Academic Performance of Young

Original
XXX
Blackwell
Oxford,
International
IMRE
©
0197-9183
2007 byArticles
UKPublishing
the
International
Academic Migration
Center for
Performance
Migration Ltd
of Review
Migration
Review
Children of Studies of New York. All rights reserved
Immigrants

Children in Immigrant Families:

The Significance of Race, Ethnicity,
and National Origins 1
Jennifer E. Glick
Arizona State University

Bryndl Hohmann-Marriott
Pennsylvania State University

Children of immigrants come from diverse backgrounds and enter school

with different family migration experiences and resources. This paper
addresses two basic questions: (1) to what extent does generation status
exert an independent effect on early school performance net of race/
panethnicity, language proficiency, and the family resources available to
children as they enter formal schooling? and (2) to what extent do these
broad conceptualizations of children in immigrant families mask variation
by national origins? We take advantage of longitudinal data on a kinder-
garten cohort from the Early Childhood Longitudinal Study to examine
children from diverse backgrounds. Considerable variation in academic
performance persists across racial/panethnic groups as well as by country-
of-origin background and linguistic ability even when adjusting for family
background, resources, and previous academic performance. We find some
intriguing evidence of early “segmentation” among children from various
groups, suggesting some convergence within race and ethnicity for some
children. However, this conclusion should not be overstated, because the
results also point to the great diversity by national origins that are masked
by reliance on racial/panethnic groupings.

INTRODUCTION

The growth of immigration over the past three decades represents the second-
largest flow of international migrants to the United States following the
massive European migration of the early 20th century. These recent arrivals
come from diverse backgrounds representing multiple countries, linguistic
origins, and ethnic groups. Nowhere is the impact of these traits more apparent
1Thisresearch is supported by a grant from the National Institute of Child Health and Human
Development, number RO3 HD044006.

© 2007 by the Center for Migration Studies of New York. All rights reserved.
DOI: 10.1111/j.1747-7379.2007.00072.x

IMR Volume 41 Number 2 (Summer 2007):371–402 371

372 I M R

than among the youth in the United States. Today one in five children in
primary and secondary school have at least one foreign-born parent ( Jamieson,
Curry, and Martinez, 2001). The school is the first major formal organization
the child encounters and serves as a major conduit in the U.S. stratification
system. Thus, one of the most revealing settings in which to view the relative
success or failure of newcomers and their children is the school (Entwisle and
Alexander, 1993).
Two difficulties that emerge when examining the relative progress of
children in immigrant families are determining what characteristics are
inherent to their status and determining to whom these children should be
compared. We know that children of immigrants today are likely to be members
of ethnic minority groups in the United States, to come from homes in which
a non-English language is spoken, and to come from a wide variety of national
origins. All of these characteristics are related to generation status and all have
been associated with differential success and opportunity in the United States.
When it comes to determining the progress of these children in U.S. schools,
it is not always clear which is the salient reference group for determining their
relative achievement.
One view is that that these children are appropriately compared to the
majority (i.e., non-Hispanic White) children so that their adaptation to the
American mainstream (e.g., their “assimilation”) can be assessed. This classic,
linear view suggests children from immigrant families may start out behind
their peers with U.S.-born parents but that over time, through the socialization
they receive in American schools, these children of immigrants catch up. In the
case of the “model minority” stereotype applied to recent Asian migrants, this
story may become even more dramatic if children of immigrants surpass their
native peers (Kao, 1995).
However, another view points to the inherent disadvantages associated
with position in the American racial/ethnic hierarchy that may inhibit progress
through formal institutions. In this case, those children of new immigrants
who are from historically disadvantaged groups in the United States may need
to overcome structural obstacles in order to have the same academic success as
their U.S.-born peers. In this “segmented” or “divergent” assimilation perspec-
tive, children of immigrants are likely to be affected by the context of reception
of their immigrant parents (Fernandez-Kelly and Schauffler, 1994). In this
case, one might expect that immigrant adaptation would vary according to
perceived position in the ethnic hierarchy of the United States (Portes and
Zhou, 1993; Fernandez-Kelly and Schauffler, 1994). Racial identity and
panethnic group membership in the United States could play an important
 A P  C  I 373

role in any generation status differences in educational outcomes. Thus, it may

be most appropriate to examine the extent to which children in immigrant
families resemble their co-ethnic or co-racial peers from non-immigrant
families, who, presumably, face similar structural obstacles in the United States.
Deciding to whom children of immigrants should be compared is also
complicated by their diverse origins, which may or may not map closely with U.S.
racial or ethnic labels. Researchers have repeatedly found important specific
country-of-origin differences that are masked by the broad panethnic identity
captured in survey data in the United States (Portes and Rumbaut, 2001). If
there are important country-of-origin differences within racial/ethnic group-
ings, specifying national origins may go further toward explaining nativity
differences in outcomes than merely examining race/ethnicity. Thus, the
academic trajectories of children from some national origins may converge
with minority racial/ethnic groups while other children from similar racial or
panethnic origins but different national origins maintain trajectories more sim-
ilar to their majority peers.
Both the straight-line and segmented assimilation frameworks are
concerned with proximity to the migration experience whether measured in
terms of generational progression or in terms of individual duration in the
receiving context and orientation to the sending community and culture.
However, when using the theoretical framework of assimilation (whether
straight-line or segmented) to study children, it is the family’s migration status
and orientations that are likely to be important. Immigrant families (raising
both first- and second-generation children) may provide resources to children
that buffer the effects of structural barriers to success in the receiving context.
The primary goals of this paper are (1) to investigate the effects of
generation status and family experience in the United States on early school
performance net of family background; and (2) to investigate the relative
performance of children by national origin when compared with children of
U.S.-born parents in different racial/ethnic groups. We pay particular attention
to the role of English proficiency, familial resources, and family-school connec-
tivity in accounting for variation across groups.

BACKGROUND

Previous research on high-school students reveals differences in school

performance by race, ethnicity, and immigrant status (e.g., Kao and Tienda,
1995; White and Kaufman, 1996; Glick and White, 2003). Results of analyses
with the High School and Beyond Survey, National Educational Longitudinal
374 I M R

Study, Adolescent Health Survey, and Children of Immigrants Longitudinal

Survey have all demonstrated important nativity, racial/ethnic, and country-
of-origin differences in academic performance and educational attainment
(e.g., Duran and Weffer, 1992; Kao, 1995; Portes and McLeod, 1996; Driscoll,
1999; Glick and White, 2003). Some studies point to downward trajectories
in school, particularly for Mexican and other Hispanic youth (Lopez and
Stanton-Salazar, 2001; Portes and Rumbaut, 2001: ch. 9). Others point to
greater educational attainment by immigrant youth when compared to youth
in the third and higher generations (Glick and White, 2004).
Fewer large-scale studies have followed children through the transition
to school. Thus, while we are gaining knowledge about the educational
attainment and labor market outcomes for new adult arrivals and adolescents,
we have less understanding of the progress made by very young children in
immigrant families and the influence of family resources on these children’s
outcomes. This paper begins this process by focusing specifically on the early
school performance of the children of immigrants and the relative importance
of generation status, racial identification, and national origins in determining
outcomes. Here we review the literature addressing these multiple factors
associated with being a child of immigrants.

Generation Status and Family Experience in the United States

The term “first generation” is broadly applied to all born outside of the United
States. Those who arrive as very young children, however, face different
adaptation processes from those who arrive as adolescents or young adults.
Indeed, earlier studies with adolescents reveal that those who arrived in the
United States as very young children (i.e., before school age or in the first year
or two of formal schooling; sometimes referred to as the 1.5 generation)
perform better on academic tests than those who arrived after schooling
was well under way (Glick and White, 2003). Both the children of this 1.5
generation and second generation have immigrant parents. These children,
therefore, all share the unique position of being socialized in the United States,
a social and cultural context potentially quite different from that faced by their
parents when they were children. Children in the 1.5 and second generations
may also share access to immigrant communities and resources in the United
States that promote achievement in school, resources that might not be
available to their third-or-higher-generation counterparts (Kao and Tienda,
1995). Further, many children with limited English proficiency in school today
are not immigrants themselves but the U.S.-born children of immigrant
 A P  C  I 375

parents. This study focuses on 1.5-generation children and their second-

generation peers. They are compared to their third-or-higher-generation (i.e.,
U.S.-born children of U.S.-born parents) counterparts.
While the children of the 1.5 and second generations share socialization
in the United States, we may find some variation between them due to differences
in families’ experience in the United States. Although the 1.5 and second
generation both have foreign-born parents, there may be considerable variation in
immigrant parents’ age at arrival. This could be a greater marker of experience
in the United States than the children’s own generation status. Few studies of
the educational trajectories of immigrant youth have disaggregated by the
parents’ age at arrival so our expectations are guided by previous work on
assimilation. If we adopt a straight-line assimilation approach, then families
with longer experience in the United States (i.e., parents who migrated as
children) should be associated with higher achievement among children than
families with less experience in the United States (Alba and Nee, 2003). Parents
who were educated and socialized in the United States may be better able to
guide their children through U.S. social institutions like schools. But, this is in
contrast to the finding that children of immigrant parents do well in school
because they benefit from their parents’ more recent arrival via their optimism
about opportunities in the United States (Kao and Tienda, 1995). The more
pessimistic view may expect children of immigrant parents who migrated
as children themselves to not do as well academically if, as the segmented
assimilation perspective implies, their parents are in a more disadvantaged
position in the United States. This also implies an interaction between parents’
age at arrival and race or ethnicity.

Racial Identification and Country of Origin

Children of immigrants today are inserted into the racial/ethnic hierarchy of

the United States and they are likely to be influenced by the context of
reception faced by their family based on this identification (Fernandez-Kelly
and Schauffler, 1994). In this view, the children of immigrants from those
groups that have been historically disadvantaged in the United States may face
different opportunities or barriers than other children from immigrant families
(Zhou, 1997a). Discrimination may spur some groups to seek success through
school (i.e., Sue and Okazaki, 1990) while others may deemphasize formal
education if these routes are perceived as blocked.
While studies of early education point to the importance of stratification
for shaping even the very initial academic trajectories of children in the United
376 I M R

States (Entwisle and Alexander, 1993; Lee and Burkam, 2002), we have less of
an understanding of when in the schooling process the children of immigrants
are most likely to be impacted by this disparity. Differences in the effect of
generation status on academic performance trajectories among students from
different racial/ethnic backgrounds imply important interactions between
race and generation status. “Regardless of their class origin or knowledge of
English, nonwhite immigrants face greater obstacles in gaining access to
the white middle-class mainstream” (Portes and Rumbaut, 2001:47).
According to this perspective, we should find interactions of race and
generation status whereby children of immigrants of some racial/panethnic
identities outperform their third-and-higher-generation peers while others lag
behind.
If insertion into a racial/ethnic hierarchy is the primary influence,
variation in academic performance should be greater by race/ethnicity than by
generation status or national origins. However, studies that compare children
across national origins suggest the “segmented” feature of adaptation may not
take place across entire panethnic groups but is rather evidenced among
particular national origin groups. National origins reflect very different
selectivity processes and the pre-migration characteristics of these groups may
explain differences in attainment among the children of immigrants (Feliciano,
2005). The story of immigrant academic achievement is often predicated on
the selection of immigrants with specific cultural values and beliefs that
enhance or detract from their achievement in the United States (Zhou, 1997b;
Goyette and Xie, 1999; Kim, 2002). Again, findings are mixed but for some
first- and particularly second-generation youth, performance is lower than those
in the third-or-higher-order generation even in the presence of controls for
differences in the stock of human capital (Portes and Rumbaut, 2001). Mexican,
Nicaraguan, Haitian, and Cambodian youth have all been found to exhibit
lower academic performance than their U.S.-born peers while Vietnamese- and
Chinese-origin youth have been cited as surpassing their third-or-higher-
generation peers (see Portes and Rumbaut, 2001; Suarez-Orzoco and Suarez-
Orzoco, 2001; Kim, 2002). In this case, a complete picture of variation in
academic achievement will only emerge once country of origin is taken into
account. We examine children’s outcomes, therefore, not only according to
their own racial/ethnic identity (as indicated by their parents) but also according
to their family’s national origins. It is quite likely that the racial/ethnic categories
so often used in the United States mask considerable variation by these national
origin groups. Further, we expect national origin differences may be larger than
differences by generation status.
 A P  C  I 377

Language Proficiency

One of the most frequently mentioned traits shared by many children of

immigrants is their lack of readiness to learn in English upon entering formal
schooling in the United States. A lack of English proficiency is cited as a
primary reason for poor school performance among many first- and second-
generation children (Rosenthal, Baker, and Ginsburg, 1983; Cosden et al.,
1995). While language proficiency and ethnic background are closely
intertwined with immigrant status, both traits have an independent effect on
educational attainment (Glick and White, 2003). Some of this may be due to
differences in the prevalence of non-English homes among youth of later
generations in some ethnic groups rather than others. Groups may vary in the
extent to which native language use persists across generations. Some groups
experience rapid turnover to English-only households by the second or at least
third generations while others see native languages persevere with and
sometimes without English use in the home (Alba et al., 2002). In other words,
English proficiency may be more directly linked to generation status for some
groups than others. Given its close link to academic performance, English
proficiency may capture most of the variation in academic performance
observed by generation status and possibly by country of origin.

Family Resources

Certainly most studies of status attainment in the United States point to the
importance of human capital and socioeconomic status in determining
subsequent outcomes across the life course. Few analyses of the progress of
immigrants’ children would fail to point to the importance of income and
parental education in enhancing children’s academic progress. Much of the
debate over the potential downward progress for these children has focused on
the extent to which impoverished immigrant families are unable to provide
needed resources that support children’s schooling.
To gain a better understanding of which characteristics of immigrant
families work as advantages for their children’s transition into formal schooling
in the United States, it may be useful to consider resources or investments that
fall under the broad construct of family capital, social capital, or even “familial
social capital” (Muller, 1993; Stanton-Salazar, 1997; Teachman, Paasch, and
Carver, 1997; Hao and Bonstead-Bruns, 1998). This network of relations pro-
vides a web of support for the student that may encourage and support achieve-
ment (Lareau, 1989). Nativity differences in access to social capital may be
378 I M R

essential to understanding the subsequent nativity differences in outcomes

over the early life course (Fuligni, 1997). In early childhood the family serves
as the primary social environment for children, and families may behave in
ways that promote academic success for their children. For example, reading
to preschool-age children, taking a child to the library, or engaging in other
academically oriented activities can support subsequent educational perform-
ance (Griffin and Morrison, 1997; Christian, Morrison, and Bryant, 1998;
Senechal and LeFevre, 2002; Sy and Schulenberg, 2005). Overall, such within-
family social capital may buffer students against attitudes or conditions
that discourage academic achievement (Hao and Bonstead-Bruns, 1998).
For young children, the school may serve as the primary extra-familial
context encountered, and the interactions of family and school are expected to
be another key resource for children. Some parents are better able to convert
their involvement to positive outcomes for their children. For example, parents
of lower-class backgrounds are not perceived as having the same level of interest
or understanding by teachers as parents whose backgrounds more closely
resemble the teacher’s background (Lareau, 1989). We address the possibility
that immigrant parents are less likely to become directly involved through contact
with the school but may still provide additional educational opportunities to
children beyond regular schooling. Enrollment in additional academic, artistic,
or culturally relevant programs may all provide an environment supportive of
academic performance and school progress. We expect this involvement
outside of school will be more common and involvement with the school to
be less common for children of immigrants when compared to their third-
and-higher-generation counterparts. But both types of involvement are expected
to enhance children’s academic achievement.

THE CURRENT STUDY

Academic performance in the early years of school forms a strong basis for
predicting later achievement (Entwisle and Alexander, 1993). Mastery of early
academic skills is an important predictor of ultimate educational attainment.
In our analyses, we consider parental age at arrival, national origins, family
activities, and resources provided to children outside of school as well as family
interactions with the schools as determinants of subsequent academic
achievement. If the educational performance of the children of immigrants was
influenced solely by the racial/ethnic hierarchy in which they are inserted in the
United States, then accounting for the racial/ethnic composition of these
children would be sufficient. Alternatively, if immigrant parents retain an
 A P  C  I 379

optimistic outlook while their children face blocked opportunities within

formal social institutions such as schools, the interaction of race and generation
status may be important for explaining actual academic performance. Finally,
because the selection of immigrants varies by national origins, we may observe
greater variation in school performance among children when we examine
national origins than when we rely solely on measures of generation status or
language proficiency to capture the “immigration” effect.
The analyses in this paper rely on new national-level data that are quickly
becoming a primary source of information on young children’s academic tra-
jectories for researchers and policymakers (Lee and Burkam, 2002). The Early
Childhood Longitudinal Study Kindergarten cohort is a rich source of
information from families, teachers, and schools. The analyses presented
here will examine children’s performance on math tests and the extent to
which variation in performance is limited to broad racial/panethnic groups
or is variable by national origins within these groups. We focus on family
characteristics and behaviors that may also explain variation in academic
performance among young children in immigrant families and non-immigrant
families.

DATA AND METHODS

The data for this paper come from The Early Childhood Longitudinal Study
– Kindergarten Cohort (ECLS-K). The ECLS-K is an ongoing data collection
effort by the U.S. Department of Education, National Center of Education
Statistics (NCES). The survey began with a cohort of 22,000 children in
kindergarten. Children are followed longitudinally and will continue to be
followed throughout their elementary schooling and beyond. The data have
the advantage of recording not only the birthplace of the child and one parent
but also allow for a very detailed analysis of racial and ethnic variation in school
outcomes. Further, the ECLS-K allows respondents to indicate more than one
race, which will allow for the inclusion of mixed race children. By focusing on
one cohort of children who are all approximately the same age, the analysis will
be better able to delineate the consequences of generation status net of
developmental stage than studies examining children of all different ages
(Garcia Coll and Magnusun, 1997; Fuligni, 2001). By focusing on a younger
cohort with nationally representative data, this paper is able to extend those
studies with adolescents (i.e., those based on High School and Beyond,
NELS:88, CILS, and Add Health data) to young children as they enter school
(Crosnoe, 2005). Because the ECLS-K is a school-based sample, all of the
380 I M R

descriptive and regression analyses are weighted and adjusted for the clustered
sampling design in the ECLS-K.2
The data employed here are limited in the sense that only children who
enroll in kindergarten are included in the sample. Since kindergarten is not
compulsory in all states some children will necessarily be missed by this criterion.
However, all types of kindergartens including full-day, half-day, public, and
private programs are included. Further, children remain in the sample even if
they change schools or withdraw from school. Data for the analyses presented
here come from participants in the Spring 1999 and Spring 2002 waves of
ECLS-K, by which time most of the children are in the third grade. Our final
sample includes 13,618 children who participated in all three waves of the
study and who have valid test score data in the 2002 wave of the survey. Basic
demographic information on children and their families are available from the
Spring 1999 panel of the data. Data on the place of birth of at least one parent
are available beginning in the Spring 2000 parent interview with additional
items included in the Spring 2002 interview. Unfortunately, this still leaves
1,453 cases with missing information on generation status or parents’ arrival in
the United States. There is some reason to suspect that these cases missing
such information are more likely to be in immigrant families; they are dis-
proportionately likely to have been given the Oral Language Development
Scale (OLDS) test, indicating that their home language is non-English, and
their math scores, like those of immigrant children, are substantially lower than
those of White children in nonimmigrant households. Simply dropping these
cases, as is often done with missing data, may bias our sample in favor of high-
performing children from English-speaking immigrant families. Therefore, we
retain these cases and indicate whether they are missing country-of-origin
information.

Outcome Measures

There are many ways to assess progress in school in the United States and
all have their merits and disadvantages. For example, grades have real
consequences for students. Poor performance in class reflected with poor
grades determines opportunities to proceed to higher grade levels. However,

2We employ the survey commands in SAS for this purpose. We also note that there are several
different weighting options available in the data. We employ longitudinal weights for the
Kindergarten –Third Grade sample. Analyses with alternate weights produced different sample
sizes but substantively similar results.
 A P  C  I 381

grades are likely to vary across schools and teachers within schools, making it
difficult to determine the extent to which academic performance varies due to
differences among teachers and schools or other factors. To make matters more
complicated, grades are more variable in the earlier school years where they are
more likely to reflect behavioral issues as well as academic ability. For these
reasons, analysis of academic progress is measured here with standardized math
test scores from the spring of 2002 and change in test scores from 2000 to
2002. The comparison of the academic achievement of students from diverse
social and educational settings is facilitated when one standardized measure is
available (Bankston and Caldas, 1996). The item response theory (IRT)
adjusted standardized math score is employed.3 We convert these scores into
Z-scores to facilitate comparison of effects across models. Because there is some
temporal variation in the administration of the test, the models also include a
variable for the date of the test.4

Independent Variables

Generation status is measured using the parent survey in the 2000 and 2002
waves. One parent, most often the mother, reported her and the child’s places
of birth.5 Children born outside the United States are counted as children of
the 1.5 generation (n = 298). Those born in the United States to at least one
foreign-born parent are categorized as second-generation children (n = 2,272).
Closely correlated with the children’s generation status is their parents’ age at
arrival in the United States. In our data, the vast majority of the 1.5-generation
children have a parent who arrived in the United States after age 15 (95%) and
was therefore much less likely to experience schooling or peer socialization in

3
The ECLS-K staff evaluated all children from non-English homes for language proficiency.
Those children who were not sufficiently proficient in either English or Spanish were not given
the achievement tests. There were 1,567 children who were not administered the English or
Spanish version test in Kindergarten and only 350 children in first grade who were not assessed.
By third grade, all children are assessed (Rathbun and West, 2004). We keep all children in the
sample. Those children who do not have first-grade math tests are coded as missing on the test.
Analyses eliminating these cases revealed substantively similar results to those we present here.
4This variable is not significant in any models.
5This is due to the survey design, where only one parent was asked about their country of birth.

This parent was usually the mother, since the order of preference for the parent interview was 1)
the child’s mother, 2) the child’s father, and 3) other adult household member, who was asked
information about the mother. We estimate that this focus on mother’s nativity misclassifies the
3% of all 4–6-year-olds in the United States who have a U.S.-born mother and a foreign-born
father (authors’ tabulations for 2000 Census data).
382 I M R

the United States. In contrast, just over 30% of the second-generation children
have a parent who arrived in the United States before age 15 and was therefore
likely to experience schooling and peer socialization in the United States. We
identify second-generation children by the age of migration of their parents,
creating three categories of generation status that approximate the family’s
experience in the United States: 1.5 generation; second generation, parent
arrived after age 15; and second generation, parent arrived before age 15. These
children are compared to those born in the United States whose parent also
reported being born in the United States (generation 3+; n = 9,900). We also
include a category for unknown generation status.
Age at migration and generation status are closely associated with English
language proficiency (Stevens, 1999). Young children may perform less well in
school when they are limited in English proficiency (Stiefel et al., 2003) but
often make rapid progress in English acquisition as well (Espenshade and Fu,
1997). Given its close link to academic performance, we expect that English
proficiency will capture some of the variation in academic performance
observed by generation status and possibly by country of origin. We include a
measure of English language proficiency in the kindergarten year. The Oral
Language Development Scale was administered to all children whose school
record indicated that they spoke a non-English language at home (in the
absence of a school record, information was provided by the teacher). The test
scores range from 0 to 60, with scores of 37 or above indicating English
proficiency. Thus, our measure of English language proficiency is a simple
dichotomy of whether the child took and subsequently failed this English
proficiency test in the fall of the kindergarten year (n = 896). We note that all
of these children pass the test by third grade.
The racial/ethnic identification of the child also comes from the parent
survey. Parents are asked to classify children according to racial and panethnic
group and then may identify children by specific ethnicity within the Hispanic
and Asian panethnic categories. So, for example, all children whose parents
selected “Mexican” also selected the broader “Hispanic” label. We use those
categories with large enough numbers of cases for comparison across all three
generation groups (i.e., there are sufficient cases from the first, second, and
third+ generation to compare within each category6): Non-Hispanic White
(n = 7,713), non-Hispanic Black (n = 1,712), Mexican origin (n = 1,242),

6
The exception to this is American Indian children, for whom the generation designation is not
meaningful.
 A P  C  I 383

Puerto Rican origin (n = 170), other Hispanic origins (n = 1,026), Asian

(n = 1,123), Pacific Islander (n = 202), American Indian or Alaskan native
(n = 230), or of mixed racial identity (n = 355).
There are some studies of children of immigrants focusing on specific
countries of origins. The Children of Immigrants Longitudinal Study (CILS)
conducted in Miami and San Diego has provided information on several of
these groups (Rumbaut and Portes, 2001). Less information by country of origin
has been available at a nationally representative level beyond Census data. To
compare the importance of the racial/ethnic identification of the children to
that of their country-of-origin background, the children of immigrant families
(first and second generation) are also examined separately by country-of-origin
grouping. Children are grouped according to the birthplace of their mother.
No analogous ancestry information is available for the third and higher
generation (i.e., children whose mothers are born in the United States) but
analyses including country of origin also include race/ethnic categories for the
U.S.-born children of U.S.-born parents. This has the advantage of leaving
non-Hispanic white third-and-higher-generation children as the reference
group and demonstrating the extent to which racial/ethnic variation among
the third and higher generation remains significant. The results presented here
include single country of origin when there were more than 40 cases from that
country and regional groupings for smaller numbers of cases. Appendix Table
A specifies which countries are included when categories consist of more than
one country. The following countries and country-groups are included:
Mexico, Puerto Rico, Central America, South America, Cuba, Caribbean, China,
Vietnam, Laos and Cambodia, the Philippines, India, Other East Asian, Other
Asian, Eastern Europe, Western Europe, and other countries (predominately
from Africa or the Middle East, with a few cases each).
To illustrate the overlap between country-of-origin and the racial/
panethnic categories so often employed in survey data, Appendix Table B
reports the racial/ethnic composition of children in the various country-
of-origin groupings. As one might expect, the more countries included in a
group, the greater the racial/ethnic diversity of the children as reported by their
parents. Nonetheless, some interesting patterns are observable. Those children
whose families come from Mexico, Puerto Rico, and Central America are
considerably more likely to be labeled as “Hispanic” than are those from South
America. Similarly, children with at least one parent from China, Vietnam,
Laos, and Cambodia are more likely to be labeled as “Asian” by their parents
than are those from the Philippines or other “Asian” countries. In addition,
considerable attention should be paid to those children from the “Caribbean,”
384 I M R

who are far more likely to be “Black” than any other group of first- or second-
generation children.

Family Resources and Involvement

Family resources are critical to children’s school success, and measures for both
financial resources and family involvement and engagement are included in the
analyses. The distribution of these measures is presented in Table 1. These
measures include the age (in months) and gender of the child, family structure
(two parents, one parent and a parent’s partner including stepparents or
cohabiting partner, single mother, and “other family type”), and number of
siblings. Two measures are included to capture the family’s socioeconomic
status. First, we include family income. This measure is logged in the regression
analyses although the mean is presented in Table 1. Second, we include a
measure of the parent’s education. For the majority of children, this measure
refers to the mother’s education. Father’s education is substituted only if
the mother’s education is not reported. These measures all come from the
kindergarten year measure unless it is missing, in which case the first-grade
measure is substituted.
We next include measures of the early educational experiences that may
influence academic performance. First, we measure childcare arrangements
prior to kindergarten leaving children who were predominately in parental care
only as the reference category. Overall, immigrant families are more likely to
use kin care and less likely to use center-based care than non-immigrant fami-
lies (Brandon, 2002). Our measure includes those receiving care in a home by
a relative, care in a home by a nonrelative, children who attended Head Start
programs, those who attended center-based programs, and finally, children
whose childcare experiences involved multiple settings. The second indicator
of previous experience is for students who are repeating kindergarten in 1999
(n = 488). Although previous research has suggested that immigrant children
may be more subject to grade retention than their U.S.-born peers at higher
grades (Harker, Guo, and Harris, 2001), this does not appear to be the case for
these very young children. A larger proportion of those in the third or higher
generation are repeating kindergarten in 2000 than are the children of immi-
grants in this sample. We also include a measure for whether the child is enrolled
in a full-day or half-day kindergarten program. There is considerable variation
in the availability of full-day kindergarten programs (Lee et al., 2006).
The final set of variables is intended to measure the extent to which families
are involved in activities at the child’s school and the extent to which families
 A P  C  I 385

TABLE 1
SAMPLE CHARACTERISTICS, ECLS-K PARTICIPANTS KINDERGARTEN-THIRD GRADE
Percent Percent
or Mean or Mean
Child Characteristics Pre-Kindergarten Care Arrangements
Age in 1999 (in months) 86.93 Parental care only 34%
(4.2) Home-based relative care 11%
Gender Home-based non-relative care 9%
Female 49% Head Start program only 8%
Male 51% Other center care 34%
Multiple care arrangements 4%
Race/Ethnicity Kindergarten School Characteristics
Non-Hispanic White 56% Private school 13%
Non-Hispanic Black 16% Public school 87%
Mexican origin 10%
Puerto Rican origin 1% Half-day K program 46%
Other Hispanic 8% Full-day K program 54%
Asian origin 4%
Pacific Islander 1% Child Is Repeating Kindergarten (1999) 4%
American Indian 2%
Mixed race and/or ethnicity 2% Kindergarten Activities/Involvement
Generation Status School Involvement (parent)
1.5 generation 2% Attended an open house 68%
2nd gen., parent arrived after age 15 5% Attended a PTA meeting 31%
2nd gen., parent arrived pre age 15 9% Attended a parent/teacher conference 78%
Third or higher generation 73%
Generation status unknown 11%
English Proficiency Non-School Activities (year)
Failed kindergarten OLDS test 6% Sports or arts classes/activitiesa 57%
Passed or did not need test 94% Non-English language instruction 5%
Family Structure in Kindergarten Child Taken on Outings in Last Month
Both parents 69% Went to library in last monthb 50%
Parent & partner 8% Other outings in last month 72%
Single mother 21%
Other family type 2%
Family Socioeconomic Status
Family Income 50,043
(1,076)
Parent Education
Less than high school 10%
High school graduate 27%
Some college 31%
Four-year college degree 16%
More than four-year degree 12%
Number of Cases 13,618
Source: ECLS-K cohort.
Note: Sample weighted and adjusted for design effects; unweighted sample size present.
aChild was enrolled in an organized sport, took dance, music, or arts classes.
bFamily/household member took child to one of these in previous month: museum, zoo, aquarium, play or concert,

game.
386 I M R

engage the child in other non-school activities that could also enhance
academic performance. First, we use three dichotomous measures of parental
involvement in the kindergarten year: attendance at an open house event at the
school; attendance at a parent/teacher organization meeting at the school, and
attendance at a parent/teacher conference at the school.
We also use several measures of non-school activities. We expect that
immigrant parents could face difficulty becoming involved with the school but
may still use their resources to enhance their children’s academic opportunities.
These activities could include being enrolled in a variety of sports or arts classes
outside of the school in their kindergarten year. We include a separate variable
for enrollment in a non-English language class; these classes may be used by
parents seeking to teach their children the traditions and practices of the
country of origin. And, it is the case that children of immigrants are more likely
to be enrolled in these non-English language classes than their peers with U.S.-
born parents (11% vs. 4%). In addition to formal enrollment in classes, we
include two variables for outings taken with any member of the family and
the focal child in the previous month. The first variable represents outings to
the library and the second other outings, including trips to zoos, museums,
aquariums, concerts, plays, or sporting events.

RESULTS

Children’s early academic performance varies by mother’s nativity, race/

ethnicity, and English proficiency, suggesting considerable diversity from the
outset of formal schooling. Figure I presents the average math scores (IRT
scaled, z-scores) in the spring of 2002, four years after introduction to formal
schooling. While there are statistically significant differences between children
of immigrant mothers (the 1.5 and second generation) and children of U.S.-
born mothers, we observe greater differences by English proficiency and race/
ethnicity. Children from all racial/ethnic groups have lower scores when compared
to non-Hispanic White children. American Indian and non-Hispanic Black
children score the lowest. Likewise, children who did not pass the OLDS test
in kindergarten perform less well on the math test three years later when
compared to those who were not identified as non-English proficient. Clearly,
the diversity of origins of immigrants in the United States makes it difficult to
tell a single “immigrant” story of academic success or failure.
These descriptive results do not control for the diversity of family and
socioeconomic backgrounds of children in immigrant families, leaving the
question of whether differences in academic performance will persist in the face
 A P  C  I 387

Figure I. Math Scores in Third Grade (z-Scores) by Generation Status, Race/

Ethnicity, and English Proficiency

Source: ECLS-K longitudinal data, Kindergarten-Third Grade sample.

Note: Y-axis represents mean for overall sample.
*Indicates significant difference from overall mean (p < 0.05).

of controls for socioeconomic status, family structure, previous school

experiences, and family involvement in school and non-school activities. It
seems plausible that these family characteristics may account for differences
in academic performance among children in immigrant families. To examine
this possibility, results of the regression models predicting the math test scores
of spring 2002 (z-scores) are presented in Table 2.
The first model demonstrates that family structure and socioeconomic
status are important predictors of early academic performance. Here we see
results consistent with previous work on academic performance. Boys are
already outperforming girls on the standardized math test. Students who are
older also outperform those who are younger. Children who come from more
affluent families and have parents with higher levels of education also evidence
higher scores. Likewise, children from homes with other than two biological (or
adoptive) parents do not perform as well on the math test. With these controls,
second-generation children still score significantly below their peers with
388 I M R

TABLE 2
REGRESSION MODELS PREDICTING SPRING 2002 MATH TEST SCORES (IRT; Z-SCORES)
Model 1 Model 2 Model 3 Model 4
Male 0.16*** 0.16*** 0.16*** 0.11**
Child’s Age (in months) 0.02*** 0.02*** 0.02*** −0.01***
Parent Age (in years) 0.01*** 0.01*** 0.00 0.00
Family Structure (vs. Both Parents)
Parent & partner −0.16*** −0.15*** −0.13*** −0.05*
Single parent −0.10* 0.00 0.00 0.02
Neither parent −0.51*** −0.27** −0.20* −0.06
Number of siblings −0.07*** −0.06*** −0.04*** −0.02**
Family Income (log) 0.11*** 0.08*** 0.05*** 0.03***
Parent Education (vs. More than College)
Less than high school −0.81*** −0.76*** −0.64*** −0.22***
High school graduate −0.57*** −0.54*** −0.47*** −0.14***
Some college −0.29*** −0.27*** −0.27*** −0.06**
Four-year degree −0.02 −0.03 −0.09* −0.01
Generation Status (vs. Third+ Generation)
1.5 generation 0.04 0.10 0.18* 0.17**
2nd gen., parent arrived after age 15 −0.09* −0.01 0.05 0.05
2nd gen., parent arrived pre age 15 −0.17** −0.12* −0.09 −0.04
Race/Ethnicity (vs. Non-Hisp. White)
Non-Hispanic Black −0.56*** −0.51*** −0.21**
Mexican origin −0.23*** −0.19*** −0.09*
Puerto Rican origin −0.28* −0.22 0.04
Other Hispanic −0.25*** −0.22*** −0.06
Asian origin 0.03 0.09 0.08**
Pacific Islander −0.20* −0.20* 0.01
American Indian −0.64*** −0.53*** −0.20
Mixed race and/or ethnicity −0.12 −0.12 0.00
Not Proficient in English (vs. Proficient in Base Year) −0.08 −0.06 0.07
Pre-K Childcare Arrangements (vs. Only Parents)
Relative in home care −0.01 −0.02
Non-relative in home care 0.19*** 0.11***
Head Start program −0.14** −0.06*
Other center care 0.16*** 0.07***
Multiple arrangements 0.05 0.05
Parental School Involvement (Kindergarten Year)
Attended open house at school 0.15*** 0.05*
Attended PTA meeting 0.04 0.02
Attended parent/teacher conference −0.01 0.00
Non-School Involvement (Kindergarten Year)
Sports or arts classes/activities (a) 0.12*** 0.00
Non-English language instruction 0.00 0.00
Went to library in last month (b) 0.07** 0.01
Other outings in last month 0.08** 0.04
School Characteristics (Kindergarten Year)
Private school (vs. public) −0.05 −0.09***
Half-day K program (vs. full-day) −0.05 0.02
Child Repeating K in 1999 −0.30*** −0.09*
Math Test Score in 2000 0.73***
Intercept −2.73 −2.11 −2.41 0.57
R square 0.19 0.22 0.25 0.65
Source: ECLS-K Third Grade longitudinal sample (n = 13,618).
Note: Regression models weighted and adjusted for design effects. Models include measures for missing parent age, number
of siblings, or generation status and a measure for the day the test was administered.
*p < 0.05.
**p < 0.01.
***p < 0.001.
 A P  C  I 389

U.S.-born parents. However, there is no evident difference between the 1.5-

generation children and children with U.S.-born parents.
The second model in Table 2 adds measures of race/ethnicity and En-
glish proficiency in the kindergarten year (1999). Limited English proficiency
at the outset does not significantly impact subsequent academic performance
net of family background and socioeconomic status. However, racial/ethnic
differences are strong. Children identified as Black, Mexican origin, other
Hispanic, Pacific Islander, or American Indian do not perform as well on the third-
grade math test as those identified as non-Hispanic White. We also examined
the possibility that there are divergent patterns of academic performance
among children of immigrant families from different racial/ethnic minority
groups by testing interactions of generation status and race/ethnicity (model
not shown).7 We find a significant negative interaction for 1.5-generation children
of Mexican and Pacific Island origin, suggesting their performance lags even
further behind others. However, it seems likely that there could be other
differences we do not observe in the interactions of larger panethnic groups.
For example, there is more variation between “Asian” first- and second-
generation children in language background and family SES, for example, than
between “Hispanic” first- and second-generation children. We explore this
possibility in subsequent analyses by identifying children by their foreign-born
parent’s national origins.
Our third model in Table 2 addresses the possibility that generation status
differences are reduced by other family resources or behaviors that enhance
children’s academic performance. We observe that children’s pre-kindergarten
care arrangements do influence outcomes as late as third grade. Children in
center-based care and children cared for by non-relatives in their own home
performed better than children cared for by only their parents, while children
in Head Start programs had lower scores, perhaps picking up on unmeasured
traits that selected children into eligibility for Head Start in the first place.
Further, parental attendance at an open house event at the child’s school is
associated with higher math scores. We also observe higher scores among
children who participated in non-school activities and outings. Children who
were retained in kindergarten also lag behind their peers who made normative
grade progression, but we observe no differences for the type of school or length
of school day in the kindergarten year. With all of the measures in the model,

7Additional models (not shown) test the interaction between language and generation status, but

these interactions are not significant.

390 I M R

there is some advantage for children of the 1.5 generation and no longer
a deficit among the second-generation children. The 1.5 generation children’s
scores are not higher until we adjust for the fact that their families tend to be less
involved in the school. Interactions of school involvement and participation in
non-school activities with generation status reveal no significant differences
in the effects of these activities for children of immigrants or children of U.S.-
born parents. Thus, parental involvement appears beneficial for all.
The models discussed thus far do not adjust for previous academic
performance. To evaluate whether children of immigrants have improvement
in academic performance relative to their third-and-higher-generation peers,
we also estimated models including previous math test scores (also IRT scaled)
from the spring 2000 wave of the survey. We mean-substitute the scores for
children missing the 2000 test and include a dichotomous variable indicating
they are missing the 2000 test. Thus, we maintain the same sample of children
in all models.8 The results are presented in Model 4. Consistent with previous
research, higher family income and higher parental education are associated
with greater gains in test scores over time. We also observe an improvement in
scores among children of the 1.5 generation. Racial/ethnic differences also
persist over time as Black and Hispanic children’s scores are lower and Asian-
origin children’s scores are higher by the third-grade test when we control for
prior performance.
The results thus far suggest some differences among children of
immigrants and children of U.S.-born parents but the directions are not clear.
The results suggest some advantage to children from more recently arrived
families. However, the strong racial/ethnic effects with some modest interactions
with generation status (suggesting worse outcomes for Mexican and Pacific
Islander children from the 1.5 generation) also suggest a need to further
disaggregate.
Generation status is correlated with national origin. Some groups are
more recently arrived than others and some are more likely to have parents
arriving as adults. To explore the possibility that children of immigrants differ
in their academic performance when national origin groups are divided rather
than relying on generation status alone, Table 3 presents the results of the same
regression models predicting math test scores in spring 2002 with an expanded
measure of race and national origin replacing the measure of race/ethnicity
and generation status from the previous models. Third-and-higher-generation

8
Models using only children with valid test scores in first grade (n = 13,379 vs. 13,618) are not
substantively different from those presented here.
 A P  C  I 391

TABLE 3
REGRESSION MODELS PREDICTING 2002 MATH SCORES (IRT, Z-SCORES); RACE/ETHNICITY
AND COUNTRY OF ORIGIN

Model 1 Model 2 Model 3 Model 4

Race/Ethnicity or Country of Origin (vs. non-Hispanic White; 3rd+ Generation)
Black, 3rd+ generation −0.73*** −0.53*** −0.48*** −0.20***
Mexican origin, 3rd+ generation −0.27** −0.11 −0.09 0.00
Puerto Rican origin, 3rd+ generation −0.26 −0.21 −0.15 0.06
Other Hispanic origin, 3rd+ generation −0.19** −0.09 −0.08 0.02
Asian, 3rd+ generation 0.05 −0.05 −0.02 0.09
Pacific Islander, 3rd+ generation −0.30* −0.25** −0.23* 0.02
American Indian −0.79*** −0.60*** −0.48*** −0.18
Mixed Race/Ethnicity, 3rd+ generation −0.14 −0.07 −0.07 0.01
First or Second Generation
Mexico −0.69*** −0.30*** −0.22*** −0.13***
Puerto Rico −0.55** −0.43 −0.33 0.07
Central American −0.29** −0.05 0.02 0.14*
South American −0.09 −0.08 −0.02 0.13*
Cuba 0.25 0.19 0.10 0.07
Caribbean −0.73*** −0.65*** −0.54** −0.12
China 0.76*** 0.45*** 0.54*** 0.37**
Other East Asian 0.51*** 0.26* 0.30* 0.21**
Vietnam 0.40*** 0.42*** 0.58*** 0.38***
Laos/Cambodia −0.53*** −0.27* −0.11 −0.02
The Philippines −0.11 −0.26 −0.16 0.01
India 0.41** 0.13 0.21 0.18
Other Asian 0.05 0.12 0.16 0.08
Eastern Europe 0.36** 0.18 0.31* 0.24**
Western Europe 0.30** 0.20* 0.20* 0.16*
Other countries −0.05 −0.15 −0.05 0.05
Not Proficient in English (vs. Proficient in Base Year) −0.05 −0.02 0.10*
Math Score in 2000 0.73***
Intercept 0.11*** −2.21*** −2.55*** 0.53***
R square 0.09 0.22 0.25 0.65
Source: ECLS-K Third Grade longitudinal sample (n = 13,618).
Notes: All models are weighted and adjusted for design effects.
Model 1 = Race/ethnicity and country variables only.
Model 2 = Model 1 + sex, age, parents’ age, parents’ education, income, family structure, number of siblings, and
kindergarten language proficiency.
Model 3 = Model 2 + pre-kindergarten care arrangements, Parental school involvement, non-school involvement,
kindergarten school characteristics.
Model 4 = Model 3 + test score in 2000.
*p < 0.05.
**p < 0.01.
***p < 0.001.

children are identified by their race or panethnicity. Children in the 1.5 and
second generations are identified by their foreign-born parent’s country of
origin. This means that non-Hispanic White children in the third and higher
generation serve as the reference group for all others. The models in Table 3
correspond to those in Table 2. The effects of socioeconomic status and family
392 I M R

resources are similar in both tables, so we do not report these coefficients in

Table 3.
We first estimate the model with only the variables for race/ethnicity and
national origins. This first model in Table 3 demonstrates considerable variation
in academic performance by national origins and race/ethnicity. Third-and-
higher-generation minority children do not perform as well as their non-
Hispanic White counterparts. Looking at the first- and second-generation
children by national origins reveals significant differences not apparent when
the single indicator of generation status is employed. For example, Mexican-
and Caribbean-origin children do not perform as well on the third grade test
as non-Hispanic White children in the third and higher generation while
children of Chinese and Vietnamese origins pull ahead. We also note the positive
effect for the children of Eastern and Western European immigrants, mostly
identified as non-Hispanic White by their parents, who also outperform non-
Hispanic White children of U.S.-born parents.
Model 2 replicates the analyses with controls for family structure,
income, parents’ education and age, number of siblings, and language proficiency
(corresponding to Model 2 from Table 2). Some of the variation by race/
panethnicity and national origins are reduced here but several groups remain
quite distinct. Even when we include family involvement and non-school
activities in Model 3 (corresponding to Model 3 from Table 2), we observe
persistent differences.
Finally, we estimate the same change model with an adjustment for
previous math test scores. The change model demonstrates the persistence of
racial/panethnic and national-origin differences over time. Some groups are on
a “negative” trajectory over time that is not explained by their socioeconomic
status or family resources. For example, Black third-and-higher-generation
children and Mexican-origin children see test scores decline over time when
compared to non-Hispanic White third-and-higher-generation children. Like-
wise, the higher performance of children of Chinese, Vietnamese, and other
East Asian migrants persists over time indicating even greater divergence from
non-Hispanic White third-and-higher-generation children.
The results by national origins are strongly suggestive of some emerging
“panethnic” patterns of academic performance. In other words, it seems that
children of “Hispanic” immigrant origins lag behind and children of some, but
not all, “Asian” immigrant origins pull ahead. Does this mean that there is a
divergent trajectory for children of immigrants that will see them ultimately
converge with third-and-higher-generation peers from the same racial or
panethnic groups? Does the color line become reified at such an early age?
 A P  C  I 393

Figure II. Predicted Average 2002 Math Test Scores by Race/Ethnicity for Third+
Generation or Parent’s Country of Origin for Children of Immigrants

Source: ECLS-K, Kindergarten Class of 1998–1999; weighted and adjusted for design effects.
Note: *Predicted value is significantly different from non-Hispanic White, 3rd+ generation.

One way to address these questions is to compare children not only to

non-Hispanic White children but to their third-generation co-ethnic peers as
well. We illustrate this in Figure II, where we present predicted math test scores
(from the full Model 4 of Table 3) for all racial/ethnic third-and-higher-
generation groups along with national origin groups of first- and second-
generation children. The predicted math test score for Whites in the third and
higher generation is presented as the axis so that each bar demonstrates predicted
distance (either positive or negative) from this reference group, making clusters
by ethnicity or national origins, where they exist, more visible.
One cluster that emerges is made up of non-Hispanic Black third-and-
higher-generation children, Pacific Islander third-and-higher-generation children,
and American Indian children along with children of Caribbean-origin
immigrants. These groups all lag behind non-Hispanic Whites and yet are not
statistically significantly different from one another. A second cluster emerges
among several of the Hispanic groups. Children of Mexican-origin immigrants
do score below the non-Hispanic White reference group but they also score
significantly above this first disadvantaged cluster. Thus, a second group emerges
394 I M R

among Mexican-origin children (particularly third and higher generation),

Puerto Rican-origin children (all generations), and other “Hispanic”-origin
children of immigrants whose predicted scores are not statistically significantly
different from one another. Finally, there is a third clear cluster of scores among
Chinese, East Asian, Vietnamese, and European children of immigrants who all
have scores significantly higher than every other racial/ethnic third-generation
group including Asian-origin children of the third and higher generation.
Although these may be substantively small differences in terms of individual
learning trajectories, they do suggest significantly different patterns by group
that could continue into the children’s future. Such a result would certainly
coincide with previous research findings on adolescents. The inclusion of
European immigrants’ children in this highest cluster also suggests that divergent
outcomes by nativity are not confined to racial or ethnic minority groups in the
United States and may point to the significance of selection of these immigrant
parents (Feliciano, 2005).

DISCUSSION

Children of immigrants come from diverse backgrounds and enter school with
different experiences and resources. This paper addresses two basic questions:
(1) To what extent does generation status or family experience in the United
States exert independent effects on early school performance even in the face
of controls for family background, race/panethnicity in the U.S., and language
proficiency? (2) To what extent do these broad conceptualizations of children
in immigrant families mask variation by national origins? We also assessed the
extent to which family behaviors, such as school involvement or non-school
activities and outings, explain differential outcomes for children of immigrant
parents. Our analyses take advantage of a large sample of young children in
U.S. schools to ask which characteristics so closely entwined with immigrant
status today predict early school performance. Our focus on race/ethnicity in
particular is designed to address current debates over the continued significance
of race for children’s academic performance and questions regarding the extent
to which children of immigrants are affected by being embedded in the U.S.
racial/ethnic hierarchy. Further, we take advantage of the longitudinal nature
of the data to ask whether deficits present in the first year or two of formal
school persist two years later.
The results support the findings of numerous studies that have continually
emphasized the importance of considering diversity within the immigrant
population (i.e., Portes and Rumbaut, 2001, among others). Our results point
 A P  C  I 395

to a continued significance of race/ethnic origins that persists net of family

background or even language proficiency. Further, our results suggest that
some children of immigrants do have divergent academic trajectories such that
some groups see scores move toward those of third-and-higher-generation
minority groups while others surge ahead. These divergent trajectories are not
explained away by differential socioeconomic status or other family background
characteristics measured after arrival in the United States. We cannot directly
assess selectivity of migration with these data.
The results also suggest that parental behaviors and involvement are
associated with academic outcomes among young children. Family resources
and behaviors are also beneficial to all children regardless of nativity, supporting
calls to increase parental involvement at home and at school.
Parents who attend school events, enroll children in classes outside of
school, or take children on outings may be more motivated or better informed
about their children’s needs or better able to seek assistance for children’s
academic skill acquisition. Regardless of whether these measures indicate some
dimension of parental motivation or are directly related to children’s academic
success, it is clear that these activities alone cannot completely account for
disadvantages associated with race/ethnic position in the United States.
The analyses presented here also demonstrate the importance of looking
within generation status at national origins. We find some intriguing support
for the idea that adaptation in the United States really takes on a “segmented”
feature along racial lines but that not all national origin groups can be so easily
categorized. When we divide children according to national origins and
compare their test scores to third-and-higher-generation children from various
racial/ethnic backgrounds, we are able to observe clustering of scores for Black
third-and-higher-generation children, children of Caribbean immigrants, and
American Indian children. Children of Mexican immigrants lag behind their
third-generation co-ethnic peers but still perform significantly better than this
lowest scoring group. It remains to be seen whether these children will lose
ground over time or rise to the scores of the third-and-higher-generation
Mexican-origin children. We are also able to replicate other studies suggesting
Asian origin immigrants are a diverse group with divergent paths through
school. While children of Vietnamese, Chinese, and other East Asian-origin
immigrants surpass all of their peers on the math test score, children of Laotian
or Cambodian immigrants, Indian, and Filipino immigrants have lower scores.
And, when we adjust for prior test performance, we still see some divergence and
clustering, suggesting that these group differences may lead to quite different
academic trajectories as children move through school.
396 I M R

Our findings are consistent with those found for adolescents. Racial/
ethnic minority children, particularly black and Hispanic youth from all
generations, perform less well on academic achievement tests than their
non-Hispanic white counterparts. In adolescence, significant racial/ethnic dif-
ferences in academic achievement persist through schooling while generation
status appears to exert less of an impact on subsequent academic performance
once previous performance is taken into account (Glick and White, 2003).
It seems likely that the effect of racial/panethnic group could eventually
predominate over the national origin differences we find here when predicting
academic performance into the future. Analyses with later waves of the survey
will help elucidate the long-term trajectories and relative importance of all of
these factors so closely intertwined with one another.
There are some limitations to the analyses presented here. Most notably,
the country-of-origin background for children in the third and higher
generation is unavailable. This means all third-and-higher-generation children
are compared within the same racial and panethnic groups that mask such
variation for first- and second-generation children. While the racial/ethnic
groupings presented may have some real consequences for the lives of these
children, they may still mask important diversity within the third-and-higher-
generation groups. In addition, while we can observe some differences by
parent’s age at migration, we are unable to determine the mode of entry to
the United States by the immigrant parents. It seems likely that some of the
differences we observe that are not captured by controls for income or parental
education are still due to differential selectivity of migrants across countries
(Feliciano, 2005). Finally, race/ethnicity cannot be used as a sole proxy for con-
text of reception. Future analyses including other contextual variables will be
better able to assess the importance of context of reception variation than the
analyses presented here.
Overall, the results presented here suggest that academic performance
diverges by nativity and, more specifically, by national origins, in the early years
of formal schooling in the United States. This is of particular concern when we
consider the children in immigrant families who may face blocked opportunities
to their subsequent achievement that keep them on similar trajectories as children
from the same minority panethnic group in the third or higher generation
(Portes and Rumbaut, 2001). Should this occur, we can expect recent immi-
gration to simply feed the system of racial/ethnic inequality in the United
States. The analyses presented here provide an initial view of school performance;
further waves of data will allow us to observe the extent to which groups
converge over time.
 A P  C  I 397

APPENDIX TABLE A
COUNTRY-OF-ORIGIN GROUPINGS FROM ECLS DATA
Number
Country Grouping Label Countries Included in Group of Cases
Mexico Mexico 765
Puerto Rico Puerto Rico 59
Central American Belize, Costa Rica, El Salvador, Guatemala, Honduras, Nicaragua 138
South American Argentina, Chile, Uruguay, Bolivia, Peru, Ecuador, Colombia, 139
Venezuala, Guyana, Brazil, Panama
Cuba Cuba 41
Caribbean Aruba, Bahamas, Bermuda, Dominica, Haiti, Jamaica, 139
Trinidad & Tobago, Virgin Islands, Dominican Republic
China China, Taiwan, Hong Kong 92
Vietnam Vietnam 87
Laos/Cambodia Laos, Cambodia 108
The Philippines The Philippines 212
India India 89
Other East Asian North Korea, South Korea, Japan 68
Other Asian Malaysia, Singapore, Sri Lanka, Thailand, Burma 63
Eastern Europe Former Macedonia, Bulgaria, Romania, Croatia, Slovakia, 71
Poland, Albania, Malta, Russia, Ukraine, Georgia, Turkey,
Armenia, Moldova
Western Europe Canada, New Zealand, Austria, Australia, Andorra, Belgium, 185
France, Germany, Luxembourg, Cyprus, Greece, Netherlands,
Portugal, San Marino, Spain, United Kingdom, Finland, Sweden,
Iceland, Switzerland, Italy/Vatican City
Other Countries Pakistan, Afghanistan, Iran, Iraq, Kuwait, Jordan, Oman, Qatar, 149
Saudi Arabia, United Arab Emirates, Yemen, Lebanon, Israel,
Syria, South Africa, Zimbabwe, Burundi, Kenya, Ethiopia,
Djibouti, Nigeria, Togo, Guinea, Ghana, Liberia, Cape Verde,
Sierra Leone
Source: ECLS-K, Kindergarten Class of 1998–1999.
 398
APPENDIX TABLE B
RACIAL/ETHNIC IDENTITY OF FIRST- OR SECOND-GENERATION CHILDREN BY PARENTS’ COUNTRY OF ORIGIN
US/3+ Generation Mexico Puerto Rico Central American South American Cuba Caribbean Eastern Europe Western Europe
Non-Hispanic White 68.0 0.4 4.4 10.4 17.0 4.8 3.8 91.0 74.9
Non-Hispanic Black 15.9 0.1 2.2 7.0 2.7 0.0 38.1 1.3 6.4
Mexican Origin 5.4 91.0 1.4 7.6 2.5 0.0 0.3 4.6 4.0

I M R

Puerto Rican Origin 0.9 0.2 75.9 0.6 3.2 1.5 6.2 0.0 2.0
Other Hispanic 0.4 7.9 15.3 73.0 62.6 93.8 42.9 0.0 5.4
Asian Origin 1.4 0.5 0.8 0.0 9.2 0.0 6.2 0.4 4.6
Pacific Islander 0.5 0.2 0.0 0.0 1.1 0.0 1.1 0.0 1.6
American Indian 1.9 0.0 0.0 0.0 0.5 0.0 0.5 0.0 0.0
Mixed Race/Ethnic. 2.5 0.0 0.0 0.0 0.7 0.0 0.8 2.7 1.0
China Other East Asian Vietnam Laos/Cambodia The Philippines India Other Asian Other Countries
Non-Hispanic White 0.0 28.5 2.1 0.5 2.5 2.7 4.2 29.8
Non-Hispanic Black 0.0 0.0 0.0 0.0 0.0 0.0 36.6 25.1
Mexican Origin 0.0 1.9 0.0 0.0 0.0 0.0 2.0 0.9
Puerto Rican Origin 0.0 0.0 0.0 0.0 0.0 0.0 0.0 3.3
Other Hispanic 0.0 0.0 0.0 0.0 3.0 0.0 0.0 31.1
Asian Origin 100.0 65.5 96.6 96.6 75.3 97.3 36.2 8.4
Pacific Islander 0.0 0.0 1.3 2.9 13.8 0.0 20.6 0.1
American Indian 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
Mixed Race/Ethnic. 0.0 4.1 0.0 0.0 5.4 0.0 0.4 0.3
Source: ECLS-K, Kindergarten Class of 1998–1999.
Note: Sample weighted and adjusted for design effects.
 A P  C  I 399

REFERENCES

Alba, R., J. Logan, A. Lutz, and B. Stults

2002 “Only English by the Third Generation? Loss and Preservation of the Mother Tongue
Among the Grandchildren of Contemporary Immigrants.” Demography 39(3):467–484.
Alba, R., and V. Nee
2003 Rethinking the American Mainstream: Assimilation and Contemporary Immigration. Cam-
bridge MA: Harvard University Press.
Bankston, C. L. III, and S. J. Caldas
1996 “Majority African American Schools and Social Injustice: The Influence of De Facto
Segregation on Academic Achievement.” Social Forces 75:535–555.
Brandon, P. D.
2004 “The Child Care Arrangements of Preschool-Age Children in Immigrant Families in
the United States.” International Migration 42(1):65 –87.
———
2002 “The Child Care Arrangements of Preschool Children in Immigrant Families in the
United States.” Working Paper. New York: The Foundation for Child Development.
Christian, K., F. J. Morrison, and F. B. Bryant
1998 “Predicting Kindergarten Academic Skills: Interactions among Child Care, Maternal
Education and Family Literacy Environments.” Early Childhood Research Quarterly
13(3):501–521.
Cosden, M., J. Zimmer, C. Reyes, and M. del Rosario Gutierrez
1995 “Kindergarten Practices and First Grade Achievement for Latino Spanish-speaking,
Latino English-speaking and Anglo Students.” Journal of School Psychology 33:123–141.
Crosnoe, R.
2005 “Double Disadvantage or Signs of Resilience? The Elementary School Contexts of
Children from Mexican Immigrant Families.” American Educational Research Journal
42(2):269–303.
Driscoll, A. K.
1999 “Risk of High School Dropout Among Immigrant and Native Hispanic Youth.” Inter-
national Migration Review 33(4):857– 875.
Duran, B. J., and R. E. Weffer
1992 “Immigrants’ Aspirations, High School Process and Academic Outcomes.” American
Educational Research Journal 29(1):163 –181.
Entwisle, D., and K. L. Alexander
1993 “Entry into School: The Beginning School Transition and Educational Stratification in
the United States.” Annual Review of Sociology 19:401– 423.
Espenshade, T. J., and H. Fu
1997 An Analysis of English-Language Proficiency among U.S. Immigrants. American Socio-
logical Review 62(2):288 – 305.
Feliciano, C.
2005 “Does Selective Migration Matter? Explaining Ethnic Disparities in Educational Attain-
ment among Immigrants’ Children.” International Migration Review 39(4):841–871.
Fernandez-Kelly, M. P., and R. Schauffler
1994 “Divided Fates: Immigrant Children in a Restructured U.S. Economy.” International
Migration Review 28:662 – 689.
400 I M R

Fuligni, A.
2001 “A Comparative Longitudinal Approach to Acculturation Among Children from
Immigrant Families.” Harvard Educational Review 71:566–577.
———
1997 “The Academic Achievement of Adolescents from Immigrant Families: The Roles of
Family Background, Attitudes and Behavior.” Child Development 68(2):351–363.
Garcia Coll, C., and K. Magnuson
1997 “The Psychological Experience of Immigration: A Developmental Perspective.” In
Immigration and the Family: Research and Policy on U.S. Immigrants. Ed. A. Booth,
A. C. Crouter, and N. Landale. Mahwah NJ: Lawrence Erlbaum. Pp. 91–131.
Glick, J. E., and M. J. White
2004 “Parental Aspirations and Post-Secondary School Participation Among Immigrant and
Native Youth in the United States.” Social Science Research 33:272–299.
———
2003 “The Academic Trajectories of Immigrant Youth: Analysis Within and Across Cohorts.”
Demography 40:759 –784.
Goyette, K., and Y. Xie
1999 “Educational Expectations of Asian American Youths: Determinants and Ethnic Differ-
ences.” Sociology of Education 72:22 – 36.
Griffin, E. A., and F. J. Morrison
1997 “The Unique Contribution of Home Literacy Environment to Differences in Early
Literacy Skills.” Early Child Development and Care 127–128:233–243.
Hao, L., and M. Bonstead-Bruns
1998 “Parent-Child Differences in Educational Expectations and the Academic Achievement
of Immigrant and Native Students.” Sociology of Education 71:175–198.
Harker, K., G. Guo, and K. M. Harris
2001 “Grade Retention among Generations of Immigrant Adolescents.” Paper presented at the
Annual Meeting of the Population Association of America, Washington DC.
Jamieson, A., A. Curry, and G. Martinez
2001 “School Enrollment in the United States-Social and Economic Characteristics of
Students.” Current Population Reports. Washington DC: U.S. Department of Commerce.
Pp. 20–533.
Kao, G.
1995 “Asian Americans as Model Minorities? Look at Their Academic Performance.” American
Journal of Education 103:121–159.
Kao, G., and M. Tienda
1995 “Optimism and Achievement: The Educational Performance of Immigrant Youth.” Social
Science Quarterly 76(1):1–19.
Kim, R. Y.
2002 “Ethnic Differences in Academic Achievement Between Vietnamese and Cambodian
Children: Cultural and Structural Explanations.” The Sociological Quarterly 43(2):213 –235.
Lareau, A.
1989 Home Advantage: Social Class and Parental Intervention in Elementary Education. London:
Falmer Press.
Lee, V. E., and D. T. Burkam
2002 Inequality at the Starting Gate: Social Background Differences in Achievement as Children
Begin School. Washington DC: Economic Policy Institute.
 A P  C  I 401

———, et al.
2006 “Full-Day versus Half-Day Kindergarten: In Which Programs Do Children Learn More?”
American Journal of Education 112:163 –208.
Lopez, D. E., and R. D. Stanton-Salazar
2001 “Mexican Americans: A Second Generation at Risk.” In Ethnicities: Children of Immi-
grants in America. Ed. R. G. Rumbaut and A. Portes. Berkeley: University of California
Press. Pp. 57–90.
Muller, C.
1993 “Parental Involvement and Academic Achievement: An Analysis of Family Resources
Available to the Child.” In Parents, Their Children and the Schools. Ed. B. Schneider and
J. S. Coleman. Boulder: Westview Press. Pp. 77–113.
Portes, A., and D. MacLeod
1996 “Educational Progress of Children of Immigrants: The Roles of Class, Ethnicity and
School Context.” Sociology of Education 69:255 –275.
———, and R. G. Rumbaut
2001 Legacies: The Story of the Immigrant Second Generation. Berkeley: University of California Press.
———, and M. Zhou
1993 “The New Second Generation: Segmented Assimilation and Its Variants.” The Annals of
the American Academy of Political and Social Science 530:74–96.
Rathbun, A., and J. West
2004 From kindergarten through third grade: Children’s beginning school experiences. (NCES
2004–2007). U.S. Department of Education, National Center for Education Statistics.
Washington DC: Government Printing Office.
Rosenthal, A. S., K. Baker, and A. Ginsburg
1983 “The Effect of Language Background on Achievement and Learning among Elementary
School Students.” Sociology of Education 56:157–169.
Rumbaut, R., and A. Portes
2001 Ethnicities: Children of Immigrants in America. Berkeley: University of California Press.
Senechal, M., and J. LeFevre
2002 “Parental Involvement in the Development of Children’s Reading Skill: A Five-Year
Longitudinal Study.” Child Development 73(2):445 – 460.
Stanton-Salazar, R. D.
1997 “A Social Capital Framework for Understanding the Socialization of Racial Minority
Children and Youth.” Harvard Educational Review 67:1–40.
Stevens, G.
1999 “Age at Immigration and Second Language Proficiency Among Foreign-born Adults.”
Language in Society 28:555 –578.
Stiefel, L., A. E. Schwartz, and D. Conger
2003 “Language Proficiency and Home Languages of Students in New York City Elementary
and Middle School.” Working Paper. New York: Taub Urban Research Center, New York
University.
Suarez-Orozco, C., and M. Suarez-Orozco
2001 Children of Immigration. Cambridge MA: Harvard University Press.
Sue, S., and S. Okazaki
1990 “Asian American Educational Achievement: A Phenomenon in Search of an Explana-
tion.” American Psychologist 45:913 –920.
402 I M R

Sy, S. R., and J. E. Schulenberg

2005 “Parent Beliefs and Children’s Achievement Trajectories during the Transition to School
in Asian American and European American Families.” International Journal of Behavioral
Development 29(6):505 –515.
Teachman, J. D., K. Paasch, and K. Carver
1997 “Social Capital and the Generation of Human Capital.” Social Forces 75(4):1343–1359.
White, M., and G. Kaufman
1996 “Language Usage, Social Capital, and School Completion among Immigrants and
Native-Born Ethnic Groups.” Social Science Quarterly 78(2):385–398.
Zhou, M.
1997a “Segmented Assimilation: Issues, Controversies and Recent Research on the New Second
Generation.” International Migration Review 31(4):975–1008.
———
1997b “Growing Up American: The Challenge Confronting Immigrant Children and Children
of Immigrants.” Annual Review of Sociology 23:63 –95.