) 50.74 (9.99)
Spring kindergarten 50.98 (10.21) 51.19 (10.67) 52.16 (9.19) 47.14 (9.96) 49.01 (9.82) 50.76 (9.84) 47.98 (10.13) 50.84 (9.79)
Spring first grade 51.01 (9.24) 50.52 (9.68) 52.30 (8.85) 46.79 (10.39) 48.90 (9.45) 50.08 (9.61) 47.55 (10.67) 50.74 (9.56)
Math skills
Fall kindergarten 48.88 (10.21) 48.08 (10.82) 53.31 (9.34) 46.66 (8.72) 47.36 (9.56) 51.37 (9.61) 47.55 (10.01) 50.77 (9.97)
Spring kindergarten 49.25 (10.12) 48.71 (10.78) 53.23 (9.03) 46.12 (9.37) 48.01 (9.54) 50.64 (9.20) 48.35 (9.62) 50.84 (9.83)
Spring first grade 49.47 (9.63) 49.17 (9.80) 52.87 (8.79) 45.68 (9.93) 48.52 (9.04) 49.49 (9.04) 47.83 (9.62) 50.64 (9.55)
Child Characteristics
Boy (%) 48.85 50.97 51.41 50.30 51.32 52.35 51.15 51.14
Child age in months, fall 68.65 (4.74) 67.46 (4.23) 68.91 (4.41) 68.25 (4.44) 68.16 (4.38) 67.52 (4.49) 69.02 (4.90) 68.51 (4.45)
kindergarten
Low birth weight (<2500 g) (%) 18.37 10.64 8.70 15.34 11.07 8.38 9.47 10.36
Premature (>=2 weeks early) (%) 15.74 15.39 16.54 18.74 17.34 10.64 15.32 16.54
Height, fall kindergarten 44.46 (2.35) 44.35 (2.13) 44.76 (2.15) 45.09 (2.24) 44.27 (2.10) 43.81 (2.18) 44.83 (2.15) 44.67 (2.18)
Weight, fall kindergarten 45.30 (8.22) 46.43 (9.63) 46.25 (8.15) 47.77 (9.78) 46.51 (9.38) 43.55 (9.28) 47.16 (9.11) 46.43 (8.78)
Number of moves since birth 2.54 (1.21) 1.94 (1.08) 2.07 (1.34) 2.14 (1.28) 2.23 (1.34) 1.93 (1.06) 2.31 (1.45) 2.10 (1.30)
Center-based care before entering 39.76 37.17 49.91 34.50 34.22 29.66 31.08 43.37
kindergarten (%)
Parent Characteristics
Mother’s age 32.82 (6.40) 33.88 (6.07) 34.01 (5.87) 32.18 (8.53) 31.67 (6.77) 34.44 (7.52) 32.61 (7.51) 33.47 (6.56)
Parent’s education (whichever is
higher) (%)
Below high school (<12) 18.86 23.75 3.43 14.86 16.50 11.17 8.45 9.66
High school degree (=12) 18.86 23.30 22.55 37.64 31.39 26.75 27.54 25.62
Some college (>=13 and <15) 20.76 21.51 33.74 34.57 36.08 30.91 42.10 32.30
College and plus (>=15) 41.53 31.44 40.28 12.93 16.03 31.17 21.91 32.42
Mother married at birth (%) 84.38 75.72 83.94 29.70 60.10 76.08 48.12 72.56
Mother currently works full-time 33.58 41.35 43.72 59.12 47.46 47.65 44.04 45.54
(%)
Family Characteristics
292 Educ. Res. Rev.

Table 2. Continued

Number of persons age <18 in 2.40 (1.21) 2.54 (1.14) 2.38 (1.00) 2.66 (1.42) 2.50 (1.21) 2.75 (1.66) 2.77 (1.49) 2.47 (1.15)
the household
Social economic status (SES) 0.00 (0.92) -0.15 (0.88) 0.22 (0.73) -0.38 (0.75) -0.24 (0.70) -0.05 (0.75) -0.16 (0.84) 0.02 (0.80)
prestigious score
Home language is not English 60.46 60.49 0.71 0.75 27.69 44.14 2.29 14.27
(%)
Region of residence
Northeast 17.95 19.37 21.12 13.26 13.90 7.98 9.92 18.28
Mid-West 15.87 10.38 32.32 17.59 15.53 13.29 46.11 25.37
South 35.07 25.34 32.33 61.05 26.73 16.77 17.86 33.81
West 31.11 44.92 14.23 8.10 43.84 61.96 26.11 22.55
Location of residence
Rural 4.59 2.46 17.34 9.27 5.60 3.07 29.77 12.89
Small town 7.10 4.24 11.12 3.51 4.53 21.47 13.74 8.85
Large town 1.46 0.61 4.15 4.68 1.45 0.20 1.83 3.22
Mid-size suburban 5.85 4.06 9.00 5.55 4.21 4.70 5.80 7.10
Large-size suburban 30.27 32.43 29.41 23.09 28.93 27.40 16.95 28.41
Mid-size city 23.80 18.07 20.15 25.78 20.75 16.97 19.24 20.66
Large-size city 26.93 38.13 8.83 28.12 34.53 26.18 16.27 18.87
Percent of sample (%) 2.78 13.44 54.46 13.42 9.24 2.84 3.81 100

Note. Skills in Reading, Math, and General Knowledge are standardized scores with mean of 50 and standard deviation of 10. Standard deviations are in parentheses. See Appendix Table for
detailed definitions of sample characteristics.
 Wei-Jui Han 293

Table 3. Mean Academic Skills by Country of Origin/Immigrant Generational Status and Race/Ethnicity of Children in Third and Later
Generation.

Reading Skills Math Skills

Fall Spring Fall First- Fall Spring Fall First-
kindergarten kindergarten grade kindergarten kindergarten grade
First generation
North America 54.24 55.61 53.59 52.37 52.73 53.22
(11.46) (9.04) (7.84) (8.31) (9.62) (7.45)
Europe (including Russia) 52.30 51.58 54.05 51.49 50.87 52.98
(10.34) (10.10) (7.79) (8.60) (8.57) (7.41)
Caribbean 42.90 44.28 48.15 41.38 43.87 47.72
(3.05) (6.07) (6.43) (7.22) (11.28) (4.07)
Puerto Rico 46.72 42.23 44.09 44.74 43.10 41.57
(15.36) (14.75) (15.39) (12.97) (9.74) (12.88)
Central America 51.85 53.92 51.42 48.52 49.54 49.87
(4.82) (6.82) (8.21) (8.48) (7.12) (7.50)
South America 45.32 46.78 47.97 47.01 47.24 45.32
(9.86) (8.40) (9.41) (10.73) (7.38) (11.36)
Dominican Republic 45.68 48.18 44.95 42.85 42.74 44.22
(5.14) (7.84) (5.92) (6.80) (7.29) (6.79)
Mexico 47.95 45.34 45.17 42.95 43.75 46.29
(12.95) (9.72) (8.46) (9.12) (10.25) (8.98)
Cuba 42.67 46.37 46.07 45.21 49.12 48.25
(7.72) (6.65) (3.75) (5.93) (3.44) (7.91)
East Asia 55.09 55.67 56.42 55.48 56.11 54.68
(9.40) (7.83) (8.14) (9.08) (7.88) (10.62)
Vietnam/Thailand/Cambodia/Laos 56.02 53.15 54.99 59.39 59.34 54.21
(5.47) (10.65) (7.44) (9.71) (12.17) (9.15)
Other Southeast Asia 52.25 53.83 55.20 52.50 51.43 49.28
(10.55) (10.40) (7.98) (9.86) (9.24) (8.37)
India 59.22 55.92 55.69 55.02 54.79 52.14
(9.24) (10.48) (8.49) (9.23) (9.16) (8.69)
South-Central/Western Asia 50.60 52.15 37.91 40.84 43.50 41.11
(0.00) (0.50) (0.64) (13.88) (5.48) (7.26)
Africa 55.59 55.79 57.81 59.49 57.86 53.62
(4.54) (7.97) (6.96) (2.70) (8.70) (7.27)
Oceania (excluding Australia) 52.26 50.92 50.88 50.78 50.78 49.85
(9.26) (9.24) (7.17) (8.69) (9.39) (9.40)
Second generation
North America 52.00 52.84 52.02 54.03 52.17 50.88
(9.16) (10.59) (10.35) (7.07) (9.80) (10.49)
Europe (including Russia) 52.74 52.31 52.69 53.28 53.75 53.09
(10.01) (9.90) (9.69) (10.03) (10.14) (9.18)
Caribbean 51.40 51.56 50.31 47.89 47.66 46.33
(9.13) (9.65) (9.53) (8.10) (9.60) (9.36)
Puerto Rico 47.81 47.54 47.38 44.74 44.74 45.48
(9.26) (10.91) (9.58) (9.08) (12.16) (11.67)
Central America 48.01 49.67 48.63 44.60 47.02 46.69
(9.62) (9.79) (9.11) (9.43) (9.42) (9.27)
South America 51.54 52.22 51.80 51.17 51.44 50.76
(10.43) (9.22) (8.74) (9.33) (9.09) (9.42)
Dominican Republic 47.29 45.18 46.28 43.18 42.40 43.52
(8.63) (10.23) (10.69) (9.10) (10.44) (9.91)
Mexico 43.84 45.72 45.88 42.39 43.47 46.26
(9.23) (9.57) (8.73) (8.97) (9.66) (9.50)
Cuba 51.53 55.08 52.17 50.15 53.77 52.60
(10.45) (8.25) (8.68) (9.62) (8.53) (9.03)
294 Educ. Res. Rev.

Table 3. Cont’d.

East Asia 59.23 58.84 56.81 58.80 57.82 54.93

(11.58) (10.01) (7.64) (9.86) (8.26) 7.78)
Vietnam/Thailand/Cambodia/Laos 48.12 50.14 50.69 50.92 51.51 50.45
(10.64) (9.50) (9.61) (9.97) (9.62) (9.07)
Other Southeast Asia 53.18 54.57 54.26 53.00 52.62 51.04
(11.73) (10.12) (8.57) (9.67) (9.08) (8.75)
India 60.34 57.87 56.46 56.63 54.81 54.10
(11.34) (10.13) (8.46) (8.97) (10.58) (8.31)
South-Central/Western Asia 53.35 52.93 54.14 51.69 50.89 51.98
(13.59) (11.68) (10.00) (12.00) (10.72) (10.02)
Africa 53.03 54.04 53.13 51.25 51.74 51.44
(9.64) (7.07) (5.93) (8.51) (8.12) (7.75)
Oceania (excluding Australia) 49.57 48.66 49.90 50.75 48.04 46.44
(13.19) (11.64) (10.52) (9.69) (10.44) (9.21)
Third and later generation
Non-Hispanic white 52.23 52.16 52.30 53.31 53.23 52.87
(9.54) (9.19) (8.85) (9.34) (9.03) (8.79)
Non-Hispanic black 47.39 47.14 46.79 46.66 46.12 45.68
(9.12) (9.96) (10.39) (8.72) (9.37) (9.93)
Hispanic 47.80 49.01 48.90 47.36 48.01 48.52
(9.91) (9.82) (9.45) (9.56) (9.54) (9.04)
Asian 50.89 50.76 50.08 51.37 50.64 49.49
(9.61) (9.84) (9.61) (9.61) (9.20) (9.04)
Other 46.61 47.98 47.55 47.55 48.35 47.83
(10.15) (10.13) (10.67) (10.01) (9.62) (9.62)
Note. Outcome measures of Reading and Math are standardized scores with mean of 50 and standard deviation of 10. Standard deviations
are in parentheses. See Appendix Table 1 for detailed definitions of sample characteristics.

different resilience) to various environments. with sounds at the end of words, (4) recognizing common words by
Regression results of teacher reported data are not presented sight, and (5) reading words in context.
but were similar to those presented here. It is worth noting that The mathematics test consisted of 64 items measuring skills in
because children who did not complete the assessment and thus conceptual and procedural knowledge and problem solving and
were excluded may have valid teacher-reported data, it may then were grouped into five proficiency levels: (1) identifying some one-
be possible to use teacher’s assessments to evaluate the possibility digit numbers, recognizing geometric shapes, and counting up to
of “biased” coefficients estimated from direct assessments. The raw ten objects by ones; (2) reading all one-digit numerals, counting
data suggest that children who were not included in the direct beyond ten, recognizing a sequence of patterns, and using
assessment analyses had significantly lower scores on teacher- nonstandard units of length to compare objects; (3) reading two-
reported reading and math outcomes compared to their digit numerals, recognizing the next number in a sequence,
counterparts. However, similar estimates were obtained for children identifying the ordinal position of an object, and solving a simple
of Hispanic and Asian origin, indicating that the direct assessment word problem; (4) solving simple addition and subtraction problems;
results may not be seriously biased or underestimated due to the and (5) solving simple multiplication and division problems and
exclusion of children with limited English proficiency. recognizing more complex number patterns.
The standardized T-scores (with a full sample mean of 50 and
standard deviation of 10) developed by the ECLS-K were used in
this analysis. Thus, the scores represent children’s abilities relative Mediating Factors
to their peers, and the change in mean T-scores over time would
reflect a change in the child’s abilities relative to their peers. Test To test the three hypotheses described above, information collected
reliabilities were high -- between .92 and .95 for all assessment from parents, teachers, and school administrators as well as a
points for reading and math. Average reading and math outcomes facility checklist completed by the study’s field supervisors were
at each assessment point are reported in Table 2 by children’s included in the analyses (variables described and detailed in
immigrant generation status and additionally by detailed country of Appendix Table 1). Selected characteristics are provided in Table 2
origin in Table 3. for the full sample as well as separately by children’s generation
The language and literacy (reading) assessment contained 72 status and racial/ethnic groups (for children of third and later
questions designed to measure basic skills (letter and word generations). To allow children with missing values to be included
recognition), receptive vocabulary, and comprehension (listening in the analyses, a set of dummy variables was constructed for
and words in context). It covered five proficiency levels: (1) covariates with missing variables (1=missing; 0=not missing), and
identifying upper and lower case letters by name, (2) associating
letters with sounds at the beginning of words, (3) associating letters
 Wei-Jui Han 295

the missing values were replaced with a value of zero.6 Rates of children) to children’s learning experiences. Controls for school
missing data were less than 1% for the demographic, family, and characteristics were proxied by the school’s student minority
home environment characteristics measured in the fall and 3% for composition and average student performance compared to the
the spring of kindergarten. Rates of missing data were higher for national mean. Information collected from school administrators
school characteristics, but generally below 20%. included the number of years served as principal, school
neighborhood quality, and school safety rated by the field
supervisor. School neighborhood quality is a composite variable
METHODS pertaining to safety, drugs, gangs, and tension stemming from
racial/ethnic/religious differences. These variables are detailed in
Ordinary Least Square (OLS) regression was used to estimate the Appendix Table 1.
associations between generation status and children’s academic In addition to assessing the impact of all these mediators, it may
achievements while controlling for an extensive set of child, parent, be equally important to understand the learning paces of children of
family, and school and neighborhood characteristics. Because immigrants, how they might shape long term developmental
schools were the primary sampling unit in the survey, the Huber- trajectories, and the extent to which changes in academic
White method was used to correct for standard errors in all achievements over time are associated with the mediating factors.
analyses. To test each set of hypotheses, the characteristics of the Examining these learning paces may shed light on how certain
child and family, home environment (including maternal depression, groups of children lag behind or catch up to each other. In addition,
parent-child relationships, and parental educational practices and examining the influence of each set of mediating factors may
expectations), and school and neighborhood environments were provide insight into different responses to the same environment.
added increasingly to the regression models. Additionally, these For example, previous studies have suggested that disadvantaged
factors may help to avoid some potential bias and more fully explain populations (e.g., youth from racial/ethnic minority groups and poor
the relationships under study. youth) may benefit to a greater extent from positive school
The first model includes only generation status by country of characteristics compared to their counterparts (Bryk, Lee, and
origin and race/ethnicity (for third and later generations) variables, Holland, 1993; Johnson et al., 2001) because these students may
without any other covariates. Thus, the coefficients represent the be more reactive to school contexts. If so, assessing the changes in
mean differences between children who were third and later academic achievements over time would reveal a particular school
generation non-Hispanic white and those who were not. The context’s initial and later effects on student performance.7 To
second model adds controls for child, parental, and family answer these questions, a residual-change model to relate changes
characteristics including: child’s gender, age in months, birth in children’s academic achievements over time is utilized (e.g.,
weight, premature status, height and weight, number of moves school following the same procedure described above (models 1 to
since birth, and attending center-based care before kindergarten; 4) with each model also controlling for the cognitive skills children
mother’s age, marital status at child’s birth, current employment safety) had already acquired by the end of kindergarten.
status, and parental education; and number of persons under 18 in The full set of OLS coefficients is presented in Appendix Tables 2
the household, family socioeconomic status (SES), home language and 3. Consistent with previous literature on immigrant families,
(a dummy variable of English or not), and location and region of non-Hispanic white children (US native born third and later
residence. generation) were used as the reference group.8 Table 4 evaluates
The third model is the same as Model 2, but adds controls for the the extent to which each set of factors may explain the association
home environment information that was collected from parental between generation status and children’s academic achievements,
surveys in the fall or spring of kindergarten. The home environment and Table 5 evaluates the extent to which each set of factors may
is proxied by considering parental emotional well-being (i.e., explain the association between generation status and children’s
depression) and educational practices (i.e., educational progress over time (the residual-change model). In addition,
expectations, importance of having skills before entering whenever appropriate, effect sizes, d (i.e., coefficients divided by
kindergarten, participating in and difficulty attending school events), standard deviation – standard deviation units with a mean of 0 and
and the general home environment (i.e., the home learning pooled standard deviation of 1) are reported to reveal what is
materials and activities, attending extracurricular activities, important beyond the rather arbitrary standards of statistical
frequency of spanking, and family routines) (see Appendix Table 1). significance (which are influenced by sample sizes). A commonly
The fourth model is the same as Model 3, but adds controls for the used set of standards based on Cohen (1977) is that an effect size
quality of school and neighborhood environments. Neighborhood of .20 is “small,” .50 “moderate,” and .80 “large.”
quality is a composite score derived from parental reports on the
prevalence of drugs, crimes, and abandoned buildings in the
family’s residential neighborhood. School environment was
measured by surveying teachers in the fall and spring of
7
kindergarten and school administrators in the spring of It is worth noting that gain scores are typically negatively correlated
kindergarten. Covariates were included for teacher surveys to with initial status (so the children who started the lowest may artificially
account for the efforts devoted by parents (parental involvement in “gain” the most). Thus, factors that are positively associated with
school activities) and teachers (i.e., communicating to the parent children’s initial scores may be negatively associated children’s change
about the child; efforts to ease transition into kindergarten for scores. For example, attending center-based care before kindergarten
may be positively associated with children’s initial scores, but may be
negatively associated with children’s progress over time and thus would
6
There are several options as to how to handle missing values, lead to a narrowed achievement gap.
8
including dropping cases with missing values, keeping those cases but Previous literature on immigrant families has also questioned the
indicating that the information is missing by use of a dummy variable, appropriateness of comparing immigrant families with a group that
and imputing the missing values. The second option was chosen possesses markedly different cultural and historical backgrounds (e.g.,
because it is both the most conservative and widely agreed upon. In non-Hispanic whites). Unfortunately, because the data set does not
future work, it would be of interest to explore imputing missing values on provide country of origin data for third and later generation children, it
all of the key variables used in the analysis; however, this is beyond the would not be possible to categorize them into the 16 immigrant groups
scope of the current paper. as detailed in the measures section.
296 Educ. Res. Rev.

RESULTS from South-Central/Western Asia and Oceania and thus

were not presented here.
Table 3 presents the descriptive statistics for the average
kindergarten and first grade academic achievements by Latin American children: Looking at the figures from
generation status by country of origin. Two general trends Models 1 to 4 in Panel A of Table 4, results consistently
are revealed. The first is that although we tend to see show that Latin American children had significantly lower
lower scores in academic achievements for the first- and reading and math scores compared to non-Hispanic
second-generation children in Table 2, there are white children. In addition, first-generation children
tremendous differences within these groups by country of tended to have lower reading and math scores compared
origin. Compared to third and later generation non- to their second-generation counterparts (except for first-
Hispanic white children (hereafter, non-Hispanic white generation children from Central America on math
children), first-generation children from East Asia, scores). Specifically, compared to non-Hispanic white
Vietnam/Thailand/Cambodia/Laos, India, and Africa, as children, the effect sizes for reading scores ranged from
well as second-generation children from East Asia, other 0.4 (for first-generation children from South America and
Southeast Asia, and India tended to have higher scores second-generation from Central America) to 0.8 (for first-
on reading and math at all assessment points, while both generation from Puerto Rico and from the Dominican
first- and second-generation children from Latin American Republic); and the effect sizes for math scores ranged
regions tended to have relatively lower scores in reading from 0.2 (for second-generation from South America) to
and math at all assessment points. Second, generally 0.9 (for first-generation from Puerto Rico and first- and
speaking, second-generation children whose parents second-generation from the Dominican Republic). These
originally came from Latin-American regions (except significant and large differences were accounted for
Central America) as well as Asian regions (except either wholly (such as with the reading scores of second-
Vietnam/Thailand/Cambodia/Laos) tended to be doing generation children from Puerto Rico and Central
better academically compared to their first-generation America, and with the math scores of first-generation
counterparts. In addition, while the first- and second- children from Mexico) or in large part (about 75%, such
generation children from Latin American regions tended as with the reading scores of first-generation children
to have relatively lower scores in reading and math from the Dominican Republic and Mexico, and with the
compared to third and later generation Hispanic children math scores of second-generation children from Central
(hereafter, Hispanic children), the first- and second- America) by child and family characteristics. In addition,
generation children from Asian regions tended to have first-generation children from Cuba (d = 0.6 on reading)
relatively higher scores in reading and math compared to and South America (d = 0.6 on math) performed
third and later generation Asian children (hereafter, Asian significantly worse than their second-generation
children). counterparts. These significant differences persisted
even after considering the three sets of mediating factors
(although each set of mediating factors did partially
Academic Achievements in the Spring of First-Grade reduce the magnitudes of the differences).
In addition to child and family characteristics, home
Table 4 presents the differences in regression estimates environment was also important for some groups of
(from Appendix Table 2) of generation status by country children. Specifically, after controlling for home
of origin on reading and math outcomes in the spring of environment, the significantly lower reading scores for (1)
first-grade (panel A for Latin American regions, panel B first-generation children from Puerto Rico compared to
for Asian regions, and panel C for North America, non-Hispanic white children became non-significant, and
Europe, the Caribbean, and Africa) using Models 1 to 4, (2) first-generation children from Cuba compared to their
which allows us to determine the influence of each set of second-generation counterparts changed from 1% to 5%
mediating factors on the estimates of generation status (the magnitude of the reduction in coefficients was 26%),
by country of origin on children’s academic achievements and compared to non-Hispanic white children changed
by the end of first grade. For example, the first number (- from 0.1% to 1%.
3.29) in the first column, first row represents the Controlling for school and neighborhood environments
difference in reading scores between first- and second- also partially explained the significantly lower reading
generation children from Puerto Rico: a positive number scores, compared to non-Hispanic white children, for
indicates that first-generation children performed better second-generation children from Mexico (the magnitude
than the second-generation children and a negative of the reduction in coefficients was about 46%) and for
number indicates otherwise. T-tests were used to first-generation children from Cuba (the magnitude of the
determine the significance level of the differences in reduction in coefficients was about 14%). Similarly, these
academic achievements between the two comparison factors to some extent explained the significantly lower
groups. No significant results were found for children math scores, compared to non-Hispanic white children,
 Wei-Jui Han 297

Table 4. Differences in Academic Achievements in the Spring of First-Grade by Country of Origin

and Generation Status Compared to Third and Later Generation non-Hispanic White Children.

A. Latin America Model 1 Model 2 Model 3 Model 4

Reading
Puerto Rico
st nd
1 vs. 2 generation -3.29 -4.89 -4.17 -3.93
st rd
1 vs. 3 + generation -8.20 * -6.29 * -5.61 -4.69
nd rd
2 vs. 3 + generation -4.91 *** -1.40 -1.44 -0.76
Central America
st nd
1 vs. 2 generation 3.61 0.31 0.45 -0.10
st rd
1 vs. 3 + generation -0.05 0.98 0.34 0.33
nd rd
2 vs. 3 + generation -3.66 *** 0.28 -0.11 0.43
South America
st nd
1 vs. 2 generation -3.27 -2.16 -1.54 -1.32
st rd
1 vs. 3 + generation -3.76 * -2.31 -2.14 -1.63
nd rd
2 vs. 3 + generation -0.49 -0.15 -0.60 -0.31
Dominican Republic
st nd
1 vs. 2 generation -1.22 -1.11 -1.57 -1.71
st rd
1 vs. 3 + generation -7.24 *** -2.65 -2.93 -2.52
nd rd
2 vs. 3 + generation -6.02 *** -1.54 -1.36 -0.81
Mexico
st nd
1 vs. 2 generation -0.70 -0.47 -0.19 -0.09
st rd
1 vs. 3 + generation -7.12 *** -1.76 -1.64 -0.88
nd rd
2 vs. 3 + generation -6.42 *** -1.29 ** -1.45 *** -0.79
Cuba
st nd
1 vs. 2 generation -5.67 ** -6.94 ** -5.13 * -4.71 *
st rd
1 vs. 3 + generation -5.79 *** -6.96 *** -5.95 ** -5.14 *
nd rd
2 vs. 3 + generation -0.12 -0.02 -0.82 -0.43
Math
Puerto Rico
st nd
1 vs. 2 generation -3.90 -5.89 * -5.16 * -4.86
st rd
1 vs. 3 + generation -11.29 *** -9.83 *** -8.97 *** -8.33 ***
nd rd
2 vs. 3 + generation -7.39 *** -3.94 ** -3.81 ** -3.47 *
Central America
st nd
1 vs. 2 generation 5.22 * 2.88 2.69 2.09
st rd
1 vs. 3 + generation -0.96 0.90 0.47 0.35
nd rd
2 vs. 3 + generation -6.18 *** -1.98 ** -2.22 ** -1.74 *
South America
st nd
1 vs. 2 generation -5.05 * -4.87 * -4.25 * -4.10 *
st rd
1 vs. 3 + generation -7.15 *** -5.77 ** -5.35 ** -5.04 **
nd rd
2 vs. 3 + generation -2.10 ** -0.90 -1.10 -0.94
Dominican Republic
st nd
1 vs. 2 generation -0.32 -0.55 -1.09 -1.35
st rd
1 vs. 3 + generation -9.66 *** -4.88 -5.13 -5.10
nd rd
2 vs. 3 + generation -9.34 *** -4.33 *** -4.04 *** -3.75 ***
Mexico
st nd
1 vs. 2 generation 0.03 0.30 0.64 0.66
st rd
1 vs. 3 + generation -6.58 *** -0.76 -0.40 0.10
nd rd
2 vs. 3 + generation -6.61 *** -1.06 * -1.04 * -0.56
Cuba
st nd
1 vs. 2 generation -3.56 -4.28 -2.31 -2.15
st rd
1 vs. 3 + generation -3.83 -3.91 -2.54 -2.12
nd rd
2 vs. 3 + generation -0.27 0.37 -0.23 0.03
298 Educ. Res. Rev.

Table 4. Continued.

B. Asia Model 1 Model 2 Model 3 Model 4

Reading
East Asia
st nd
1 vs. 2 generation -0.42 -0.48 -0.82 -0.95
st rd
1 vs. 3 + generation 4.10 *** 3.48 ** 3.55 ** 3.49 **
nd rd
2 vs. 3 + generation 4.52 *** 3.96 *** 4.37 *** 4.44 ***
Vietnam/Thailand/Cambodia/Laos
st nd
1 vs. 2 generation 4.49 1.85 2.29 2.55
st rd
1 vs. 3 + generation 2.89 3.85 4.79 * 5.25 *
nd rd
2 vs. 3 + generation -1.60 2.00 * 2.50 *** 2.70 ***
Other South East Asia
st nd
1 vs. 2 generation 1.12 1.63 2.25 2.79 *
st rd
1 vs. 3 + generation 3.09 * 4.14 *** 4.83 *** 5.62 ***
nd rd
2 vs. 3 + generation 1.97 *** 2.51 *** 2.58 *** 2.83 ***
India
st nd
1 vs. 2 generation -0.77 -1.44 -0.69 -0.42
st rd
1 vs. 3 + generation 3.40 0.94 1.89 2.36
nd rd
2 vs. 3 + generation 4.17 *** 2.38 *** 2.58 ** 2.78 ***
Math
East Asia
st nd
1 vs. 2 generation -0.19 -0.36 -0.72 -0.83
st rd
1 vs. 3 + generation 1.87 1.94 2.02 1.94
nd rd
2 vs. 3 + generation 2.06 *** 2.30 *** 2.74 *** 2.77 ***
Vietnam/Thailand/Cambodia/Laos
st nd
1 vs. 2 generation 1.45 -2.08 -1.65 -1.51
st rd
1 vs. 3 + generation -0.96 -0.73 0.20 0.48
nd rd
2 vs. 3 + generation -2.41 ** 1.35 1.85 * 1.99 **
Other South East Asia
st nd
1 vs. 2 generation -1.41 -0.83 -0.26 0.12
st rd
1 vs. 3 + generation -3.24 * -1.48 -0.58 -0.02
nd rd
2 vs. 3 + generation -1.83 ** -0.65 -0.32 -0.14
India
st nd
1 vs. 2 generation -1.97 -2.02 -1.27 -1.11
st rd
1 vs. 3 + generation -0.73 -2.09 -0.97 -0.65
nd rd
2 vs. 3 + generation 1.24 -0.07 0.30 0.46

for second-generation children from Puerto Rico, Central Child and family characteristics seemed to make little
America, and Mexico (the magnitudes of the reductions in difference in accounting for the significantly better
coefficients were about 10%, 22%, and 46%, reading scores for children of Asian origin; in some
respectively). cases, the differences became even larger after
controlling for these factors (e.g., for the reading skills of
Asian children: Panel B of Table 4 presents the reading second-generation children from
and math results for children of Asian origin. Generally Vietnam/Thailand/Cambodia/Laos, and first- and second-
speaking, first- and second-generation children generation children from other Southeast Asia). The
performed significantly better than non-Hispanic white significant differences in math scores between second-
children on reading skills, the effect sizes ranged from 0.2 generation children from East Asia and non-Hispanic
(for first- and second-generation children from South- white children also increased with each set of mediators.
Central/Western Asia) to 0.5 (for second-generation Child and family background, however, were able to
children from East Asia). In contrast, all groups of account wholly for the significantly lower math scores of
children tended to have significantly lower math scores the first- and second-generation children from other
compared to non-Hispanic white children (except for Southeast Asia and of second-generation children from
second-generation children from East Asia who Vietnam/Thailand/Cambodia/Laos as compared to non-
performed significantly better). No significant differences Hispanic white children.
were found between first- and second-generation children After controlling for home environment, the differences
on reading and math. in reading scores between first- and second-generation
 Wei-Jui Han 299

Table 4. continued.

C. North America, Europe, Caribbean, and Africa Model 1 Model 2 Model 3 Model 4
Reading
North America
st nd
1 vs. 2 generation 1.11 2.11 1.74 1.50
st rd
1 vs. 3 + generation 0.84 -1.01 -1.23 -1.67
nd rd
2 vs. 3 + generation -0.27 -3.12 -2.97 -3.17
Europe (including Russia)
st nd
1 vs. 2 generation 1.36 1.02 1.33 1.45
st rd
1 vs. 3 + generation 1.76 1.14 1.55 1.65
nd rd
2 vs. 3 + generation 0.40 0.12 0.22 0.20
Caribbean
st nd
1 vs. 2 generation -2.16 0.56 0.15 1.27
st rd
1 vs. 3 + generation -4.14 0.09 -0.44 1.09
nd rd
2 vs. 3 + generation -1.98 -0.47 -0.59 -0.18
Africa
st nd
1 vs. 2 generation 4.69 5.33 * 6.28 * 7.52 **
st rd
1 vs. 3 + generation 5.52 * 5.56 * 6.42 * 7.85 **
nd rd
2 vs. 3 + generation 0.83 0.23 0.14 0.33
Math
North America
st nd
1 vs. 2 generation 2.69 3.36 3.34 3.20
st rd
1 vs. 3 + generation 0.70 -0.96 -0.88 -1.30
nd rd
2 vs. 3 + generation -1.99 -4.32 * -4.22 * -4.50 **
Europe (including Russia)
st nd
1 vs. 2 generation -0.11 -0.22 0.10 0.20
st rd
1 vs. 3 + generation 0.11 -0.07 0.38 0.41
nd rd
2 vs. 3 + generation 0.22 0.15 0.28 0.21
Caribbean
st nd
1 vs. 2 generation 1.39 3.51 2.99 4.01 *
st rd
1 vs. 3 + generation -5.14 ** -1.00 -1.49 -0.34
nd rd
2 vs. 3 + generation -6.53 *** -4.51 *** -4.48 *** -4.35 ***
Africa
st nd
1 vs. 2 generation 2.19 2.59 3.82 4.76
st rd
1 vs. 3 + generation 0.76 0.89 2.13 3.15
nd rd
2 vs. 3 + generation -1.43 -1.70 -1.69 -1.61

Note. Model 1 controls only for immigrant generation status by country of origin and race/ethnicity (for the 3rd and later generation)
variables without any other covariates. Model 2 adds controls for child’s gender, age, being low birth weight, being at least 2 weeks
premature, current weight and height in the fall of kindergarten, number of moves since birth, attending center-based care before
kindergarten, mother’s age in the fall of kindergarten, parental education (either mother or father, whichever is higher), mother
married at child’s birth, number of people age < 18 in household, mother working full-time, family’s SES prestigious score, home
language is not English, region of residence, location of residence, and a set of dummy variables indicating missing values for
controlled covariates. Model 3 adds controls for the home environment. Model 4 adds controls for the quality of neighborhood and
school environment. Additional details about covariates are presented in Appendix Table 1. T-tests were used to determine the
significance level of the differences in estimated coefficients between generations (based on the regression results shown in
Appendix Table 2). * p < .05. ** p < .01. *** p < .001.

children from Vietnam/Thailand/Cambodia/Laos and non- white children, and in the math scores between second-
Hispanic white children, and the difference in math generation children from
scores between second-generation children from Vietnam/Thailand/Cambodia/Laos and non-Hispanic
Vietnam/Thailand/Cambodia/Laos and non-Hispanic white children.
white children, became larger and significant. Similarly,
after controlling for the school and neighborhood North American, European, Caribbean, and African
environments, the differences became larger in the children: Panel C of Table 4 presents the reading and
reading scores between first- and second-generation math results for four groups of children. Regarding the
children from Vietnam/Thailand/Cambodia/Laos and non- reading results as shown in the top panel of Panel C,
Hispanic white children and between first-generation first-generation children from Africa performed
children from other Southeast Asia and non-Hispanic significantly better than both non-Hispanic white children
300 Educ. Res. Rev.

Table 5. Changes in Academic Achievements during Kindergarten and the Spring of First-Grade by Country
of Origin and Generation Status, Compared to Third and Later Generation non-Hispanic White Children

A. Latin ca Model 1 Model 2 Model 3 Model 4

Reading
Puerto Rico
st nd
1 vs. 2 generation -1.26 -1.78 -1.67 -1.61
st rd
1 vs. 3 + generation -0.81 -1.24 -1.30 -0.96
nd rd
2 vs. 3 + generation 0.45 0.54 0.37 0.65
Central America
st nd
1 vs. 2 generation 2.71 2.45 2.48 2.37
st rd
1 vs. 3 + generation 3.29 * 3.32 * 3.18 * 3.15 *
nd rd
2 vs. 3 + generation 0.58 0.87 0.70 0.78
South America
st nd
1 vs. 2 generation 0.54 0.43 0.46 0.54
st rd
1 vs. 3 + generation 0.54 0.22 0.01 0.19
nd rd
2 vs. 3 + generation 0.00 -0.21 -0.45 -0.35
Dominican Republican
st nd
1 vs. 2 generation -5.49 ** -4.18 * -4.73 * -4.57 *
st rd
1 vs. 3 + generation -5.10 ** -4.17 * -4.34 * -4.00 *
nd rd
2 vs. 3 + generation 0.39 0.51 0.39 0.57
Mexico
st nd
1 vs. 2 generation -1.40 -1.44 -1.50 -1.38
st rd
1 vs. 3 + generation -1.13 -0.91 -1.11 -0.87
nd rd
2 vs. 3 + generation 0.27 0.53 0.39 0.51
Cuba
st nd
1 vs. 2 generation -0.35 -0.77 -0.83 -0.67
st rd
1 vs. 3 + generation -0.97 -1.49 -1.76 -1.57
nd rd
2 vs. 3 + generation -0.62 -0.72 -0.93 -0.90
Math
Puerto Rico
st nd
1 vs. 2 generation -3.34 -3.95 -3.94 -3.80
st rd
1 vs. 3 + generation -3.59 -4.16 * -4.15 * -4.13 *
nd rd
2 vs. 3 + generation -0.25 -0.21 -0.21 -0.33
Central America
st nd
1 vs. 2 generation 1.74 1.89 1.85 1.84
st rd
1 vs. 3 + generation 1.42 1.71 1.59 1.50
nd rd
2 vs. 3 + generation -0.32 -0.18 -0.26 -0.34
South America
st nd
1 vs. 2 generation -0.43 -0.69 -0.68 -0.77
st rd
1 vs. 3 + generation -0.95 -1.18 -1.20 -1.36
nd rd
2 vs. 3 + generation -0.52 -0.49 -0.52 -0.59
Dominican Republican
st nd
1 vs. 2 generation -0.73 -0.45 -0.54 -0.68
st rd
1 vs. 3 + generation -1.90 -1.20 -1.27 -1.51
nd rd
2 vs. 3 + generation -1.17 -0.75 -0.73 -0.83
Mexico
st nd
1 vs. 2 generation -0.13 -0.01 0.01 -0.01
st rd
1 vs. 3 + generation 1.30 * 1.49 * 1.47 * 1.28 *
nd rd
2 vs. 3 + generation 1.43 *** 1.50 *** 1.46 *** 1.29 ***
Cuba
st nd
1 vs. 2 generation -1.22 -1.62 -1.50 -1.59
st rd
1 vs. 3 + generation -0.46 -1.07 -1.05 -1.35
nd rd
2 vs. 3 + generation 0.76 0.55 0.45 0.24
 Wei-Jui Han 301

Table 5. continued.

C. North America, Europe, Caribbean, and Africa Model 1 Model 2 Model 3 Model 4
Reading
North America
1st vs. 2nd generation 0.50 0.59 0.46 0.38
1st vs. 3rd+ generation -0.25 -0.51 -0.72 -0.83
2nd vs. 3rd+ generation -0.75 -1.10 -1.18 -1.21
Europe (including Russia)
1st vs. 2nd generation 1.35 1.27 1.28 1.33
1st vs. 3rd+ generation 1.71 * 1.52 * 1.45 * 1.52 *
2nd vs. 3rd+ generation 0.36 0.25 0.17 0.19
Caribbean
1st vs. 2nd generation 3.65 3.77 * 3.74 * 4.15 **
1st vs. 3rd+ generation 2.50 2.91 2.75 3.38 *
2nd vs. 3rd+ generation -1.15 -0.86 -0.99 -0.77
Africa
1st vs. 2nd generation 3.00 * 2.79 * 2.64 * 3.05 *
1st vs. 3rd+ generation 2.75 * 2.45 * 2.22 2.69 *
2nd vs. 3rd+ generation -0.25 -0.34 -0.42 -0.36
Math
North America
1st vs. 2nd generation 3.72 * 3.20 * 3.26 * 3.32 *
1st vs. 3rd+ generation 2.13 1.79 1.80 1.75
2nd vs. 3rd+ generation -1.59 -1.41 -1.46 -1.57
Europe (including Russia)
1st vs. 2nd generation 1.54 * 1.47 * 1.51 * 1.54 *
1st vs. 3rd+ generation 1.58 * 1,42 * 1.43 * 1.42 *
2nd vs. 3rd+ generation 0.04 -0.05 -0.08 -0.12
Caribbean
1st vs. 2nd generation 5.34 5.57 5.53 5.70
1st vs. 3rd+ generation 3.29 3.97 3.91 3.98
2nd vs. 3rd+ generation -2.05 ** -1.60 * -1.62 * -1.72 *
Africa
1st vs. 2nd generation -3.38 * -4.07 * -3.99 * -3.75 *
1st vs. 3rd+ generation -3.53 * -4.24 ** -4.17 ** -4.00 **
2nd vs. 3rd+ generation -0.15 -0.17 -0.18 -0.25
Note. Model 1 controls only for immigrant generation status by country of origin and race/ethnicity (for the 3rd and plus generation) variables and
earlier measure of corresponding outcome without any other covariates. Model 2 adds controls for child’s gender, age, being low birth weight, being at
least 2 weeks premature, current weight and height in the fall kindergarten, number of moves since birth, attending center-based care before
kindergarten, mother’s age in fall kindergarten, parental education (either mother or father, whichever is higher), mother married at child’s birth,
number of people age < 18 in household, mother working full-time, family’s SES prestigious score, home language is not English, region of residence,
location of residence, and a set of dummy variables indicating missing values for controlled covariates. Model 3 adds controls for the home
environment. Model 4 adds controls for the quality of neighborhood and school environment. Additional details about covariates are presented in
Appendix Table 1. T-tests were used to determine the significant level of the differences in estimated coefficients between generations (based on the
regression results shown in Appendix Table 3). * p < .05. ** p < .01. *** p < .001.

and their second-generation counterparts after North America performed significantly worse than non-
considering child and family backgrounds. These Hispanic white children after considering child and family
significant performance gaps became larger with each characteristics, and the significance persisted after
set of mediating factors (the effect sizes increased from considering the other two sets of mediating factors (d =
0.6 to 0.8). There is some indication from the data that 0.5). Second, first-generation children from the Caribbean
the one African child who is white may be driving the performed significantly better than their second-
significantly better performance by first-generation generation counterparts after considering the three sets
children from Africa. Still, even if the analyses are limited of mediating factors (d = 0.4), while both generations
to only black children from Africa, the significant results performed significantly worse than non-Hispanic white
hold for the reading comparisons between first- and children (d = 0.5 and d = 0.7, respectively). Child
second-generations. andfamily characteristics were able to account wholly for
Two patterns are evident from the math results (second the first-generation’s lower math scores, while all three
panel of Panel C). First, second-generation children from sets of mediators were only able to partially account for
302 Educ. Res. Rev.

Appendix Table 1. Definitions of Covariates Used in Analyses.

Constructs and variables Definitions

Child characteristics, fall kindergarten
Race and ethnicity 5 dummy variables (1=yes; 0=no) for Non-Hispanic white (as
reference group), Non-Hispanic black, Hispanic, Asia, or Other
(including Native American, Pacific Islander, Native Hawaiian, ore
more than one race).
Gender Dummy variable (1=boy; 0=girl).
Age Continuous variable, age in months, ranges from 45.77 to 96.50.
Low birth weight Dummy variable for birth weight <2500 grams (1=yes; 0=no).
Premature Dummy variable for 2 weeks or more early (1=yes; 0=no).
Height, fall kindergarten Average of two interviewer-assessed measurements in inches, ranges
from 35 to 60.
Weight, fall kindergarten Average of two interviewer-assessed measurements in lbs, ranges
from 22.5 to 100.
Number of moves since child’s birth, fall kindergarten Continuous variable. Ranges from 1 to 20.
Attending center care the year before entering Dummy variable (1=yes; 0=no) if the child ever attended center-based
kindergarten care.
Parental characteristics reported by parents
Mother’s age, fall kindergarten Continuous variable, ranges from 18 to 83.
Parental education, fall kindergarten 4 dummy variables (1=yes; 0=no) for less than high school (<12), high
school degree (=12), some college (>=13 and <15), and college and
plus (>=15) with less than high school as the reference group. Derived
from either mother or father, whichever is higher.
Mother married at child’s birth Dummy variable (1=yes; 0=no).
Mother’s employment status, fall kindergarten 3 dummy variables (1=yes; 0=no) for full-time (35 or more hours per
week), part-time (less than 35 hours per week), and not working (as
the reference group).
Family characteristics reported by parents
Number of persons age <18 in the household, fall Ordinal variable. Ranges from 1 to 11.
kindergarten
Family SES scores Continuous variable. Socioeconomic status was computed at the
household level using the following components that were collected
mainly in the fall of kindergarten, except that the income was collected
from the spring of kindergarten: father/male guardian’s education,
other/female guardian’s education, father/male guardian’s occupation,
mother/female guardian’s occupation, and household income.
Home language is not English Dummy variable for whether the home primary language of the child is
not English (1=yes; 0=no).
Region of residency 4 dummy variables (1=yes; 0=no) for northeast (reference group),
midwest, south, and west.
Location of residency 7 dummy variables (1=yes; 0=no) for:
large city (reference group)--a central city of consolidated Metropolitan
Statistical Areas (CMSA) with a population greater than or equal to
250,000, mid-size city--a central city of CMSA or Metropolitan
Statistical Areas (MSA) with a population less than 250,000, urban
fringe of large city--any incorporated place, Census Designated Place,
or nonplace territory within CMSA or MSA of a large city and defined
as urban by the Census Bureau, urban fringe of mid-size city-- any
incorporated place, Census Designated Place, or nonplace territory
within CMSA or MSA of a mid-size, large town--an incorporated place
or Census Designated Place with a population greater than or equal to
250,000 and located outside a CMSA or MSA.
small town--an incorporated place or Census Designated Place with a
population less than 250,000 and greater than 2,500, and located
outside a CMSA or MSA.
rural-- any incorporated place, Census Designated Place, or nonplace
territory and defined as rural by the Census Bureau
 Wei-Jui Han 303

Appendix Table 1. Continued.

Parental well-being and educational practices

Maternal depression composite, spring kindergarten Continuous variable. Sum of 12 items such as “how often during the
past week have you felt that you had trouble keeping your mind on
what you were doing.” Responses range from 0 (never) to 3 (most of
the time). These 12 items were extracted from Center for
Epidemiologic Studies – Depression Mood Scale (CES-D; Radloff,
1977). Cronbach Alpha of .86.
Parental educational expectations for child, fall Ordinal variable. Responses range from 1 (receive less than high
kindergarten school diploma) to 6 (to get Ph.D., MD, or other higher degree).
Importance of child having skills by entrance to Continuous variable. Mean of 5 ordinal variables -- counting, sharing,
kindergarten, fall kindergarten communication, draws, and knows letters. Responses range from 1
(essential) to 5 (not important). Cronbach Alpha of .77.
Participating in school events, fall kindergarten Ordinal variable. Sum of 7 dummy variables (1=yes; 0=no) for
attending PTA meetings, open houses, parent groups, parent advisory
meetings, volunteering at school, and participating in school
fundraisers. Ranges from 0 to 7. Cronbach Alpha of .58.
Difficulty in attending school events, spring kindergarten 8 dummy variables (1=yes; 0=no) for “inconvenient meeting times,”
“no child care,” “cannot get off from work,” “problems with safety going
to school,” “the school does not make your family feel welcome,”
“problems with transportation going to school,” “problems because
you or your family members speak a language other than English and
meetings are only conductedg in English,” and “don’t hear about
things going on at school that you might want to be involved in.” The
higher the score, the more difficulty in attending school events.
Cronbach Alpha of
Home environment reported by parent
Home learning activities, fall and spring kindergarten Average score of home learning activities from the fall and spring of
kindergarten.
Fall Kindergarten: Continuous variable. Sum of 14 items including
“learning activities in home including reading books to child” with
1(everyday) and 0 otherwise, “building things, teaching about nature,
playing sports, doing art, doing chores, singing songs, playing games”
with responses 1 (at least 3 times a week) and 2 (less than 3 times a
week),” “number of children’s books in the home” with 1 (10 or more
books) and 0 (less than 10 books), “number of music tapes, CDs, or
records in home” with 1 (5 or more) and 0 (less than 5), “child looking
at picture books outside of school” with 1 (at least once or twice a
week) and 0 (never), “child reading books outside of school” with 1 (at
least once or twice a week) and 0 (never), and “watches Sesame
Street at home or someplace else at least once a week for a period of
three months or more” with 1 (yes) and 0 (no). Cronbach Alpha of
.68.

Spring kindergarten: Continuous variable. Sum of 12 items including

“family members bring child to the library, concerts, the museum, the
zoo, and to sporting events” with 1 (yes) and 0 (no), “child looks at
picture books outside of school” with 1 (at least once or twice a week)
and 0 (never), “child reads books outside of school” with 1 (at least
once or twice a week) and 0 (never), “have home computer that child
uses” with 1 (yes) and 0 (no), “frequency of using computer” with 1 (at
least once or twice a week) and 0 (never), and “uses computer to
learn skills, to learn to draw, or for Internet” with 1 (yes) and 0 (no).
Cronbach Alpha of .51.
Attending extracurricular activities, spring kindergarten Continuous variable. Sum of 9 dummy variables (1=yes; 0=no) for the
activities in which the child participates outside of school hours
including dance lessons, athletic events, organized clubs, music
lessons, drama classes, art lessons, organized performing arts
programs, crafts lessons, and non-English language instruction.
Cronbach Alpha of .58.
304 Educ. Res. Rev.

Appendix Table 1. Continued.

Frequency of spanking child in past week, spring Ordinal variable. Ranges from 0 to 30.
kindergarten
Family routines, spring kindergarten Continuous variable. Standardized score of 8 items including “number
of days eats breakfast together” with responses ranging from 0 to 7,
“number of days child eats breakfast at regular time” with responses
ranging from 0 to 7, “number of days eats dinner together” with
responses ranging from 0 to 7, “number of days child eats dinner at
regular time” with responses ranging from 0 to 7, “goes to bed same
time each night” with 1 (yes) and 0 (no), “how often family talks about
ethnic or racial heritage” with 1 (at least several times a year) and 0
(never or almost never), “how often family discusses family’s religious
beliefs or traditions” with 1 (at least several times a year) and 0 (never
or almost never), and “how often someone in the family participates in
special cultural events or traditions connected” with racial or ethnic
background with 1 (at least several times a year) and 0 (never or
almost never). The higher the score, the more family routine activities.
Cronbach Alpha of .52.

School and neighborhood characteristics

Student minority composition, spring kindergarten Continuous variable. Standardized score of 9 items including “percent
of Hispanic students in class,” “percent of minorities in class,” “having
students who speak non-English languages,” “having students with
limited English proficiency,” “teachers speak only English,” “providing
in-class ESL,” “percent of white students at school,” “percent of LEP
students at school,” and “percent receiving both ESL and bilingual
services.” The higher the score, the more diversity the student
composition. Cronbach Alpha of .88.
Average student performance, spring kindergarten Continuous variable. Standardized score of 4 items including “percent
of school students having reading and verbal skills at or below grade
level nationally,” “percent of school students having math or
quantitative skills at or below grade level nationally,” “percent of
students having reading skills below grade level in the class,” and
“percent of students having math skills below grade level in the class.”
Cronbach Alpha of .71.
Number of years served as principal, spring kindergarten Ordinal variable. Ranges from 0 to 30.
School communication about the child to the parent, Continuous variable. Standardized score (with mean of 0 and standard
spring kindergarten deviation of 1) of 6 items such as “the school lets you know between
report cards how your child is doing in school,” “the school helps you
understand what children at your child’s age are like,” “the school
makes you aware of chances to volunteer at the school,” “the school
provides workshops, materials, or advice about how to help you child
learn at home,” “the school provides information on community
services to help your child or your family,” “the child’s teacher has sent
home ideas for things to do with your child at home (reverse coded)”
with 1 (does this very well) to 3 (does not do this at all). Cronbach
Alpha of .68.
Efforts to ease the transition into kindergarten for children, 4 dummy variables (2 each for fall and spring kindergarten; 1=yes;
fall and spring kindergarten 0=no) including “school days are shortened in the beginning of the
school year,” and “teacher visited the home of the child at the
beginning of the school year.” Cronbach Alpha of .66.
 Wei-Jui Han 305

Appendix Table 1. Continued.

Teacher-reported parental involvement in school Continuous variable. Mean of 4 ordinal variables (0=none; 1=1-25
activities, spring kindergarten percent; 2=26-50 percent; 3=51 to 75 percent; 4=76 or more percent)
asking about the percentage of children in the class whose parents
participate in the following activities: teacher-parent conferences,
volunteers regularly, attends open house, and attends art/music
events. Cronbach Alpha of .78.
School neighborhood quality composite, spring Continuous variable. Mean of 8 ordinal variables (1=big problem;
kindergarten 2=somewhat of a problem; 3=no problem) asking about how much of
a problem the following are in the neighborhood where this school is
located: tension based on racial, ethnic, or religious differences,
garbage/litter/broken glasses in the road or street or on the sidewalk
or in the yards, selling or using drugs or excessive drinking in public,
gangs, heavy traffic, violent crimes such as drive-by shootings, vacant
houses and buildings, and crimes in the neighborhood, as well as 7
dummy variables (1=yes; 0=no) including “visitors must sign in,”
“limited restroom use,” “teachers patrol hallways,” “hall pass required,”
“students bringing weapons to school,” “children or teachers being
physically attacked or involved in fights,” “children or teachers having
things taken by force or threat of force on the way to or from school.”
The higher the score, the worse the school environment and
neighborhood. Cronbach Alpha of .81.
Observers’ overall rating of school safety Ordinal variable. Ranges from 1 (very unsafe) to 4 (very safe).
Parental reports of neighborhood quality composite, Continuous variable. Standardized score of 6 items asking about
spring kindergarten problems in the area around the house or apartment –
garbage/litter/broken glasses, selling/using drugs, burglary/rubbery in
the area, violent crimes like drive-by shootings, and vacant houses
with responses ranging from 1 (big problem) to 3 (no problem), and
the overall rating of safety with 1 (not at all safe) to 3 (very safe).
Cronbach Alpha of .77.
Note. Details about missing data and missing data dummy variables are available from the author upon request.

Appendix Table 2. OLS Regression Estimates of Country of Origin by Generation on Academic Achievements in the Spring of
First-Grade.

A. Reading Model 1 Model 2 Model 3 Model 4

First generation
North America 0.84 (1.73) -1.01 (1.43) -1.23 (1.42) -1.67 (1.48)
Europe (including Russia) 1.76 (1.00) 1.14 (0.98) 1.55 (1.02) 1.65 (1.04)
Caribbean -4.14 (2.79) 0.09 (2.85) -0.44 (2.29) 1.09 (2.40)
Puerto Rico -8.20 (3.97)* -6.29 (3.06)* -5.61 (2.90) -4.69 (2.83)
Central America -0.05 (2.01) 0.98 (1.98) 0.34 (2.03) 0.33 (1.83)
South America -3.76 (1.87)* -2.31 (1.90) -2.14 (1.76) -1.63 (1.80)
Dominican Republic -7.24 (1.76)*** -2.65 (2.42) -2.93 (2.56) -2.52 (2.56)
Mexico -7.12 (0.92)*** -1.76 (0.94) -1.64 (0.92) -0.88 (0.94)
Cuba -5.79 (1.33)*** -6.96 (1.92)*** -5.95 (1.87)** -5.14 (2.02)*
East Asia 4.10 (1.16)*** 3.48 (1.24)** 3.55 (1.21)** 3.49 (1.18)**
Vietnam/Thailand/Cambodia/Laos 2.89 (2.40) 3.85 (2.12) 4.79 (2.16)* 5.25 (2.16)*
Other South East Asia 3.09 (1.40)* 4.14 (1.23)*** 4.83 (1.20)*** 5.62 (1.20)***
India 3.40 (1.90) 0.94 (1.66) 1.89 (1.71) 2.36 (1.71)
South-Central/West Asia -1.70 (2.38) -0.42 (2.14) 0.00 (2.08) -0.45 (2.14)
Africa 5.52 (2.60)* 5.56 (2.32)* 6.42 (2.51)* 7.85 (2.49)**
Oceania (excluding Australia) -0.90 (0.93) -0.22 (0.89) -0.12 (0.88) 0.03 (0.87)
Second generation
North America -0.27 (1.86) -3.12 (1.80) -2.97 (1.68) -3.17 (1.67)
Europe (including Russia) 0.40 (0.65) 0.12 (0.57) 0.22 (0.56) 0.20 (0.56)
Caribbean -1.98 (1.22) -0.47 (1.15) -0.59 (1.13) -0.18 (1.13)
Puerto Rico -4.91 (1.30)*** -1.40 (1.14) -1.44 (1.12) -0.76 (1.15)
Central America -3.66 (0.78)*** 0.28 (0.77) -0.11 (0.76) 0.43 (0.74)
South America -0.49 (0.75) -0.15 (0.67) -0.60 (0.66) -0.31 (0.66)
306 Educ. Res. Rev.

Appendix Table 2. Continued.

Dominican Republic -6.02 (1.25)*** -1.54 (1.13) -1.36 (1.11) -0.81 (1.10)
Mexico -6.42 (0.38)*** -1.29 (0.44)** -1.45 (0.44)*** -0.79 (0.46)
Cuba -0.12 (1.36) -0.02 (1.15) -0.82 (1.10) -0.43 (1.09)
East Asia 4.52 (0.62)*** 3.96 (0.60)*** 4.37 (0.60)*** 4.44 (0.61)***
Vietnam/Thailand/Cambodia/Laos -1.60 (0.84) 2.00 (0.81)* 2.50 (0.78)*** 2.70 (0.77)***
Other South East Asia 1.97 (0.59)*** 2.51 (0.54)*** 2.58 (0.55)*** 2.83 (0.55)***
India 4.17 (0.91)*** 2.38 (0.82)*** 2.58 (0.81)** 2.78 (0.84)***
South-Central/West Asia 1.84 (1.22) 1.63 (1.09) 1.87 (1.09) 2.06 (1.08)
Africa 0.83 (0.93) 0.23 (1.10) 0.14 (1.04) 0.33 (1.03)
Oceania (excluding Australia) -2.39 (2.11) 0.66 (1.90) 0.36 (1.83) 1.31 (1.87)
Third+ generation
Black -5.51 (0.24)*** -2.46 (0.26)*** -2.40 (0.26)*** -1.76 (0.27)***
Hispanic -3.40 (0.27)*** -0.92 (0.27)*** -1.02 (0.26)*** -0.73 (0.27)**
Asian -2.21 (0.46)*** 0.43 (0.45) 0.79 (0.44) 1.22 (0.44)**
Other (including mixed race) -4.74 (0.44)*** -2.27 (0.40)*** -2.21 (0.39)*** -1.50 (0.39)***
Child Characteristics
Boy -2.02 (0.14)*** -1.76 (0.14)*** -1.75 (0.14)***
Child age in months 0.16 (0.02)*** 0.16 (0.02)*** 0.16 (0.02)***
Low birth weight (<2500 g) -0.93 (0.26)*** -0.93 (0.26)*** -0.91 (0.26)***
Premature (>=2 weeks early) -0.31 (0.20) -0.39 (0.20)* -0.36 (0.20)
Height 0.32 (0.05)*** 0.30 (0.05)*** 0.29 (0.05)***
Weight -0.03 (0.01)* -0.03 (0.01)* -0.03 (0.01)*
Number of moves since birth 0.00 (0.06) 0.07 (0.06) 0.08 (0.06)
Center-based care before entering 0.87 (0.15)*** 0.65 (0.15)*** 0.56 (0.15)***
kindergarten
Parent Characteristics
Mother’s age 0.01 (0.01) 0.02 (0.01) 0.01 (0.01)
Highest parental education
High school degree 2.01 (0.33)*** 1.69 (0.33)*** 1.62 (0.32)***
Some college 3.10 (0.35)*** 2.38 (0.34)*** 2.24 (0.34)***
College and plus 4.06 (0.41)*** 3.14 (0.41)*** 2.88 (0.40)***
Mother married at birth 1.55 (0.20)*** 1.18 (0.20)*** 1.02 (0.20)***
Mother currently works full-time -0.61 (0.15)*** -0.50 (0.15)*** -0.47 (0.15)***
Family Characteristics
Number of persons age <18 in the -0.82 (0.07)*** -0.70 (0.07)*** -0.65 (0.07)***
household
Social economic status (SES) 2.11 (0.15)*** 1.59 (0.15)*** 1.37 (0.14)***
Home language is not English -1.46 (0.28)*** -1.16 (0.28)*** -0.84 (0.28)**
Region of residence
Mid-West 0.10 (0.21) 0.01 (0.21) -0.05 (0.21)
South 0.92 (0.20)*** 0.82 (0.20)*** 0.98 (0.21)***
West 0.66 (0.23)** 0.51 (0.23)* 0.56 (0.24)*
Location of residency
Rural -1.90 (0.28)*** -1.59 (0.28)*** -1.94 (0.29)***
Small town -1.03 (0.29)*** -0.68 (0.29)* -0.82 (0.30)**
Large town -0.17 (0.45) 0.07 (0.44) -0.22 (0.45)
Mid-size suburban -1.07 (0.31)*** -0.94 (0.30)** -1.24 (0.31)***
Large-size suburban 0.03 (0.21) -0.00 (0.21) -0.39 (0.21)
Mid-size city -0.53 (0.23)* -0.46 (0.22)* -0.69 (0.23)***
Parental well-being and educational practice
Maternal depression -0.05 (0.01)*** -0.05 (0.01)***
Parental educational expectations for child 0.48 (0.07)*** 0.48 (0.07)***
Importance of child having skills by entrance 1.41 (0.14)*** 1.46 (0.14)***
to kindergarten
Participating in school events 0.14 (0.05)** 0.09 (0.05)
Difficulty in attending school events -0.20 (0.06)** -0.18 (0.06)**
Home Environment
Home learning activities 0.39 (0.04)*** 0.37 (0.04)***
Attending extracurricular activities 0.15 (0.06)*** 0.13 (0.06)*
Frequency of spanking child in past week -0.16 (0.07)* -0.16 (0.07)*
 Wei-Jui Han 307

Appendix Table 2. Continued.

Family routines 0.32 (0.16)* 0.33 (0.16)*

School and neighborhood characteristics
Student minority composition in class -0.42 (0.14)**
Average student performance 0.82 (0.12)***
Number of years served as principal 0.02 (0.01)*
School communication about the child -0.30 (0.12)**
Teachers’ efforts to ease the transition into -0.26 (0.10)**
kindergarten for children
Teacher-reported parental involvement in 0.53 (0.10)***
school activities
School neighborhood quality 0.64 (0.17)***
Observers’ overall rating of school safety 0.11 (0.11)
Residential neighborhood quality 0.11 (0.11)

Observations 16102 16102 16102 16102

Adjusted R-squared 0.07 0.21 0.24 0.24

those of the second-generation (the magnitude of the academic performance but also the progress in academic
reduction in coefficients was 34% with controls for child achievements of children of immigrants. The exceptions
and family backgrounds). were children who attended center-based care before
kindergarten, had a non-English home language, and
attended extracurricular activities. Specifically, children
Change Model who attended center-based care before kindergarten or
attended more extracurricular activities had slower
Table 5 presents the estimates of generation status by learning paces compared to those who did not, although
country of origin (obtained from Appendix Table 3) for the former groups had significantly higher initial and first-
academic achievements in the spring of first grade for grade scores compared to the latter. In other words, the
children with non-missing academic achievement data difference in academic achievements between these
during the kindergarten year. Models 1 to 4 were groups may become narrower over time. In contrast,
analyzed with controls for earlier measures from the although children with non-English home language had
kindergarten year to evaluate the additional effects that significantly lower initial scores than their counterparts,
extend beyond their individual initial effects on the the former had a faster learning pace, thus narrowing the
changes in academic achievements over time. The gap between them over time.
regions in Table 5 were grouped the same as in Table 4. Regarding results for children of Latin American origin,
Generally, adding each set of mediating factors to the first-generation children from Central America had a
analyses did not change the results substantially for all significantly faster learning pace for reading compared to
groups, with a few exceptions. Child and family non-Hispanic white children (and thus may have better
characteristics (i.e., child’s health status, number of scores over time), while first-generation children from the
moves since birth, mother’s education, marital status at Dominican Republic had a significantly slower learning
child’s birth, work hours, and family SES), parental pace for reading compared to their second-generation
educational practices and home environments (i.e., counterparts and to non-Hispanic white children
parental educational expectations, home learning (suggesting they may lag behind over time). Also,
activities, frequency of spanking, and family routines), compared to non-Hispanic white children, first-generation
and school and neighborhood environments (i.e., student children from Puerto Rico not only had significantly lower
minority composition, average student performance, initial math scores, but also a significantly slower learning
number of years served as principal, teachers’ efforts to pace for math, suggesting this score gap may widen over
ease the transition into kindergarten for children, teacher- time. In contrast, first- and second-generation children
reported parental involvement, school neighborhood from Mexico had lower initial scores but significantly
safety, and observer ratings of overall school safety faster learning paces than non-Hispanic white children,
measures) all have the same directions of the effects as suggesting their score gaps may narrow over time.
in Appendix Table 2 when the academic achievements by Of the results for children of Asian origin, the
first-grade were examined. Thus, for example, school significantly higher reading scores by first-generation
diversity may adversely affect not only the first-grade children from East Asia compared to non-Hispanic white
308 Educ. Res. Rev.

Appendix Table 2. continued.

B. Math Model 1 Model 2 Model 3 Model 4

First generation
North America 0.70 (1.63) -0.96 (1.58) -0.88 (1.52) -1.30 (1.55)
Europe (including Russia) 0.11 (0.95) -0.07 (0.91) 0.38 (0.95) 0.41 (0.97)
Caribbean -5.14 (1.77)** -1.00 (1.49) -1.49 (1.96) -0.34 (1.57)
Puerto Rico -11.29 (3.32)*** -9.83 (2.30)*** -8.97 (2.26)*** -8.33 (2.22)***
Central America -0.96 (2.08) 0.90 (2.10) 0.47 (2.06) 0.35 (1.99)
South America -7.15 (2.12)*** -5.77 (2.08)** -5.35 (2.00)** -5.04 (1.98)**
Dominican Republic -9.66 (2.50)*** -4.88 (3.45) -5.13 (3.73) -5.10 (3.77)
Mexico -6.58 (0.83)*** -0.76 (0.84) -0.40 (0.84) 0.10 (0.85)
Cuba -3.83 (2.64) -3.91 (3.17) -2.54 (3.20) -2.12 (3.25)
East Asia 1.87 (1.51) 1.94 (1.54) 2.02 (1.50) 1.94 (1.49)
Vietnam/Thailand/Cambodia/Laos -0.96 (3.62) -0.73 (3.72) 0.20 (3.71) 0.48 (3.83)
Other South East Asia -3.24 (1.50)* -1.48 (1.31) -0.58 (1.25) -0.02 (1.19)
India -0.73 (1.94) -2.09 (1.83) -0.97 (1.94) -0.65 (2.00)
South-Central/West Asia -1.08 (2.58) 0.30 (2.35) 0.94 (2.34) 0.66 (2.34)
Africa 0.76 (2.71) 0.89 (2.06) 2.13 (2.29) 3.15 (2.26)
Oceania (excluding Australia) -2.29 (1.21) -1.30 (1.14) -1.21 (1.16) -1.14 (1.13)
Second generation
North America -1.99 (1.89) -4.32 (1.83)* -4.22 (1.76)* -4.50 (1.71)**
Europe (including Russia) 0.22 (0.64) 0.15 (0.57) 0.28 (0.57) 0.21 (0.57)
Caribbean -6.53 (1.19)*** -4.51 (1.08)*** -4.48 (1.06)*** -4.35 (1.07)***
Puerto Rico -7.39 (1.55)*** -3.94 (1.38)** -3.81 (1.36)** -3.47 (1.36)*
Central America -6.18 (0.77)*** -1.98 (0.72)** -2.22 (0.72)** -1.74 (0.71)*
South America -2.10 (0.80)** -0.90 (0.72) -1.10 (0.70) -0.94 (0.70)
Dominican Republic -9.34 (1.15)*** -4.33 (1.07)*** -4.04 (1.02)*** -3.75 (1.03)***
Mexico -6.61 (0.36)*** -1.06 (0.43)* -1.04 (0.44)* -0.56 (0.45)
Cuba -0.27 (1.41) 0.37 (1.20) -0.23 (1.20) 0.03 (1.20)
East Asia 2.06 (0.63)*** 2.30 (0.59)*** 2.74 (0.60)*** 2.77 (0.60)***
Vietnam/Thailand/Cambodia/Laos -2.41 (0.79)** 1.35 (0.76) 1.85 (0.76)* 1.99 (0.76)**
Other South East Asia -1.83 (0.60)** -0.65 (0.55) -0.32 (0.54) -0.14 (0.54)
India 1.24 (0.90) -0.07 (0.84) 0.30 (0.82) 0.46 (0.83)
South-Central/West Asia -0.88 (1.22) -0.57 (1.04) -0.13 (1.02) 0.09 (1.00)
Africa -1.43 (1.21) -1.70 (1.24) -1.69 (1.21) -1.61 (1.19)
Oceania (excluding Australia) -6.43 (1.85)** -2.95 (1.54) -3.21 (1.48)* -2.62 (1.46)
Third+ generation
Black -7.19 (0.23)*** -4.38 (0.25)*** -4.18 (0.25)*** -3.74 (0.26)***
Hispanic -4.35 (0.25)*** -1.55 (0.26)*** -1.56 (0.25)*** -1.34 (0.26)***
Asian -3.38 (0.43)*** -0.48 (0.42) -0.04 (0.42) 0.25 (0.43)
Other (including mixed race) -5.04 (0.40)*** -2.87 (0.38)*** -2.76 (0.37)*** -2.26 (0.37)***
Child Characteristics
Boy 0.14 (0.13) 0.46 (0.13)*** 0.47 (0.13)***
Child age in months 0.24 (0.02)*** 0.24 (0.02)*** 0.24 (0.02)***
Low birth weight (<2500 g) -1.25 (0.26)*** -1.21 (0.25)*** -1.19 (0.25)***
Premature (>=2 weeks early) -0.43 (0.20)* -0.51 (0.19)** -0.49 (0.19)*
Height 0.37 (0.05)*** 0.36 (0.05)*** 0.35 (0.05)***
Weight -0.03 (0.01)** -0.04 (0.01)*** -0.03 (0.01)**
Number of moves since birth 0.07 (0.06) 0.14 (0.06)* 0.15 (0.06)*
Center-based care before entering 0.92 (0.15)*** 0.67 (0.15)** 0.60 (0.15)**
kindergarten
Parent Characteristics
Mother’s age 0.04 (0.01)** 0.04 (0.01)** 0.03 (0.01)**
Highest parental education
High school degree 1.02 (0.31)*** 0.70 (0.31)* 0.64 (0.31)*
Some college 2.24 (0.33)*** 1.51 (0.33)*** 1.40 (0.33)***
College and plus 3.43 (0.40)*** 2.43 (0.39)*** 2.23 (0.39)***
Mother married at birth 1.18 (0.20)*** 0.79 (0.20)*** 0.67 (0.19)***
Mother currently works full-time -0.24 (0.15) -0.11 (0.15) -0.09 (0.15)
Family Characteristics
 Wei-Jui Han 309

Appendix Table 2. continued.

Number of persons age <18 in the household -0.45 (0.07)*** -0.31 (0.07)*** -0.27 (0.07)***
Social economic status (SES) 1.92 (0.15)*** 1.34 (0.15)*** 1.17 (0.15)***
Home language is not English -1.68 (0.27)*** -1.34 (0.27)*** -1.09 (0.28)***
Region of residence
Mid-West 1.07 (0.21)*** 1.02 (0.21)*** 1.00 (0.21)***
South 1.56 (0.20)*** 1.51 (0.20)*** 1.51 (0.21)***
West 0.78 (0.22)*** 0.62 (0.22)** 0.60 (0.23)**
Location of residency
Rural -1.27 (0.27)*** -0.97 (0.27)*** -1.13 (0.28)***
Small town -1.20 (0.29)*** -0.84 (0.29)** -0.89 (0.30)**
Large town -0.71 (0.44) -0.51 (0.44) -0.80 (0.44)
Mid-size suburban -0.17 (0.29) -0.02 (0.28) -0.14 (0.29)
Large-size suburban 0.34 (0.20) 0.30 (0.20) 0.03 (0.21)
Mid-size city -0.12 (0.22) -0.07 (0.22) -0.23 (0.22)
Parental well-being and educational practice
Maternal depression -0.05 (0.01)*** -0.05 (0.01)***
Parental educational expectations for child 0.45 (0.07)*** 0.44 (0.07)***
Importance of child having skills by entrance to 1.09 (0.14)*** 1.14 (0.14)***
kindergarten
Participating in school events 0.24 (0.05)*** 0.19 (0.05)***
Difficulty in attending school events -0.20 (0.06)*** -0.19 (0.06)**
Home Environment
Home learning activities 0.37 (0.04)*** 0.35 (0.04)***
Attending extracurricular activities 0.32 (0.05)*** 0.30 (0.05)***
Frequency of spanking child in past week -0.28 (0.07)*** -0.28 (0.07)***
Family routines -0.06 (0.16) -0.07 (0.16)
School and neighborhood characteristics
Student minority composition in class -0.28 (0.14)*
Average student performance 0.88 (0.12)***
Number of years served as principal 0.01 (0.01)
School communication about the child -0.04 (0.12)
Teachers’ efforts to ease the transition into -0.03 (0.09)
kindergarten for children
Teacher-reported parental involvement in school 0.33 (0.10)***
activities
School neighborhood quality 0.20 (0.17)
Observers’ overall rating of school safety 0.27 (0.11)*
Residential neighborhood quality 0.07 (0.11)
Observations 16382 16382 16382 16382
Adjusted R-squared 0.10 0.22 0.25 0.25
Note. Reference group is Non-Hispanic whites. Standard errors shown in parentheses are corrected for school clustering using
Huber-White methods. Details about covariates are presented in Appendix Table 1.
* p < .05. ** p < .01. *** p < .001.

children (from the top panel of Panel B of Table 4) compared to non-Hispanic white children (as shown in
became wider over time judging by the significant positive the top panel of Panel B of Table 4), but also had a
difference in the top panel of Panel B of Table 5. significantly faster learning pace for reading compared to
Although there were no significant differences in first- non-Hispanic white children as shown by the significant
grade reading scores between first- and second- positive difference in the top panel of Panel B of Table 5.
generation children from East Asia as shown in the top Although first- and second-generation children from other
panel of Panel B of Table 4, the former appeared to have Southeast Asia had a slower learning pace for math
a significantly faster learning pace than the latter as compared to non-Hispanic white children, this was
shown by the significant positive difference in the top accounted for entirely by child and family backgrounds.
panel of Panel B of Table 5. Likewise, first- and second- Panel C of Table 5 presents results for children from
generation children from other Southeast Asia not only North America, Europe, the Caribbean, and Africa. First,
had significantly higher reading scores by first grade with respect to the reading results, first-generation
310 Educ. Res. Rev.

Appendix Table 3. OLS Regression Estimates of Changes in Academic Achievements during Kindergarten and First-Grade by
Country of Origin and Generational Status.

A. Reading Model 1 Model 2 Model 3 Model 4

First generation
North America -0.25 (0.85) -0.51 (0.86) -0.72 (0.87) -0.83 (0.91)
Europe (including Russia) 1.71 (0.69)* 1.52 (0.69)* 1.45 (0.71)* 1.52 (0.71)*
Caribbean 2.50 (1.78) 2.91 (1.59) 2.75 (1.44) 3.38 (1.42)*
Puerto Rico -0.81 (2.34) -1.24 (2.30) -1.30 (2.27) -0.96 (2.22)
Central America 3.29 (1.43)* 3.32 (1.33)* 3.18 (1.29)* 3.15 (1.28)*
South America 0.54 (1.24) 0.22 (1.18) 0.01 (1.16) 0.19 (1.15)
Dominican Republic -5.10 (1.62)** -4.17 (1.72)* -4.34 (1.79)* -4.00 (1.82)*
Mexico -1.13 (0.79) -0.91 (0.81) -1.11 (0.80) -0.87 (0.82)
Cuba -0.97 (1.09) -1.49 (1.15) -1.76 (1.21) -1.57 (1.17)
East Asia 2.66 (0.67)*** 2.57 (0.68)*** 2.52 (0.68)*** 2.47 (0.67)***
Vietnam/Thailand/Cambodia/Laos 2.04 (2.11) 2.20 (2.20) 2.33 (2.05) 2.66 (2.08)
Other South East Asia 2.38 (0.72)*** 2.21 (0.76)** 2.22 (0.78)** 2.49 (0.77)***
India 0.34 (1.30) -0.28 (1.26) -0.29 (1.22) 0.00 (1.18)
South-Central/West Asia -1.22 (1.77) -1.86 (1.72) -1.99 (1.72) -2.16 (1.75)
Africa 2.75 (1.18)* 2.45 (1.13)* 2.22 (1.15) 2.69 (1.15)*
Oceania (excluding Australia) -0.07 (0.69) 0.00 (0.66) -0.03 (0.66) 0.05 (0.66)
Second generation
North America -0.75 (1.19) -1.10 (1.16) -1.18 (1.15) -1.21 (1.17)
Europe (including Russia) 0.36 (0.41) 0.25 (0.40) 0.17 (0.40) 0.19 (0.40)
Caribbean -1.15 (0.78) -0.86 (0.77) -0.99 (0.77) -0.77 (0.77)
Puerto Rico 0.45 (0.86) 0.54 (0.88) 0.37 (0.89) 0.65 (0.89)
Central America 0.58 (0.59) 0.87 (0.60) 0.70 (0.60) 0.78 (0.60)
South America 0.00 (0.45) -0.21 (0.47) -0.45 (0.47) -0.35 (0.47)
Dominican Republic 0.39 (0.80) 0.51 (0.83) 0.39 (0.84) 0.57 (0.85)
Mexico 0.27 (0.33) 0.53 (0.37) 0.39 (0.38) 0.51 (0.38)
Cuba -0.62 (0.85) -0.72 (0.84) -0.93 (0.82) -0.90 (0.82)
East Asia -0.05 (0.40) -0.22 (0.41) -0.28 (0.41) -0.24 (0.41)
Vietnam/Thailand/Cambodia/Laos 0.76 (0.58) 1.15 (0.59) 1.16 (0.60) 1.26 (0.60)*
Other South East Asia 1.14 (0.36)** 1.04 (0.38)** 0.96 (0.38)* 1.03 (0.38)**
India -0.33 (0.62) -0.85 (0.63) -0.97 (0.63) -0.81 (0.63)
South-Central/West Asia 1.06 (0.80) 0.61 (0.82) 0.50 (0.82) 0.53 (0.82)
Africa -0.25 (0.71) -0.34 (0.74) -0.42 (0.74) -0.36 (0.73)
Oceania (excluding Australia) 0.86 (1.23) 0.46 (1.26) 0.43 (1.26) 0.77 (1.27)
Third+ generation
Black -1.63 (0.16)*** -1.16 (0.18)*** -1.20 (0.18)*** -0.98 (0.19)***
Hispanic -0.13 (0.18) 0.10 (0.19) 0.02 (0.19) 0.09 (0.20)
Asian 0.41 (0.28) 0.59 (0.30) 0.57 (0.31) 0.74 (0.31)*
Other (including mixed race) -1.20 (0.28)*** -0.91 (0.28)*** -0.94 (0.29)*** -0.75 (0.29)**
Child Characteristics
Boy -0.38 (0.10)*** -0.40 (0.10)*** -0.41 (0.10)***
Child age in months -0.09 (0.01)*** -0.08 (0.01)*** -0.08 (0.01)***
Low birth weight (<2500 g) -0.25 (0.19) -0.27 (0.18) -0.27 (0.18)
Premature (>=2 weeks early) 0.06 (0.14) 0.05 (0.14) 0.06 (0.14)
Height 0.05 (0.03) 0.05 (0.03) 0.05 (0.03)
Weight 0.00 (0.01) 0.00 (0.01) 0.00 (0.01)
Number of moves since birth 0.01 (0.04) 0.02 (0.04) 0.02 (0.04)
Center-based care before entering kindergarten -0.26 (0.11)** -0.26 (0.11)* -0.27 (0.11)**
Parent Characteristics
Mother’s age -0.03 (0.01)** -0.03 (0.01)** -0.03 (0.01)***
Highest parental education
High school degree 1.25 (0.26)*** 1.20 (0.26)*** 1.16 (0.25)***
Some college 1.56 (0.27)*** 1.48 (0.27)*** 1.41 (0.27)***
College and plus 1.59 (0.31)*** 1.49 (0.31)*** 1.41 (0.31)***
Mother married at birth 0.58 (0.14)*** 0.54 (0.14)*** 0.49 (0.14)***
Mother currently works full-time -0.22 (0.11)* -0.21 (0.11)* -0.21 (0.11)
Family Characteristics
Number of persons age <18 in the household 0.05 (0.05) 0.05 (0.05) 0.06 (0.05)
 Wei-Jui Han 311

Appendix Table 3. Continued.

Social economic status (SES) 0.24 (0.10)* 0.18 (0.10) 0.13 (0.10)
Home language is not English 0.41 (0.20)* 0.38 (0.21) 0.45 (0.21)*
Region of residence
Mid-West 0.14 (0.15) 0.15 (0.15) 0.11 (0.15)
South 0.16 (0.14) 0.19 (0.15) 0.28 (0.15)
West 0.05 (0.16) 0.05 (0.17) 0.05 (0.17)
Location of residency
Rural -0.27 (0.20) -0.24 (0.20) -0.43 (0.21)*
Small town 0.20 (0.21) 0.22 (0.21) 0.16 (0.22)
Large town 0.71 (0.31)* 0.76 (0.31)* 0.66 (0.31)*
Mid-size suburban 0.32 (0.23) 0.32 (0.23) 0.17 (0.23)
Large-size suburban 0.07 (0.15) 0.08 (0.15) -0.07 (0.16)
Mid-size city 0.31 (0.16) 0.33 (0.16)* 0.21 (0.17)
Parental well-being and educational practice
Maternal depression -0.01 (0.01) -0.01 (0.01)
Parental educational expectations for child 0.15 (0.05)** 0.15 (0.05)**
Importance of child having skills by entrance to -0.07 (0.10) -0.05 (0.10)
kindergarten
Participating in school events -0.02 (0.04) -0.04 (0.04)
Difficulty in attending school events -0.01 (0.05) -0.00 (0.05)
Home Environment
Home learning activities 0.07 (0.03)* 0.06 (0.03)
Attending extracurricular activities -0.10 (0.04)** -0.11 (0.04)**
Frequency of spanking child in past week -0.10 (0.05)* -0.10 (0.05)*
Family routines 0.28 (0.11)* 0.28 (0.11)*
School and neighborhood characteristics
Student minority composition -0.11 (0.11)
Average student performance 0.17 (0.08)*
Number of years served as principal 0.02 (0.01)*
School communication about the child -0.12 (0.08)
Teachers’ efforts to ease the transition into -0.18 (0.07)**
kindergarten for children
Teacher-reported parental involvement in school 0.25 (0.07)***
activities
School neighborhood quality 0.21 (0.13)
Observers’ overall rating of school safety -0.02 (0.08)
Residential neighborhood quality 0.09 (0.09)
Earlier measure
Average reading score of fall and spring 0.76 (0.01)*** 0.75 (0.01)*** 0.74 (0.01)*** 0.74 (0.01)***
kindergarten

Observations 15564 15564 15564 15564

Adjusted R-squared 0.60 0.60 0.61 0.61
Note. Reference group is Non-Hispanic whites. Standard errors shown in parentheses are corrected for school clustering using
Huber-White methods. Details about covariates are presented in Appendix Table 1.
* p < .05. ** p < .01. *** p < .001.

children from Europe had a significantly faster learning a significantly faster learning pace compared to their
pace compared to non-Hispanic white children, even second-generation counterparts and non-Hispanic white
though the former’s first-grade scores were not children. Results from first-generation children from Africa
significantly different from the latter. Second, first- suggests that they not only had significantly higher
generation children from the Caribbean had a reading scores by first-grade (as shown on the top panel
significantly faster learning pace compared to their of Panel C of Table 4), but also that these differences
second-generation counterparts after controlling for child may become wider over time. Regarding math results,
and family backgrounds and compared to non-Hispanic first-generation children from North America had a
white children after controlling for all three sets of significantly faster learning pace compared to their
mediators. Third, first-generation children from Africa had second-generation counterparts. First-generation children
 312 Educ. Res. Rev.

Appendix Table 3. continued.

B. Math Model 1 Model 2 Model 3 Model 4

First generation
Europe (including Russia) 1.58 (0.67)* 1.42 (0.65)* 1.43 (0.66)* 1.42 (0.66)*
North America 2.13 (1.10) 1.79 (1.01) 1.80 (1.01) 1.75 (1.02)
Caribbean 3.29 (3.14) 3.97 (3.04) 3.91 (2.97) 3.98 (3.09)
Puerto Rico -3.59 (2.06) -4.16 (1.99)* -4.15 (1.96)* -4.13 (1.96)*
Central America 1.42 (1.92) 1.71 (1.87) 1.59 (1.86) 1.50 (1.88)
South America -0.95 (1.50) -1.18 (1.52) -1.20 (1.52) -1.36 (1.52)
Dominican Republic -1.90 (2.00) -1.20 (2.18) -1.27 (2.18) -1.51 (2.15)
Mexico 1.30 (0.60)* 1.49 (0.63)* 1.47 (0.64)* 1.28 (0.64)*
Cuba -0.46 (2.02) -1.07 (1.92) -1.05 (1.96) -1.35 (1.94)
East Asia 0.13 (0.95) 0.23 (0.93) 0.26 (0.94) 0.17 (0.94)
Vietnam/Thailand/Cambodia/Laos -2.41 (1.62) -2.94 (1.88) -2.94 (1.86) -2.86 (1.89)
Other South East Asia -1.82 (0.84)* -1.53 (0.86) -1.47 (0.87) -1.43 (0.86)
India -1.07 (1.02) -1.34 (0.99) -1.23 (1.02) -1.19 (1.00)
South-Central/West Asia 1.59 (1.79) 0.78 (1.78) 0.79 (1.78) 0.76 (1.80)
Africa -3.53 (1.46)* -4.24 (1.54)** -4.17 (1.56)** -4.00 (1.56)**
Oceania (excluding Australia) -0.49 (0.84) -0.39 (0.82) -0.41 (0.82) -0.44 (0.81)
Second generation
North America -1.59 (1.25) -1.41 (1.23) -1.46 (1.23) -1.57 (1.22)
Europe (including Russia) 0.04 (0.37) -0.05 (0.37) -0.08 (0.37) -0.12 (0.37)
Caribbean -2.05 (0.80)** -1.60 (0.78)* -1.62 (0.78)* -1.72 (0.79)*
Puerto Rico -0.25 (0.92) -0.21 (0.94) -0.21 (0.94) -0.33 (0.94)
Central America -0.32 (0.52) -0.18 (0.53) -0.26 (0.53) -0.34 (0.54)
South America -0.52 (0.53) -0.49 (0.53) -0.52 (0.53) -0.59 (0.53)
Dominican Republic -1.17 (0.75) -0.75 (0.77) -0.73 (0.76) -0.83 (0.76)
Mexico 1.43 (0.26)*** 1.50 (0.32)*** 1.46 (0.32)*** 1.29 (0.34)***
Cuba 0.76 (0.82) 0.55 (0.80) 0.45 (0.80) 0.24 (0.83)
East Asia -0.78 (0.41) -0.75 (0.42) -0.76 (0.42) -0.79 (0.42)
Vietnam/Thailand/Cambodia/Laos -0.20 (0.54) -0.11 (0.56) -0.11 (0.57) -0.12 (0.56)
Other South East Asia -0.81 (0.35)* -0.68 (0.36) -0.61 (0.36) -0.64 (0.36)
India -0.39 (0.53) -0.85 (0.55) -0.85 (0.55) -0.77 (0.55)
South-Central/West Asia 1.28 (0.72) 0.95 (0.73) 0.97 (0.73) 1.00 (0.73)
Africa -0.15 (0.73) -0.17 (0.72) -0.18 (0.73) -0.25 (0.73)
Oceania (excluding Australia) -1.83 (0.83)* -1.93 (0.82)* -1.90 (0.81)* -1.89 (0.81)*
Third+ generation
Black -1.68 (0.15)*** -1.68 (0.17)*** -1.64 (0.17)*** -1.60 (0.18)***
Hispanic -0.01 (0.18) 0.19 (0.19) 0.17 (0.19) 0.10 (0.19)
Asian -0.19 (0.28) 0.01 (0.30) 0.03 (0.30) 0.03 (0.30)
Other (including mixed race) -0.90 (0.25)*** -0.67 (0.25)** -0.69 (0.25)** -0.70 (0.25)**
Child Characteristics
Boy 0.45 (0.09)*** 0.47 (0.09)*** 0.47 (0.09)***
Child age in months -0.11 (0.01)*** -0.10 (0.01)*** -0.10 (0.01)***
Low birth weight (<2500 g) -0.22 (0.17) -0.22 (0.17) -0.21 (0.17)
Premature (>=2 weeks early) -0.10 (0.13) -0.11 (0.13) -0.09 (0.13)
Height 0.08 (0.03)** 0.08 (0.03)** 0.08 (0.03)**
Weight -0.01 (0.01) -0.01 (0.01) -0.01 (0.01)
Number of moves since birth 0.06 (0.04) 0.07 (0.04) 0.07 (0.04)
Center-based care before entering kindergarten -0.28 (0.10)** -0.29 (0.10)** -0.30 (0.10)**
Parent Characteristics
Mother’s age -0.01 (0.01) -0.01 (0.01) -0.01 (0.01)
Highest parental education
High school degree 0.17 (0.22) 0.14 (0.22) 0.15 (0.22)
Some college 0.51 (0.23)* 0.43 (0.23) 0.44 (0.23)
College and plus 0.71 (0.28)** 0.61 (0.28)* 0.62 (0.28)*
Mother married at birth 0.14 (0.13) 0.09 (0.13) 0.08 (0.14)
Mother currently works full-time 0.03 (0.10) 0.03 (0.10) 0.03 (0.10)
Family Characteristics
Number of persons age <18 in the household 0.12 (0.05)** 0.13 (0.05)** 0.13 (0.05)**
Social economic status (SES) 0.10 (0.10) 0.04 (0.10) 0.03 (0.10)
 Wei-Jui Han 313

Appendix Table 3. continued.

Home language is not English 0.30 (0.20) 0.31 (0.20) 0.23 (0.21)
Region of residence
Mid-West 0.66 (0.14)*** 0.68 (0.15)*** 0.66 (0.15)***
South 1.16 (0.14)*** 1.20 (0.14)*** 1.14 (0.15)***
West 0.34 (0.16)* 0.35 (0.16)* 0.21 (0.17)
Location of residency
Rural 0.32 (0.19) 0.34 (0.19) 0.38 (0.19)*
Small town -0.28 (0.20) -0.26 (0.20) -0.16 (0.21)
Large town 0.57 (0.28)* 0.60 (0.28)* 0.55 (0.29)
Mid-size suburban 0.66 (0.20)*** 0.67 (0.20)*** 0.76 (0.20)***
Large-size suburban 0.46 (0.14)*** 0.46 (0.14)*** 0.46 (0.15)***
Mid-size city 0.64 (0.15)*** 0.64 (0.15)*** 0.67 (0.16)***
Parental well-being and educational practice
Maternal depression -0.01 (0.01) -0.01 (0.01)
Parental educational expectations for child 0.06 (0.05) 0.06 (0.05)
Importance of child having skills by entrance to -0.08 (0.10) -0.06 (0.10)
kindergarten
Participating in school events 0.06 (0.04) 0.04 (0.04)
Difficulty in attending school events 0.03 (0.04) 0.03 (0.04)
Home Environment
Home learning activities 0.02 (0.03) 0.02 (0.03)
Attending extracurricular activities -0.00 (0.04) -0.00 (0.04)
Frequency of spanking child in past week -0.16 (0.05)*** -0.16 (0.05)***
Family routines 0.01 (0.11) -0.00 (0.11)
School and neighborhood characteristics
Student minority composition 0.19 (0.10)
Average student performance 0.30 (0.08)***
Number of years served as principal -0.00 (0.01)
School communication about the child 0.09 (0.08)
Teachers’ efforts to ease the transition into kindergarten for -0.01 (0.07)
children
Teacher-reported parental involvement in school activities 0.09 (0.07)
School neighborhood quality -0.06 (0.12)
Observers’ overall rating of school safety 0.18 (0.08)*
Residential neighborhood quality 0.02 (0.08)
Earlier measure
Average math score of fall and spring kindergarten 0.78 (0.01)*** 0.78 (0.01)*** 0.78 (0.01)*** 0.78 (0.01)***
Observations 16161 16161 16161 16161
Adjusted R-squared 0.63 0.63 0.64 0.64
Note. Reference group is Non-Hispanic whites. Standard errors shown in parentheses are corrected for school clustering using Huber-White
methods. Details about covariates are presented in Appendix Table 1.
* p < .05. ** p < .01. *** p < .001.

from Europe had a significantly faster learning pace (ECLS-K), this paper examined the developmental
compared to their second-generation counterparts and to experiences of young children in immigrant families with
non-Hispanic white children. And second-generation a rich set of factors that have been theoretically and
children from the Caribbean had a significantly slower empirically related to individual, parental, family, home
learning pace compared to non-Hispanic white children, environment, and school and neighborhood
while first-generation children from Africa had a characteristics. In addition, a change model was
significantly slower learning pace compared to their examined to see whether the academic learning paces
second-generation counterparts and to non-Hispanic during kindergarten and first grade differed by generation
white children. status by country of origin.
Consistent with prior research (e.g., Duncan and
Brooks-Gunn, 1997; Shonkoff and Phillips, 2000;
DISCUSSION AND CONCLUSIONS Smolensky and Gootman, 2003), the analyses suggest
that child and family characteristics were the most
Taking advantage of a large-scale longitudinal data set important factors to these young children’s academic
314 Educ. Res. Rev.

achievements (effect from moderate to large before and disadvantaged position from the start.
small to moderate after controls), which is Moreover, home environments accounted partially for
understandable given children are heavily influenced by the lower reading scores for first-generation children from
their own characteristics and their family in the early Puerto Rico and Cuba compared to non-Hispanic white
years of life. To a lesser degree, the analyses also children. Specifically, although the former group lived
suggest that home, school, and neighborhood with higher parental educational expectations than non-
environments exerted some influences on these Hispanic white children, they tended to have parents who
children’s learning experiences. Analyses that were participated significantly less in school events, had their
conducted separately by racial/ethnic origin suggest that children attend fewer extracurricular activities, had more
(results not shown), in addition to the uniform influence of difficulty in attending school events, had less
child and family characteristics on both children of Latin communication with teachers, and provided fewer home
American and Asian origin, home and school learning activities. First-generation children from Puerto
environments seemed to matter more to the academic Rico also had mothers with higher levels of depression,
achievement of children of Latin American (i.e., while first-generation children from Cuba also had
particularly effects of maternal depression, parental mothers who spanked them more. It is worth noting that
educational expectations, participation in and difficulty all first- and second-generation children tended to live
attending school events, home learning activities, with higher parental educational expectations than non-
frequency of spanking, student body composition, Hispanic white children, but these expectations can only
average student performance, teachers’ efforts to ease partially account for the differences in children’s
the kindergarten transition, teacher-reported parental academic achievements, other child, family, and home
involvement, and school safety were significant at at least characteristics (e.g. SES and home language) matter,
p < .05) than of Asian children (i.e., effects of parental too.
educational expectations, importance of obtaining skills School and neighborhood environments were also
before kindergarten, number of years served as principal, important to some groups of Latin American children.
teachers’ efforts to ease the transition, and teacher- Compared to non-Hispanic white children, second-
reported parental involvement were significant at at least generation children from Mexico and first-generation
p < .05). The discussion below expands upon these children from Cuba had significantly lower reading
results and the mediating factors that were significant in scores, and second-generation children from Puerto Rico
the regression models by country of origin. and Central America and second-generation children
Child and family characteristics were important to the from Mexico had significantly lower math scores. These
academic differences between all first- and second- significant differences either disappeared (for second-
generation children and non-Hispanic white children. generation children from Mexico in reading and math) or
Latin American children in particular seemed to improve became less significant (for first-generation children from
through later generations, which may be largely due to Cuba on reading from 1% to 5%, and for second-
earlier generations' relatively disadvantageous family generation children from Puerto Rico and Central
socioeconomic status and to a lesser extent to home, America in math from 1% to 5%). Compared to non-
school, and neighborhood environments. Specifically, the Hispanic white children, these children tended to live in
mediating factors suggest that this difference was due in less safe neighborhoods (except for children from Central
part to lower maternal education, lower family SES, and America) and attend schools with high student minority
having a non-English language spoken in the home (and compositions, generally poor student academic
lower attendance in center-based care before performance (except for first-generation children from
kindergarten for first-generation children from the Cuba), and poor school safety. It is worth noting that
Dominican Republic and Cuba). Although previous almost all first- and second-generation children from Latin
studies have shown that families who moved more American regions tended to live in these school and
frequently may be less likely to establish and accumulate neighborhood conditions.
social capital to benefit children’s cognitive and social Comparatively, children of Asian origin performed
development (Coleman, 1988; Hagan, MacMillan, and significantly better than non-Hispanic white children on
Wheaton, 1996), the present math-score estimates for both reading and math skills (except for children from
first- and second-generation children of frequent movers other Southeast Asia who had significantly lower math
were positive, which may reflect parent(s)’ investment in scores). Even after controlling for all three sets of
their children by continually moving to better areas. First- mediators, children of Asian origin still had significantly
generation children from Mexico and Cuba also had higher scores (although the differences became smaller).
significantly higher percentages of low birth weight In some cases (particularly for children from
compared to non-Hispanic white children. On the whole, Vietnam/Thailand/Cambodia/Laos), the differences
children from Mexico tended to have the most became larger with each set of mediators. On one hand,
challenging family backgrounds, which put them in a compared to non-Hispanic white children, children of
 Wei-Jui Han 315

Asian origin tended to move more frequently, to have by first-grade, but also learned the skills faster during
mothers with higher education and more likely to be kindergarten and first grade, thus widening the initial
married at the child’s birth (except for children from performance gap. Similarly, children from Europe, North
Vietnam/Thailand/Cambodia/Laos and other Southeast America, and Africa learned reading and math skills at
Asia), to have families with higher SES (except for faster paces, which may widen the gaps in academic
children from Vietnam/Thailand/Cambodia/Laos and achievements over time.
other Southeast Asia), to have parents with higher All in all, the results in this paper indicate that the
educational expectations and who valued obtaining skills weaker academic achievements by Latin American
before entering kindergarten, to have higher family children could largely be attributed to their less
routines, and to attend schools with higher average advantageous family socio-demographic backgrounds,
student academic performance (for children from East and to a lesser extent their less stimulating home
Asia and India). On the other hand, they tended to be environments and worse school and neighborhood
less likely to speak English at home, to be spanked more environments. In other words, if all children of Latin
(for children from East Asia, other Southeast Asia, and American origins had family backgrounds and quality of
India), to have parents who had more difficulty attending schools similar to non-Hispanic white children, the former
school events and communicated less with teachers, and would have had similar, and sometimes even better,
to attend schools with a higher student minority academic achievements than the latter. In contrast
composition. Children from children of Asian origin tended to have advantageous
Vietnam/Thailand/Cambodia/Laos and other Southeast family backgrounds that contributed to their better
Asia areas in particular had less advantageous academic performance compared to non-Hispanic white
characteristics. For example, compared to non-Hispanic children. The former also had more stimulating home
white children, they were less likely to attend center- environments and attended higher quality schools
based care before kindergarten, more likely to have compared to the latter, and even after controlling for
parents who provided fewer home learning activities, and these advantages, they still performed better
to have poor school and and neighborhood safety. academically. The higher parental educational
Children from North America, Europe, and Africa expectations coupled with more stimulating home
tended to have more advantageous child and family learning activities, attending more extracurricular
characteristics (higher maternal educational attainment, activities, higher family routines (despite participating less
more pre-kindergarten center-based care, and higher in school events, having more difficulty in attending them,
family SES). The change from non-significant differences and communicating less with teachers), and
to significant and negative ones for children from North advantageous school environments have demonstrated
America compared to non-Hispanic white children may the strong emphasis parents of Asian origin have on
further be explained by the advantaged home and school children’s academic achievements (Zhou and Bankston,
and neighborhood environments experienced by the 1994, 1998). These results suggest that the degree of
former, who tended to have parents with higher influence of each mediator is different for children of
expectations and family routine, provided more home different cultures. For example, while parents in Latin
learning activities, had their children attend more American or Asian societies tend to use more physical
extracurricular activities (except for first-generation discipline than western parents (Baldwin, Baldwin, and
children), have schools with higher average student Cole, 1990; Portes, Dunham, and Williams, 1986),
academic performance and parental involvement, and children in these families do not tend to have more
safer residential neighborhoods (except for second- behavioral problems later on (McLoyd and Smith, 2002).
generation children). Children from Africa tended to have Likewise, parents' participation in school activities is
similar advantageous school and neighborhood positively associated with academic achievements, but
environments, which most likely aided them in only for native-born non-Hispanic white children; children
outperforming non-Hispanic white children. in immigrant families (such as the children of Asian origin
It is also important to note that although children of in this study) might achieve similarly even without the
immigrants tended to have either higher or lower scores same level of parent involvement, perhaps partially
by first-grade, they usually learned skills at faster paces, because of the higher parental educational expectations
which widened (for children of Asian origin) or narrowed that have been noted in prior ethnographic studies.
(for children of Latin America) the initial gaps in academic Although a rich set of factors related to child, parental,
achievements. For example, although children from family, home environment, and school and neighborhood
Mexico had significantly lower math scores than non- characteristics were considered in this study, several
Hispanic white children by first grade, the former limitations should be considered when interpreting the
narrowed the gap over time. Children from East Asia and results. First, although this analysis has distinguished
other Southeast Asian countries not only had significantly among generations for many countries of origin, ethnic
higher reading scores than non-Hispanic white children groups themselves are replete with individual differences
316 Educ. Res. Rev.

and distinct historical backgrounds, thus making it difficult allows us to better understand some of the factors
to define a culture (Bean et al., 1997; Bean and Stevens, important to the developmental experiences of children in
2003; Portes and Rumbaut, 1996; Portes and Zhou, immigrant families. The heterogeneity of these children
1993). Second, the results may be biased due to the as described above further highlights the need for more
exclusion of some children who did not complete a direct research consisting of intricate analyses of the ways that
assessment (especially for children from Mexico and to a familial and social factors intersect in shaping families’,
lesser extent from Asian countries). Such limitations also and especially, children's experiences. Specifically, future
reflect the need to include diverse cultures and research should attempt to account for as many aspects
languages in future data collection and measurement of children's culturally shaped social and educational
development. Third, given the context of the research environments as possible because they are essential
questions under examination, the importance of home pieces to our understanding of children's cognitive and
language on children’s academic learning experiences emotional development. In addition, while this paper has
and in turn on their academic achievements should be attempted to provide a general overview of the
acknowledged. Although the current analyses controlled developmental experiences of young children in many
for this factor and the results indicate that home language immigrant groups, future research focusing on the
exerts strong effects in all models – with a significant similarities and differences within particular immigrant
negative effect on initial academic achievements and a and cultural groups is needed to provide us with a more
significant positive effect on the changes in academic in-depth understanding of young children in immigrant
achievements over time – finer-grained analyses are families. Without a detailed understanding of these
needed to explain these associations. Fourth, obtaining complex interactions, we cannot construct informed and
accurate measurements of socioeconomic status for effective policies to address the needs of immigrant
immigrant families is far more complicated than simply families today.
identifying parental education, occupation, or family
income (Fuligni and Yoshikawa, 2004), especially
because it has been observed that parents in immigrant ACKNOWLEDGEMENT
families tend to have lower-level occupations in the U.S.
than they did in their native countries (Portes and The Author would like to thank Katherine Magnuson,
Rumbaut, 1996). In addition, many immigrant families Ruby Takanish, Jane Waldfogel, and participants at “The
remit a large proportion of household income to relatives Next Generation: Immigrant Youth and Families in
in their country of origin (Schiller, 1999). Thus, this Comparative Perspective” conference for valuable
study's measure of family socioeconomic status may be discussions, and Rocky Citro for wonderful editorial
invalid and biased for many immigrants. Fifth, despite assistance. In addition, the author gratefully
many covariates, the author was still unable to control for acknowledges support from the Foundation for Child
some characteristics that might help explain these Development Young Scholars Program.
associations. For instance, information on children's
interactions with their teachers and peers, as well as the
attitudes of teachers toward children from different REFERENCES
cultural backgrounds, may have proved fruitful given Baldwin AL, Baldwin C, Cole RE (1990). Stress-resistant families and
previous findings on the importance of these variables for stress-resistant children. In: JE Rolf, AS Masten (Eds.), Risk and
children’s academic achievements (e.g., Conchas, 2001; protective factors in the development of psychopathology. New York:
NICHD ECCRN, 2002b; Suárez-Orozco and Suárez- Cambridge University Press. pp. 257-280.
Bean FD, Van Hook J, Glick JE (1997). Country-of-origin, type of public
Orozco, 2001). Additionally, despite including some
assistance, and patterns of welfare recipiency among U.S.
school-level variables (e.g., student minority composition immigrants and natives. Social Science Quarterly, 78(2), 432-451.
and average student performance) that have been found Bean FD, Stevens G (Eds.). (2003). America’s newcomers and the
to be associated with inadequate educational resources dynamics of diversity. New York: Russell Sage Foundation.
Belsky J (2001). Developmental risk (still) associated with early child
and academic failure (Wang and Gordon, 1994) to proxy
care.” J. Child Psychol. Psychiatry, 42, 845-860.
the segregation and oppression, the data at hand did not Benard B (1991). Fostering resiliency in kids: Protective factors in the
allow this study to fully consider factors unique and family, school, and community. Portland, OR: Northwest Regional
important to children's daily experiences in immigrant Educational Laboratory.
Board on Children and Families, Commission on Behavioral and Social
families and likely their academic development, such as Sciences and Education, National Research Council, Institute of
racial discrimination and oppression (García Coll, 1996, Medicine. (1995). Immigrant children and their families: Issues for
2004). Future studies should utilize both qualitative and research and policy. The Future of Children, 5(2), 72-89.
quantitative methods to understand the broad patterns as Booth A, Crouter AC, Landale N (eds.). (1997). Immigration and the
family: Research and policy on U.S. immigrants. New Jersey:
well as the rich and intricate diversity of this subject area.
Lawrence Erlbaum Associates.
The results from this paper also show that examining a Borman G, Overman LT (2004). Academic resilience in mathematics
variety of individual, family, and external environments among poor and minority students. The Elementary Sch. J. 104(3),
 Wei-Jui Han 317

177-195. immigrant and involuntary minorities. New York: Garland. pp. 169-
Bornstein MH, Gist NF, Hahn CS, Haynes OM, Voigt MD (2001). Long- 204.
term cumulative effects of daycare experience on children’s mental Gibson MA, Bhachu PK (1991). The dynamics of educational decision
and socioemotional development. National Institute of Child Health making: A comparative study of Sikhs in Britain and the United
and Human Development, mimeo. States. In: MA Gibson, JU Obgu (eds.), Minority status and schooling:
Bradley RH (1995). Environment and parenting. In M.H. Bornstein A comparative study of immigrant and involuntary minorities. New
(ed.), Handbook of Parenting: Volume 2, Biology and Ecology of York: Garland. pp. 63-96.
Parenting New York: Erlbaum. pp. 235-261. Griffith J (2000). School climate as group evaluation and group
Bradley RH, Caldwell BM, Rock SL, Ramey CT, Barnard KE, Gary C, consensus: Student and parent perceptions of the elementary school
Hammond MA, Mitchell S, Gottfried AW, Siegel L, Johnson D (1989). environment. The Elementary Sch. J. 101 (1), 35-61.
Home environment and cognitive development in the first 3 years of Griffith J (2003). Schools as organizational models: Implications for
life: A collaborative study involving six sites and three ethnic groups examining school effectiveness. The Elementary Sch. J. 104(1), 29-
in North America. Developmental Psychology, 25(2), 217-235. 47.
Bronfenbrenner U (1979). The ecology of human development: Hagan J, MacMillan R, Wheaton B (1996). New id in town: Social
Experiments by nature and design. Cambridge, MA: Harvard capital and the life course effects of family migration on children.
University Press. American Sociological Review, 61, 368-385.
Bronfenbrenner U (1986). Ecology of the family as a context for human Herderson N, Milstein MM (1996). Resiliency in schools: Making it
development: Research perspectives. Developmental Psychology, happen for students and educators. Thousand Oaks, CA: Corwin.
22, 723-742. Hernandez DJ (ed.). (1999). Children of immigrants. National Research
Bryk AS, Lee VE, Holland PB (1993). Catholic schools and the common Council Committee on the Health and Adjustment of immigrant
good. Cambridge, MA: Harvard University Press. Children and Families, Washington, DC: National Academy Press.
Caplan N, Choy MH, Whitmore JK (1991). Children of the boat people: Huff R, Trump K (1996). Youth violence and gangs: School safety
A study of educational success. Ann Arbor, MI: University of initiatives in urban and suburban school districts. Educational and
Michigan Press. Urban Society, 28, 492-503.
Chao RK (1994). Beyond parental control and authoritarian parenting Johnson DJ, Jaeger E, Randolph SM, Cauce AM, Ward J, National
style: Understanding Chinese parenting through the cultural notion of Institute of Child Health and Human Development Early Child Care
training. Child Development, 65, 1111-1119. Research Network. (2003). Studying the effects of early child care
Chao RK (2001). Extending research on the consequences of parenting experiences on the development of children of color in the United
style for Chinese Americans and European Americans. Child States: Toward a more inclusive research agenda. Child
Development, 72, 1832-1843. Development, 74(5), 1227-1244.
Cohen J (1977). Statistical power analysis for the behavioral sciences Johnson MK, Crosnoe R, Elder GH, Jr (2001). Student attachment and
(revised edition). New York: Academic Press. academic engagement: The role of ethnicity. Sociology of Education,
Coleman J (1988). Social capital in the creation of human capital. 74, 318-340.
Am(r). J. Sociol., 94(supp), S95-S120. Kao G, Tienda M (1995). Optimism and achievement: The educational
Conchas GQ (2001). Structuring failure and success: Understanding the performance of immigrant youth. Social Science Quarterly, 76, 1-19.
variability in Latino school engagement. Harvard Educational Lamb ME (1998). Nonparental child care: Context, quality, correlates,
Review, 71, 475-504. and consequences. In: W Damon (general ed.) and IE Sigel, KA
Conger RD, Conger KJ, Elder GH, Jr, Lorenz FO, Simons RL, Renninger (volume eds.) Handbook of child psychology, 5th edition,
Whitbeck LB (1992). A family process model of economic hardship Volume 4: Child psychology in practice. New York: John Wiley and
and adjustment of early adolescent boys. Child Development, 63, Sons, Inc. pp. 73-133.
526-541. Lee VE, Burkham D (2002). Inequality at the starting gate: Social
Crosnoe R (2005). Double disadvantage or signs of resilience? The background differences in achievement as children begin school.
elementary school contexts of children from Mexican immigrant Washington, DC: Economic Policy Institute.
families. Am(r). Edu. Res. J. 42(2), 269-303. Lee VE, Winfield LF, Wilson TC (1991). Academic behaviors among
Denton K, West J (2002). Children’s reading and mathematics high-achieving African American students. Education and Urban
achievement in kindergarten and first grade. Washington, DC: U.S. Society, 24, 65-86.
Department of Education. Louie V (2001). Parents’ aspirations and investment: The role of social
Duncan G, Brooks-Gunn J (eds.) (1997). Consequences of growing up class in the educational experiences of 1.5- and second-generation
poor. New York: Russell Sage Foundation. Chinese Americans. Harvard Educational Review, 71, 438-474.
Elder GH, Conger RD, Foster EM, Ardelt M (1992). Families under Magnuson K, Meyers MK, Ruhm C, Waldfogel J (2004). Inequality in
economic pressure. J. Family Issues, 13, 5-37. preschool education and school readiness. Am(r) Edu. Res. J. 41(1),
Fuligni AJ (1997). The academic achievement of adolescents from 115-157.
immigrant families: The roles of family background, attitudes, and Mahoney JL (2000). School extracurricular activity participation as a
behavior. Child Development, 68, 351-363. moderator in the development of antisocial patterns. Child
Fuligni AJ, Tseng V, Lam M (1999). Attitudes toward family obligations Development, 71, 502-516.
among American adolescents with Asian, Latin American, and Masten AS (1994). Resilience in individual development: Successful
European backgrounds. Child Development, 70, 1030-1044. adaptation despite risk and adversity. In: MC Wang, FW Gordon
Fuligni AJ, Yoshikawa H (2004). Investments in children among (Eds.), Educational resilience in inner-city America: Challenges and
immigrant families. In: A Kalil, T DeLeire (eds.), Family Investment in prospects. Hillsdale, NJ: Erlbaum. pp. 3-25.
Children’s Potential. Mahwah, NJ: Lawrence Erlbaum Associates, Matute-Bianchi ME (1986). Ethnic identities and patterns of school
Inc. pp. 139-162. success and failure among Mexican-descent and Japanese-
García Coll C, Lamberty G, Jenkins R, McAdoo HP, Crnic K, Wasik BH, American students in a California high school: An ethnographic
García HV (1996). An integrative model for the study of analysis. Am(r). J. Edu. 95, 233-255.
developmental competencies in minority children. Child McLoyd VC (1990). The impact of economic hardship on black families
Development, 67(5), 1891-1914. and children: Psychological distress, parenting, and socioemotional
García Coll C, Szalacha LA (2004). The multiple contexts of middle development. Child Development, 61, 311-346.
childhood. The Future of Children, 14(2), 81-97. McLoyd VC, Smith J (2002). Physical discipline and behavioral
Gibson MA (1991). Ethnicity, gender, and social class: The school problems in African American, European American, and Hispanic
adaptation patterns of West Indian youths. In: MA Gibson, JU Obgu children: Emotional support as a moderator. J. Marriage and Family,
(eds.), Minority status and schooling: A comparative study of 64, 40-53.
318 Educ. Res. Rev.

McLoyd VC, Wilson L (1991). The strain of living poor: Parenting, social and western high school students. J. Res. Adolescence, 1, 135-154.
support, and child mental health. In: AC Huston (ed.), Children in Rumbaut RG (1994). The crucible within: Ethnic identity, self-esteem
poverty. Cambridge: Cambridge University Press. pp. 105-135. and segmented assimilation among children of immigrants.
McNeal RB, Jr (1995). Extracurricular activities and high school International Migration Review, 28, 748-794.
dropouts. Sociology of Education, 68, 62-80. Rumbaut RG (1995). The new Californians: Comparative research
McNeal R (1997). High school dropouts: A closer examination of school findings on the educational progress of immigrant children. In: RG
effects. Social Science Quarterly, 78, 209-222. Rumbaut, W Cornelius (eds.), California’s immigrant children.
Moody J (2001). Race, school integration, and friendship segregation in LaJolla, CA: Center for U.S.-Mexican Studies. pp. 17-69.
America. Am(r). J. Sociol. 107(3), 679-716. Rumbaut RG (1997). Ties that bind: Immigration and immigrant families
Moore KA, Halle T (1997). Positive youth development. Washington, in the United States. In: A Booth, AC Crouter, N Landale (eds.),
D.C. Child Trends, Inc. Immigration and the family: Research and policy on U.S. immigrants.
National Center for Education Statistics [NCES] (2002). ECLS-K New Jersey: Lawrence Erlbaum Associates. pp. 3-46.
longitudinal kindergarten-first grade public-use data files and Sampson R, Groves W (1989). Community structure and crime: Testing
electronic code book. Retrieved February 15, 2003 from social disorganization theory. American Journal of Sociology, 94,
http://nces.ed.gov/ecls/KinderInstruments.asp 774-802.
NICHD Early Child Care Research Network (1999). Child care and Schiller NG (1999). Transmigrants and nation-states: Something old
mother-child interaction in the first 3 years of life. Developmental and something new in the U.S. immigrant experience. In: C
Psychology 35(6): 1399-1413. Hirschman, P Kasinitz, J DeWind (eds.), The Handbook of
NICHD Early Child Care Research Network (2000). The relation of child International Migration. New York: Russell Sage Foundation.
care to cognitive and language development. Child Development, Shonkoff JP, Phillips DA (eds) (2000). From neurons to neighborhoods:
71(4), 960-980. The science of early childhood development. Washington, DC:
NICHD Early Child Care Research Network (2002a). Early child care National Academy Press.
and children’s development prior to school entry: Results from the Siantz dL, ML (1997). Factors that impact developmental outcomes of
NICHD Study of Early Child Care. Am(r) Edu. Res. J. 39, 133-164. immigrant children. In: A Booth, AC Crouter, N Landale (Eds.),
NICHD Early Child Care Research Network (2002b). The relation of first Immigration and the family: Research and policy on U.S. immigrants.
grade classroom environment to structural classroom features, Mahwah, New Jersey: Lawrence Erlbaum Associates, Publishers. pp.
teacher, and student behaviors. The Elementary Sch. J., 102, 367- 149-161.
387. Smolensky E, Gootman J (eds.) (2003). Working families and growing
Nord CW, Griffin JA (1999). Educational profile of 3- to 8-year-old kids: Caring for children and adolescents. Washington, DC: The
children of immigrants. In: DJ Hernadez (ed.) National Research National Academies Press.
Council Committee on the Health and Adjustment of immigrant Suárez-Orozco C, Suárez-Orozco MM (1995). Transformations:
Children and Families, Children of immigrants: Health, adjustment, Immigration, family life, and achievement motivation among Latino
and public assistance . Washington, DC: National Academy Press. adolescents. Stanford CA: Stanford University Press.
pp. 348-409. Suárez-Orozco C, Suárez-Orozco MM (2001). Children of immigration.
Ogbu JU (1978). Minority education and caste: The American system in Cambridge, MA: Harvard University Press.
cross-cultural perspective. New York: Academic Press. US Census Bureau (2004). The foreign-born population in the United
Ogbu JU (1981). Origins of human competence: A cultural ecological States: 2004. Retrieved on August 1, 2005 from
perspective. Child Development, 52, 413-429. http://www.census.gov/population/www/socdemo/foreign/ppl-
Ogbu, J. U. (1988). Cultural diversity and human development. In: D 176.html.
Slaughter (Ed.), Black children and poverty: A developmental Valencia R (2000). Inequalities and the schooling of minority students in
perspective. San Francisco: Jossey-Bass. pp. 11-28. Texas. Hispanic Journal of Behavioral Sciences, 22, 445-459.
Ogbu J (1991). Minority coping responses and school experience. J. Valenzuela A (1999). Subtractive schooling: U.S.-Mexican youth and
Psychohistory, 18(4), 433-456. the politics of caring. Albany, NY: State University of New York Press.
Ogbu JU, Simons HD (1998). Voluntary and involuntary minorities: A Wang MC, Gordon EW (Eds.). (1994). Educational resilience in inner-
cultural-ecological theory of school performance with some city America: Challenges and prospects. Hillsdale, NJ: Erlbaum.
implications for education. Anthropology and Education Quarterly, 29, Weinraub M, Jaeger E (1990). The timing of mothers’ return to the
155-188. workplace: Effects on the developing infant-mother relationship. In: J
Pérez Carre n G, Drake C, Barton AC (2005). The importance of Hyde, M Essex (eds.), Parental leave and child care. Philadelphia:
presence: Immigrant parents’ school engagement experiences. Temple University Press. pp. 307-322.
Am(r). Educ. Res. J. 42(3), 465-498. Wilson WJ (1987). The truly disadvantaged: The inner-city, the
Pessar PR (1995). A visa for a dream: Dominicans in the United States. underclass, and public policy. Chicago, IL: The University of Chicago
Boston, MA: Allyn and Bacon. Press.
Petterson SM, Albers AB (2001). Effects of poverty and maternal Zhou M, Bankston CL III (1994). Social capital and the adaptation of the
depression on early child development. Child Development, 72(6), second generation: The case of Vietnamese youth in New Orleans.
1794-1813. International Migration Review, 28(4), 821-845.
Portes A, MacLeod D (1999). Educating the second generation: Zhou M, Bankston CL III (1998). Growing up American: How
Determinants of academic achievement among children of Vietnamese children adapt to life in the United States. New York:
immigrants in the United States. J.Ethnic and Migration Studies, Russell Sage Foundation.
25(3), 373-396.
Portes A, Rumbaut RG (1996). Immigrant America: A portrait. (2nd ed).
Berkeley, CA: University of California Press.
Portes A, Zhou M (1993). The new second generation: Segmented
assimilation and its variants. Annals of the American Academy, 530,
74-96.
Protes PR, Dunham RM, Williams S (1986). Assessing child-rearing
style in ecological settings: Its relation to culture, social class, early
age intervention and scholastic achievement. Adolescence, 21, 723-
735.
Rosenthal DA, Feldman SS (1991). The influence of perceived family
and personal factors on self-reported school performance of Chinese