Untitled-4
Dual Language Proficiency
and the Educational Attainment of Latinos
Amy Lutz
Syracuse University
ABSTRACT
The relationship between Spanish maintenance and the educational attainment of
English-speaking Latino students in the United States is investigated using data
from the National Education Longitudinal Survey. The analysis indicates that biliterate
students are significantly more likely to complete high school compared to their
monolingual peers. Those with oral proficiency in Spanish and English are not sig-
nificantly different in terms of their likelihood of high school completion than were
those who spoke only English. Biliterate students are also significantly more likely to
enter college than those who speak only English. Overall, this research indicates that
bilingualism, offers Latino students an advantage in terms of educational attainment.
Keywords: 1. international migration, 2. bilingualism, 3. Spanish, 4. Latinos, 5. United
States.
RESUMEN
Usando datos de la Encuesta Logitudinal sobre Educación Nacional, se investiga la
relación entre la conservación del español y los logros educativos de los estudiantes
latinos en Estados Unidos que hablan inglés. El análisis indica que los estudiantes
bilingües significativamente tienen más probabilidad de completar el bachillerato
(high school) en comparación con sus compañeros monolingües. Aquellos estudiantes
con la habilidad de hablar bien español e inglés no son significativamente diferentes
–en términos de probabilidad– de terminar el bachillerato y tienen significativamente
más probabilidad de entrar a la universidad que aquellos que solamente hablan inglés.
En conclusión, esta investigación muestra que el bilingüismo ofrece a los estudiantes
latinos una ventaja en términos de logros educativos.
Keywords: 1. migración internacional, 2. bilingüismo, 3. español, 4. latinos, 5. Estados
Unidos.
MIGRACIONES INTERNACIONALES, VOL. 2, NÚM. 4, JULIO-DICIEMBRE DE 2004
96 MIGRACIONES INTERNACIONALES
Fueled by high levels of immigration, Latino students have become a
substantial part of the U.S. school population. Thus, the educational
attainment of U.S. Latino students, which is considerably lower than
that of their European American and Asian American peers, is a press-
ing contemporary educational issue. This article addresses the effect
of dual language proficiency on the educational attainment of fully
English-proficient Latino students. The results indicate that among
Latinos who are fluent in English, Spanish retention, particularly in
the form of Spanish literacy, has a positive impact on high school
completion, entry into any college, and entry into an accredited
bachelor’s degree program.*
A great deal of research and attention has focused on the academic
problems facing students with limited English proficiency, such as low
levels of achievement, placement into lower grade levels or educational
tracks, and high dropout levels (Bean and Tienda, 1987; Clifton et al.,
1986; Donato, Menchaca, and Valencia, 1991; Ekstrom, Goertze, and
Pollack, 1986; Fligstein and Fernández, 1985; Ready, 1991; Roscigno,
Vélez, and Ainsworth-Darnell, 2001; Rumbaut, 1997; Warren, 1996).
However, little research has addressed the impact of the level of profi-
ciency in Spanish in addition to English proficiency on educational at-
tainment, even though depressed educational attainment for English-
speaking Latinos continues to be a concern.
Further, research on the impact of bilingualism on academic out-
comes has primarily focused on aspects of educational achievement,
such as achievement test scores and grades.1 The question of the effect
of language on educational attainment, addressed in this article, is sig-
nificant because educational credentials have a great impact on life chances
and social mobility. A traditional route to socioeconomic mobility in
the United States is through educational attainment or perseverance in
the educational system. Alternative routes, such as ethnic networks,
can link one to occupational opportunities in ethnic niches or enclaves
(Zhou, 1992). However, throughout much of the United States (with a
few exceptions, such as in Miami, see for example Portes and Manning,
* This research was supported by the American Educational Research Association
grants program which receives funds for its “AERA Grants Program” from the U.S. Depart-
ment of Education’s National Center for Education Statistics and the Office of Educa-
tional Research and Improvement, and the National Science Foundation under NSF Grant
#RED-9980573. It was also supported by the Spencer Foundation grant, Advanced Re-
search Studies #200100304. The opinions expressed here reflect those of the author and
do not necessarily reflect those of the granting agencies. I thank Richard Alba, Scott
South, Nancy Denton, Pamela Bennett, Karl Alexander, and the anonymous reviewers for
their helpful comments on previous versions of this article.
1 “Achievement” generally refers to a child’s competency in the material taught in
school, as measured by grade point average or achievement test scores. “Attainment” refers
to a child’s progression through the school system to higher levels of education and the
obtaining of educational credentials.
LANGUAGE AND EDUCATION OF BILITERATE LATINOS 97
1986), these networks operate by linking co-ethnics with occupations
in the mainstream economy, which may provide social mobility but
also typically require educational credentials.
Bilingual Proficiencies and Educational Outcomes
Scholarly research about the effect of the persistence of an ethnic mother
tongue for educational outcomes has generated mixed results. Increas-
ingly, researchers are finding that the maintenance of an ethnic mother
tongue in addition to English is associated with enhanced educational
achievement (Bankston and Zhou, 1995; Fernández and Nielson, 1986;
Nielson and Lerner, 1986; Peal and Lambert, 1962; Portes and
Rumbaut, 2001; Rumberger and Larson, 1998; Zhou and Bankston,
1998). They also associate language loss with the phenomenon of “sec-
ond-generation decline,” that is, lower academic performance of U.S.-
born students compared to their immigrant co-ethnics (Baral, 1979;
Gans, 1992; Baral, 1979; Matute-Bianchi, 1991; Suárez-Orozco,
1991;Valverde, 1987; Matute-Bianchi, 1986; Valverde, 1987; Zhou,
1997). However, Francis Evans and James Anderson (1973) find that
when achievement is measured using standardized tests, Mexican Ameri-
can students who speak Spanish at home score lower than both Anglos
and Mexican Americans who speak English at home, but when achieve-
ment is measured using course grades, the researchers find that home
language has no effect. Alejandro Portes and Rubén Rumbaut (1990),
using samples of school children in San Diego and Miami, find that
bilingualism is positively associated with higher course grades for Asian
and Cuban but not Mexican students. Jennifer Glick and Michael White
(2003) find that compared to living in an English monolingual house-
hold, living in a non-English-language household is associated with
lower test scores, while living in an English-dominant bilingual house-
hold is associated with higher test scores. Eugene Kennedy and Hae-
Seong Park (1994) find that speaking a language other than English at
home is positively associated with higher grades but negatively associ-
ated with standardized achievement scores for students from Asian eth-
nic groups. For Mexican-American children, however, language was
found to have no effect when socioeconomic background, hours of home-
work, self-concept, sense of control over life, and educational expecta-
tions were included in the regression models. In their research on Asian
students, Ted Mouw and Yu Xie (1999) find that bilingualism offers
no educational advantage per se. Rather, use of an immigrant mother
tongue enhances educational achievement only for those whose parents
do not speak English.
98 MIGRACIONES INTERNACIONALES
Regarding educational attainment, Charles Hirschman (1996) finds
that longer duration in the United States is associated with greater
school enrollment. However, in a later study utilizing the same data,
Hirschman (2001) finds that immigrant adolescents are at least as
likely as their native-born peers to be enrolled in high school, al-
though he does not distinguish between the language characteris-
tics of the immigrants and native born. White and Glick (2000)
find that students from bilingual homes are significantly more likely
to remain in high school until the senior year than are those from
homes where only one language is used, and among dropouts,
bilinguals are more likely to continue some other type of training.
However, in later work, Glick and White (2003) find some rela-
tively minor, but significant, differences in the effect of home lan-
guage by student cohorts across a ten-year period, although they
note that there are few differences across language backgrounds when
controlling for socioeconomic status.
Theoretical Perspectives on the Role
of Language in Educational Attainment
Four theoretical models address educational effects of language for Latino
children: assimilation, multiculturalism, the cognitive perspective, and
segmented assimilation. Assimilation theory posits that speaking “only
English” is positively associated with educational attainment.
Multicultural and cognitive models theorize that dual language profi-
ciency, rather than English monolingualism, is positively associated with
educational success. Segmented assimilation theory implies that the
effects of language on subsequent educational attainment vary by indi-
vidual, family, and neighborhood characteristics, but that the mainte-
nance of an ethnic mother tongue is generally associated with greater
academic performance.
Assimilation
Assimilation theory anticipates a “payoff from acculturation” for immi-
grants in American institutions (Logan and Alba, 1993:252), positively
associating English monolingualism with educational attainment. De-
spite variation in progression through the stages of acculturation, par-
ticularly regarding language (Fishman, 1972; Stevens, 1986), assimila-
tion theory assumes a transition over generations to only English, which
is generally considered to be a step toward increased acceptance and
LANGUAGE AND EDUCATION OF BILITERATE LATINOS 99
full participation in all areas of society, including the educational sys-
tem (Gans, 1979/1996). The assimilation approach posits knowledge
of English as the key to academic success in the U.S. educational con-
text, and educational success is linked to future career and socioeco-
nomic success.
Multiculturalism
Multiculturalism reflects a variety of theoretical perspectives united
by a vision of U.S. society as an array of various ethnic cultures.
Multiculturalism rejects the notion of assimilation to a Eurocentric
monoculture (Goldberg, 1994) or a unified core (Zhou, 1997), and
it values retention of an ethnic mother tongue, which it associates
with increased educational attainment and achievement. Horace
Kallen’s (1915/1996) early formulations of cultural pluralism con-
tained explicit references to the importance of literacy in an ethnic
mother tongue. More contemporary varieties of multiculturalism, such
as cosmopolitanism (Hollinger, 1995), associate fluency in foreign
languages with globalism and an acceptance of diversity. As fluency in
two languages is increasingly coveted in a global economy, instruc-
tion in non-English languages is coming to be seen as an important
part of educational curriculums (see U.S. Department of Education,
1998). When language skills are part of formal or informal academic
evaluations of students, bilingualism can help students toward the
route of higher education. As such, biliteracy may be an indicator of
cultural capital, subtlely conveying that the student possesses spe-
cialized knowledge associated with a well-rounded academic prepara-
tion (Bourdieu, 2000). In a cosmopolitan, multicultural context, a
student’s interest in languages may be viewed by teachers as a cul-
tural indicator of someone who will succeed academically and move
toward college. Based on their findings that studying a foreign lan-
guage in the ninth grade is a significant factor for placement into a
college-track program in high school, Karl Alexander and Martha Cook
theorize that “foreign language represents one of the few avenues for
implementing one’s preference for an academically rigorous program
of study” (1982:633). Because educators associate knowledge of two
languages with academically focused students who have high attain-
ment goals, bilingual students’ interest in language prior to high school
may make them more likely than other students to be placed in col-
lege-track programs when they enter high school. Thus, from a
multicultural perspective, retention of Spanish is hypothesized as hav-
ing a positive effect on educational outcomes.
100 MIGRACIONES INTERNACIONALES
Cognitive Perspectives
Several researchers argue that high-level bilingualism is beneficial to
cognitive development (Bain, 1974; Bialystok, 1988; Cummins, 1976;
Duncan and De Ávila, 1979; Peal and Lambert, 1962; Willig, 1985).
In this line of research, the cognitive advantages associated with bilin-
gualism are related to having two or more codes for every concept, al-
lowing for greater mental flexibility. As such, cognitive benefits associ-
ated with “multicompetence” (Cook, 1992) may extend into other
academic pursuits. István Kecskés (1998), for example, argues that in-
tensive learning of a second language can enhance development of the
first language (see also Kecskés and Papp, 2000). Building on Vygotsky’s
(1962) work, Kecskés argues that the study of grammar and writing is
“crucial for the mental development of the child because these two help
children rise to a higher level of speech development by making them
become aware of what they are doing with language” (1998:324).
Kecskés finds that the process of mental development associated with
gaining literacy in the first language extends to the study of a second
language, which, in turn, facilitates greater capacity in the first lan-
guage. Such enhanced mental development may give students an ad-
vantage in other academic subjects, ultimately leading biliterate stu-
dents to higher levels of educational attainment. Along similar lines,
Lily Wong Fillmore (1991) argues that loss of a first language among
young children can also result in parents’ inability to teach their chil-
dren lessons appropriate for their developmental stages.
Segmented Assimilation
Segmented assimilation theory suggests that there are three different
routes of incorporation available to children of immigrants: assimila-
tion into the middle class, assimilation into an “underclass,” or preser-
vation of an ethnic identity within the immigrant community (Portes
and Zhou, 1993; Portes and Rumbaut, 2001). This implies that the
effects of language on subsequent educational and occupational attain-
ment vary by individual, family, and neighborhood characteristics
(Portes, 1995; Rumbaut, 1996). What is novel about segmented as-
similation theory is the notion that assimilation can have negative con-
sequences for immigrant children because assimilation into an underclass
is seen to restrict social mobility for those children (Portes, 1995; Zhou
and Bankston, 1998).
Segmented assimilation theory focuses on the integration of first- and
second-generation children into inner-city cultural norms that are said
LANGUAGE AND EDUCATION OF BILITERATE LATINOS 101
to devalue educational success as a route to social mobility (Fernández-
Kelly, 1995; Fernández-Kelly, 1998; Portes and Zhou, 1993; Zhou,
1997). In this view, educational success is associated with “acting white,”
something their peers reject (Fordham and Ogbu, 1986), and the chil-
dren cannot reconcile parental expectations regarding education with
the desire to fit into their peer culture. They are, in short, “both more
aware of the existence of discrimination and more dubious about their
chances to overcome it” (Portes, 1995:265). Therefore, segmented as-
similation theorists believe that these children risk falling into an
underclass. This theory suggests that for poor and urban immigrants
and their descendants, shifting to “only English” results in blocked so-
cial mobility.
Segmented assimilation theory hypothesizes that ethnic families and
communities use language to influence younger members to strive for
educational success. Bilingualism in the younger generation can allow
parents to retain authority that can be lost without effective communi-
cation with their English-speaking children (García-Coll and Magnuson,
1997). Bilingualism can also allow elder family and community mem-
bers, who may not know English well, to exert social control aimed at
ensuring that children succeed academically and persevere in the edu-
cation system so they can live out their parents’ hopes for social mobil-
ity (Portes and Zhou, 1993; Portes, 1995; Zhou and Bankston, 1998).
Given the risks of assimilation into an urban underclass, the segmented
assimilation perspective seems to imply that retention of an ethnic lan-
guage has a protective function, which is more pronounced for those
living in poor and urban areas.
The segmented assimilation perspective emphasizes the importance
of social capital networks, maintained through the ethnic mother
tongue, in fostering students’ perseverance in the educational sys-
tem (Fernández-Kelly, 1998; Zhou and Bankston, 1996). Because
oral rather than written communication sustains social networks,
one might anticipate that attainment associated with dual language
skills would accrue to both those with substantial oral proficiency
and those who are literate in Spanish. In short, students with high
levels of oral proficiency and biliterate students should both benefit
from co-ethnic social pressures to persevere in their studies and to
enter college.
In summary, the cognitive, the multicultural, and the segmented as-
similation perspectives all anticipate that Spanish maintenance, in ad-
dition to English-language fluency, enhances educational attainment,
albeit based on different criteria. In contrast, the assimilation perspec-
tive anticipates that English monolingualism is associated with enhanced
educational attainment.
102 MIGRACIONES INTERNACIONALES
Data and Methods
The data used in this research are from the restricted-use version of the
National Education Longitudinal Survey (NELS:88). To date, NELS:88 is
the only longitudinal, nationally representative dataset that follows the
academic trajectories of youth from their pre-high school years through
their mid-twenties. It includes rich data on home language use and
multiple indicators of proficiency in English and Spanish, as well as
social, demographic, and education-related information. NELS:88 was
administered in 1988 to 25,000 eighth graders and to their parents,
teachers, and principals, and it provides individual, family, and school-
level data. Surveys were again administered to the same students in
1990, 1992, 1994, and 2000. Hispanic students were over-sampled.
The research reported on in this article used a sample of students who
remained in the study from 1988 to 1994. Those records were linked
to 1990 U.S. Census data, at the zip-code level, to provide contextual
data on the neighborhoods in which the children lived.
Dependent Variables
High School Completion, a dichotomous variable, indicates whether a
student had completed high school or the equivalent (GED or high school
certificate) by 1994 (completed=1/not completed=0).
College Enrollment, a dichotomous variable, measures whether a stu-
dent was enrolled in an associate’s degree or bachelor’s degree program,
based on the student’s self-report in 1994 (enrolled=1/not enrolled=0).
Bachelor’s Degree Program Enrollment, a dichotomous variable, mea-
sures whether a student was enrolled in an accredited bachelor’s degree
program based on the student’s self-report in 1994 (enrolled=1/not
enrolled=0).
Primary Independent Variable
Dual Language Proficiency variables reflect the student’s overall language
proficiency in speaking, understanding, reading, and writing in En-
glish and Spanish in the eighth grade. Four categories were created
from student’s responses to nine questions for self-rating of proficiency
in English and Spanish: English only, English dominant, oral bilin-
gual, and biliterate.2 Using these categories frees the research from what
2 For the English proficiency measures, students were asked four questions: How well do
you do the following? How well do you…a. Understand spoken English? b. Speak English?
LANGUAGE AND EDUCATION OF BILITERATE LATINOS 103
Alejandro Portes and Lingxin Hao have described as “a dualistic frame-
work [in much of the research on bilingualism] where foreign
monolingualism is opposed to full assimilation into English” (2002,
891). A typology for the various measures of language proficiency was
built from the student self-reports: English Only: an individual who
speaks, reads, and writes only in English. English Dominant: an indi-
vidual who has high proficiency in English, limited oral proficiency in
Spanish, and is not literate in Spanish. Oral Bilingual: an individual
who has high proficiency in English and high oral proficiency in Span-
ish, but who has limited reading and writing abilities in Spanish.
Biliterate: an individual who is both highly proficienct in English and
highly literate in Spanish.
The categories of “oral bilingual” and “biliterate,” which are similar
to the notion of folk versus fluent bilingualism (Portes and Rumbaut,
1996), are useful because they distinguish between different language-
skill dimensions, namely oral proficiency and literacy. Except for a very
few cases, those students classified as biliterate were also highly profi-
cient in oral Spanish. Because this research focuses on the impact of
Spanish retention in addition to high overall English-language profi-
ciency, those with limited English were omitted. Among English-pro-
ficient Latino students, 21.2% spoke only English, 20.7% were En-
glish dominant, 20.6% were oral bilinguals, and 37.5% were biliterate.
(See table 1 for means and standard deviations for all variables used in
the multivariate analysis.)
c. Read English? d. Write English? Students could respond very well, pretty well, well, or
not very well. Items were coded as follows: Very well=4, Pretty well=3, Well=2, Not very
well=1. In the NELS:88 survey, these questions were asked of students who responded “yes”
to the question “Is any language other than English spoken in your home?” The value as-
signed to the response “very well” is imputed for monolingual English speakers.
Spanish proficiency is measured using NELS:88 survey questions: To determine Spanish
proficiency, the following questions were used: With regard to THAT LANGUAGE, how well
do you do the following? How well do you… a. Understand that language when people
speak it? b. Speak that language? c. Read that language? d. Write that language? This time
students could respond very well, pretty well, well, not very well, or not at all. The re-
sponses are coded as follows: Very Well=4, Pretty Well=3, Well=2, Not very well=1, Not
at all=0. Students were asked these four questions if they answered “yes” to the screening
question, “Is any language other than English spoken in your home?” Only those who
responded “Spanish” to the follow-up question “What language, other than English, do
you currently use most often?” were selected for inclusion in this analysis. A score of “0”
reflecting a response of “not at all” on the home language proficiency measure was assigned
to those who indicated that only English is used in their home.
Those who report high overall English proficiency and some limited oral proficiency in
Spanish, but no reading or writing abilities are categorized as English dominant. Those
who report high overall English proficiency, high oral proficiency in Spanish, but low pro-
ficiency in reading and writing (less than “well”) in Spanish are categorized as bilingual.
Those who have high overall proficiency in English and can also read and write well in
Spanish (at least “well”) are categorized as biliterate. Those who report low levels of
English proficiency are not included in this analysis.
104 MIGRACIONES INTERNACIONALES
Table 1. Descriptive statistics for variables
used in the regression analysis (N = 1616).
Educational attainment
High school completion 0.816 0.016 0.387
College entry 0.383 0.016 0.486
Entry into bachelor’s degree program 0.200 0.013 0.400
Dual language proficiency
English dominant 0.207 0.014 0.405
Bilingual 0.206 0.017 0.405
Biliterate 0.375 0.022 0.484
Ethnicity
Cuban 0.044 0.012 0.205
Puerto Rican 0.104 0.014 0.306
Other Hispanic 0.193 0.016 0.395
Race
Black 0.044 0.007 0.204
White 0.603 0.021 0.489
Generation
Second generation 0.362 0.025 0.481
First generation 0.156 0.017 0.363
Unknown generation 0.128 0.014 0.334
Religion
Catholic 0.477 0.019 0.499
Sex
Male 0.651 0.019 0.477
Time spent on homework 6.454 0.188 5.417
Family variables
Single parent 0.173 0.014 0.373
Number of siblings 2.788 0.071 1.672
Socioeconomic status - 0 . 6 1 8 0.038 0.748
School type
Catholic school 0.056 0.012 0.229
Private school 0.018 0.008 0.134
Neighborhood and geographic variables
Percent urban 7 4 . 1 7 9 4.198 4 1 . 5 6 6
Percent Hispanic 4 2 . 6 4 9 2.734 3 1 . 2 8 1
Neighborhood median household income 2 7 . 2 7 8 0.719 1 1 . 2 0 1
Source: National Educational Longitudinal Survey.
Control Variables
Ethnicity is measured by a series of dummy variables based on the
student’s response to the question, “Which of these best categorizes
LANGUAGE AND EDUCATION OF BILITERATE LATINOS 105
your background?” There are four possible responses: (1) Mexican,
Mexican-American, or Chicano, (2) Cuban, (3) Puerto Rican, or (4)
Other Hispanic. “Mexican, Mexican-American, or Chicano” is the omit-
ted category.
Race is measured by a series of dummy variables based on the student’s
response to the question, “What is your race?” There are three possible
Latino responses: Black, White, and Other. “Other” is the omitted
category.
Gender is a dichotomous variable measured by the student’s self-re-
port of sex, either male or female (male=1/female=0).
Generation is measured by a series of dummy variables constructed
from questions on student and parent birthplaces for the base year
(eighth grade):3 first generation (children born outside of the United
States), second generation (U.S.-born children with at least one for-
eign-born parent), third and later generations (U.S.-born children of
native-born parents), and unknown generation. “Third and later gen-
erations” is the omitted category. “Unknown generation” is included
because nearly 13% of the English-proficient Latino sample is missing
data only on generation status. Although there are some indications
that the group missing generation-status data is comprised mostly of
first- and second-generation children, the evidence is not compelling
enough to warrant an imputation of data on this variable.
Religion is a dichotomous variable that indicates whether the student
is Catholic or belongs to another religion (Catholic=1/non-Catholic=0).
Retention of the Catholic religion is often an indicator of retention of
ethnic cultural ties, possibly including attendance at religious services
in Spanish (see also Alba, 1985; Byun, 1990).
Time Spent on Homework is a variable derived from two ordinal scales
indicating the child’s self report of the number of hours spent on home-
work per week, at home and school. The scales were recoded to their
midpoints and added together.
Single-Parent Status is a dichotomous variable which indicates if the
child’s parents are not married or married (or in a marriage-like rela-
tionship): Not married=1/Married=0.
Number of Siblings is a variable measuring the student’s or parent’s
base-year report of the number of brothers and sisters the student has,
including any stepbrothers or stepsisters who live in the same home.
Values range from “0” to “6 and above.”
Socioeconomic Status is measured using a standardized scale created by
the National Center for Education Statistics (NCES), which includes par-
ents’ income, occupation, and education. The range for the variable is -
2.97 through 2.56.
3 See Oropesa and Landale (1997) for a similar identification of immigrant generation.
106 MIGRACIONES INTERNACIONALES
School Type is a series of dichotomous variables indicating whether the
student attends a public, Catholic, or private high school.
Neighborhood Context is measured by three variables created by link-
ing the NELS:88 data to 1990 U.S. Census data using zip-code identifi-
ers: average household income, urbanism (percentage of the zip-code
area defined by the Census as urban), and percentage of residents who
are Hispanic.4
Analytic Strategy
Selecting only those students who speak English well (N=1616) re-
gardless of Spanish ability, descriptive statistics and logistic regression
were used to examine the impact of various dual-language proficiencies
(and the control variables) on the likelihood that students will graduate
from high school (or receive a GED), attend any college, or attend a
bachelor’s degree program. A distinction is drawn between “any college
program” (community college two-year programs and four-year pro-
grams) and an accredited bachelor’s degree program because the latter
generally has more rigorous admission’s criteria. A series of regression
models were estimated for each of the dependent variables in order to
determine whether the various language abilities have an effect over and
above the control variables. Given that the segmented assimilation per-
spective anticipates variation in outcomes based on ethnicity and settle-
ment context, I also included interaction terms to test for variation in
the impact of language by ethnicity, socioeconomic status, and neigh-
borhood characteristics. All analyses were weighted and the standard
errors were adjusted for clustering, using AM Statistical Software.5
Findings
High school completion was highest for biliterate students (86.1%)
and English monolinguals (85%) (table 2), and it was somewhat lower
for English-dominant (80%) and oral-bilingual (74%) students. En-
rollment in a college program was also highest for biliterate (41%) and
English-dominant (40%) students, followed closely by oral bilingual
students (39%), and then by English monolinguals (33%). Biliterates
had higher levels of enrollment in a bachelor’s degree programs—about
one-quarter were enrolled—compared to slightly less than one-fifth of
the students from the other high-English-proficiency groups.
4 I use the term “Hispanic” in reference to the NELS:88 variables because this is the
term used in the survey. Elsewhere, I used the term “Latino,” as it is the preferred term
among those of Latin American ancestry in the United States (Oboler, 1992).
5 More information is available on this software at the American Institutes for Research
web site, http://am.air.org.
LANGUAGE AND EDUCATION OF BILITERATE LATINOS 107
Table 2. Effect of language use on educational attainment.
English English
Only Dominant Bilingual Biliterate
Dropped out (at any time) 23.0% 30.1% 39.3% 24.8%
Completed high school 85.0% 80.0% 73.8% 86.1%
Entered an associate’s or bachelor’s
degree program 33.2% 39.9% 38.8% 41.4%
Entered a bachelor’s degree program 17.6% 17.9% 19.30% 24.3%
Source: National Educational Longitudinal Survey.
Multivariate Analysis
The results of five logistic regression models address the impact of dual
language proficiency on high school completion (table 3). These re-
gression models include only Latino students who are English profi-
cient and who may or may not have additional proficiency in Spanish.
At the bivariate level, no significant association between dual language
proficiency and high school completion appears to exist. The baseline
model (Model 1) includes individual-level characteristics, while the other
models add to the baseline family structure (Model 2), socioeconomic
status (Model 3), school type (Model 4), and neighborhood character-
istics (Model 5). Models 1 and 2 indicate that dual language profi-
ciency has no significant impact on high school graduation, but there
are significant effects of ethnicity, race, and family structure. However,
when controlling for family socioeconomic status (Model 3), biliterate
students appear to be significantly more likely to complete high school
compared to those who speak English only. However, in terms of com-
pleting high school, those with oral proficiency in Spanish (English-
dominant and bilingual students) do not appear to be significantly
different from those who speak only English. Net of the effects of the
control variables in Model 3, biliterate students are more than twice as
likely to complete high school compared to those who speak only En-
glish (2.32=exp(.843)). When socioeconomic status is entered into the
model (see Model 3), the effects of family structure fall from signifi-
cance, and the effect of race falls just short of significance, although it
reemerges in Models 4 and 5. It is not entirely surprising that biliteracy
emerges as a significant predictor of high school completion once socio-
economic status is controlled. Vincent Roscigno, María Vélez, and James
Ainsworth-Darnell note that language-minority students are about
“50% more likely to live in poverty,” and overall, their research seems
to indicate that “the obstacles language-minority students face have as
Ta
bl
e
3.
E
ff
ec
ts
o
f D
ua
l L
an
gu
ag
e
P
ro
fi
ci
en
ci
es
o
n
H
ig
h
S
ch
oo
l C
om
pl
et
io
n
.
B
iv
ar
ia
te
M
od
el
1
M
od
el
2
M
od
el
3
M
od
el
4
M
od
el
5
E
st
i-
E
st
i-
E
st
i-
E
st
i-
E
st
i-
E
st
i-
m
at
e
S
E
m
at
e
S
E
m
at
e
S
E
m
at
e
S
E
m
at
e
S
E
m
at
e
S
E
D
ua
l l
an
gu
ag
e
pr
of
ic
ie
nc
y
E
n
gl
is
h
o
n
ly
(
om
it
te
d
)
E
n
gl
is
h
d
om
in
an
t
-0
.5
66
0.
30
3
0.
04
5
0.
29
3
0.
03
6
0.
30
6
0.
18
9
0.
31
3
0.
14
8
0.
29
9
0.
15
3
0.
30
7
B
il
in
gu
al
0.
23
7
0.
28
0
-0
.2
98
0.
26
3
-0
.2
69
0.
25
8
0.
01
1
0.
27
6
-0
.0
43
0.
27
5
-0
.0
05
0.
27
6
B
il
it
er
at
e
0.
82
1*
0.
40
2
0.
56
1
0.
29
5
0.
56
6
0.
29
5
0.
84
3*
*
0.
29
0
0.
81
1*
*
0.
28
3
0.
84
6*
*
0.
29
2
E
th
ni
ci
ty
M
ex
ic
an
(
om
it
te
d
)
C
ub
an
-0
.2
47
0.
26
3
0.
63
2
0.
40
3
0.
48
3
0.
39
2
0.
00
3
0.
38
8
0.
02
3
0.
41
3
0.
03
0
0.
40
8
P
u
er
to
R
ic
an
1
.3
8
8
**
*
0.
40
6
-0
.1
93
0.
24
8
-0
.1
68
0.
24
1
-0
.3
17
0.
27
3
-0
.4
70
0.
26
9
-0
.3
88
0.
30
6
O
th
er
H
is
p
an
ic
-0
.0
29
0.
49
5
1
.4
3
9
**
*
0.
41
7
1
.4
4
8
**
*
0.
42
6
1.
24
2*
*
0.
43
2
1.
21
6*
*
0.
43
4
1.
22
2*
*
0.
44
7
R
ac
e
B
la
ck
0.
65
9*
*
0.
23
5
-0
.0
78
0.
49
9
-0
.0
09
0.
46
6
0.
06
0
0.
44
4
0.
04
9
0.
44
6
0.
07
6
0.
45
7
W
h
it
e
-0
.0
37
0.
26
5
0.
63
0*
*
0.
24
5
0.
57
9*
0.
24
4
0.
48
5
0.
25
3
0.
54
7*
0.
25
5
0.
52
9
*
0.
25
0
O
th
er
r
ac
e
(o
m
it
te
d
)
G
en
er
at
io
n
T
h
ir
d
g
en
er
at
io
n
(
om
it
te
d
)
Se
co
n
d
g
en
er
at
io
n
-0
.1
0
4
0.
32
4
-0
.2
07
0.
26
3
-0
.1
49
0.
25
3
0.
12
7
0.
26
8
0.
18
1
0.
27
3
0.
16
5
0.
26
9
F
ir
st
g
en
er
at
io
n
-0
.7
56
**
0.
28
9
-0
.2
80
0.
32
5
-0
.1
78
0.
33
0
0.
23
9
0.
33
8
0.
31
9
0.
34
2
0.
34
6
0.
33
6
U
n
kn
ow
n
g
en
er
at
io
n
0.
07
0.
21
3
-0
.5
21
0.
31
1
-0
.5
86
0.
31
0
-0
.3
66
0.
32
8
-0
.3
89
0.
30
8
-0
.4
17
0.
32
0
Se
x
M
al
e
0.
38
6
0.
22
3
0.
14
6
0.
20
9
0.
11
5
0.
21
6
0.
08
7
0.
21
5
0.
04
1
0.
21
0
0.
02
5
0.
21
4
T
ab
le
3
. C
on
ti
n
ua
ti
on
.
B
iv
ar
ia
te
M
od
el
1
M
od
el
2
M
od
el
3
M
od
el
4
M
od
el
5
E
st
i-
E
st
i-
E
st
i-
E
st
i-
E
st
i-
E
st
i-
m
at
e
S
E
m
at
e
S
E
m
at
e
S
E
m
at
e
S
E
m
at
e
S
E
m
at
e
S
E
R
el
ig
io
n
C
at
h
ol
ic
0.
03
3
0.
02
7
0.
39
2
0.
23
6
0.
41
6
0.
23
7
0.
37
3
0.
25
3
0.
25
6
0.
24
5
0.
23
8
0.
25
4
T
im
e
sp
en
t
on
h
om
ew
or
k
-0
.5
6
1
*
0.
25
5
0.
03
4
0.
02
6
0.
03
4
0.
02
7
0.
01
7
0.
02
9
0.
01
3
0.
02
8
0.
01
2
0.
02
8
F
am
il
y
va
ri
ab
le
s
Si
n
gl
e
p
ar
en
t
-0
.1
64
**
0.
05
2
-0
.6
43
**
0.
25
0
-0
.3
73
0.
25
3
-0
.3
75
0.
25
3
-0
.3
39
0.
25
4
N
u
m
b
er
o
f
si
b
li
n
gs
1.
02
6*
**
0.
16
9
-0
.1
38
**
0.
05
2
-0
.0
67
0.
05
6
-0
.0
71
0.
05
6
-0
.0
67
0.
05
5
So
ci
oe
co
n
om
ic
s
ta
tu
s
2.
60
2*
1.
03
6
0.
96
9*
**
0.
17
8
0.
99
1*
**
0.
17
6
0.
95
6*
**
0.
17
1
Sc
ho
ol
t
yp
e
P
u
b
li
c
sc
h
oo
l
(o
m
it
te
d
)
2.
04
1
1.
08
1
2.
04
0
1.
07
9
C
at
h
ol
ic
s
ch
oo
l
-0
.8
20
0.
94
-2
.1
5
7
*
0.
88
3
-2
.2
04
**
0.
82
9
P
ri
va
te
s
ch
oo
l
0.
00
0
0.
00
2
N
ei
gh
bo
rh
oo
d
an
d
ge
og
ra
ph
y
va
ri
ab
le
s
P
er
ce
n
t
u
rb
an
-0
.0
05
0.
00
3
-0
.0
03
0.
00
2
P
er
ce
n
t
H
is
p
an
ic
0.
03
4*
**
0.
01
0.
00
1
0.
00
3
N
ei
gh
b
or
h
oo
d
m
ed
ia
n
h
ou
se
h
ol
d
i
n
co
m
e
0.
01
7
0.
01
3
C
on
st
an
t
0.
50
3
0.
38
5
1.
03
1*
*
0.
46
6
1.
40
5
*
0.
55
4
1.
55
7*
**
0.
43
8
1.
24
2*
0.
52
8
L
og
L
ik
el
ih
oo
d
-7
0
8
.2
0
0
-6
9
5
.4
1
4
-6
6
3
.9
6
2
-6
4
9
.0
3
4
-6
4
6
.4
3
3
So
ur
ce
: N
at
io
na
l E
du
ca
tio
na
l L
on
gi
tu
di
na
l S
ur
ve
y (
N
=
1
61
6)
.
**
*p
<
.0
01
, *
*p
<
.0
1,
*p
<
.0
5
(s
ta
nd
ar
d
er
ro
rs
ar
e r
ob
us
t)
.
Ta
bl
e
4.
E
ff
ec
ts
o
f D
ua
l L
an
gu
ag
e
P
ro
fi
ci
en
ci
es
o
n
C
ol
le
ge
E
n
tr
y.
B
iv
ar
ia
te
M
od
el
1
M
od
el
2
M
od
el
3
M
od
el
4
M
od
el
5
E
st
i-
E
st
i-
E
st
i-
E
st
i-
E
st
i-
E
st
i-
m
at
e
S
E
m
at
e
S
E
m
at
e
S
E
m
at
e
S
E
m
at
e
S
E
m
at
e
S
E
D
ua
l l
an
gu
ag
e
pr
of
ic
ie
nc
y
E
n
gl
is
h
o
n
ly
(
om
it
te
d
)
E
n
gl
is
h
d
om
in
an
t
0.
42
6*
0.
21
4
0.
60
6*
0.
24
7
0.
62
1*
0.
24
5
0.
70
0*
*
0.
24
9
0.
67
1*
*
0.
24
6
0.
64
5*
*
0.
24
4
B
ili
n
gu
al
0.
36
2*
0.
18
2
0.
55
0*
*
0.
20
9
0.
59
7*
**
0.
21
4
0.
78
6*
**
0.
21
6
0.
74
1*
**
0.
21
8
0.
72
7*
**
0.
22
0
B
il
it
er
at
e
0.
44
9*
0.
18
5
0.
59
2*
*
0.
22
5
0.
60
8*
**
0.
22
8
0.
79
5*
**
0.
22
6
0.
80
7*
**
0.
21
8
0.
78
1*
**
0.
22
5
E
th
ni
ci
ty
M
ex
ic
an
(
om
it
te
d
)
C
ub
an
0.
94
3*
*
0.
36
3
0.
86
7*
0.
35
9
0.
67
9
0.
37
3
0.
29
7
0.
36
2
0.
14
0
0.
35
6
0.
15
4
0.
36
4
P
u
er
to
R
ic
an
0.
09
5
0.
25
6
0.
20
0
0.
23
0
0.
17
2
0.
23
3
-0
.0
01
0.
22
1
-0
.1
70
0.
22
1
-0
.1
02
0.
23
0
O
th
er
H
is
p
an
ic
0.
76
8*
**
0.
18
7
1.
01
5*
**
0.
19
4
0.
96
4*
**
0.
20
1
0.
72
0*
**
0.
21
3
0.
69
6*
**
0.
19
7
0.
74
1*
**
0.
20
3
R
ac
e
B
la
ck
-0
.0
34
0.
34
8
0.
05
5
0.
35
8
0.
08
1
0.
36
2
0.
16
3
0.
35
1
0.
17
0
0.
35
3
0.
21
6
0.
35
5
W
h
it
e
0.
47
7*
**
0.
16
2
0.
51
5*
*
0.
18
1
0.
46
0*
0.
18
5
0.
34
4
0.
19
3
0.
39
9*
0.
19
1
0.
37
7*
0.
19
2
O
th
er
r
ac
e
(o
m
it
te
d
)
G
en
er
at
io
n
T
h
ir
d
g
en
er
at
io
n
(
om
it
te
d
)
Se
co
n
d
g
en
er
at
io
n
0.
17
1
0.
17
7
-0
.0
69
0.
19
3
-0
.0
21
0.
18
3
0.
17
7
0.
18
6
0.
22
0
0.
18
1
0.
14
2
0.
18
1
F
ir
st
g
en
er
at
io
n
0.
24
8
0.
18
8
0.
11
8
0.
20
3
0.
24
7
0.
20
2
0.
50
6*
0.
21
2
0.
58
3*
*
0.
20
8
0.
50
8*
0.
21
0
U
n
kn
ow
n
g
en
er
at
io
n
-0
.2
4
0
.2
2
9
-0
.0
1
5
0.
28
3
-0
.0
25
0.
28
1
0.
14
6
0.
27
8
0.
07
9
0.
26
5
-0
.0
18
0.
26
8
Se
x
M
al
e
0.
05
2
0.
14
1
0.
04
5
0.
13
6
0.
01
7
0.
13
5
-0
.0
10
0.
14
1
-0
.0
45
0.
13
9
-0
.0
42
0.
13
9
Ta
bl
e
4.
C
on
ti
n
ua
ti
on
.
B
iv
ar
ia
te
M
od
el
1
M
od
el
2
M
od
el
3
M
od
el
4
M
od
el
5
E
st
i-
E
st
i-
E
st
i-
E
st
i-
E
st
i-
E
st
i-
m
at
e
S
E
m
at
e
S
E
m
at
e
S
E
m
at
e
S
E
m
at
e
S
E
m
at
e
S
E
R
el
ig
io
n
C
at
h
ol
ic
0.
46
0*
*
0.
16
7
0.
49
0*
0.
21
8
0.
48
8*
0.
21
6
0.
42
5
0.
22
4
0.
30
9
0.
21
7
0.
25
8
0.
21
6
T
im
e
sp
en
t
on
h
om
ew
or
k
0.
05
1*
**
0.
01
2
0.
05
5*
**
0.
01
2
0.
05
5*
**
0.
01
2
0.
04
3*
**
0.
01
2
0.
04
1*
**
0.
01
2
0.
04
1*
**
0.
01
3
F
am
il
y
va
ri
ab
le
s
Si
n
gl
e
p
ar
en
t
-0
.3
2
2
0
.1
8
4
-0
.3
3
6
0.
19
0
-0
.1
36
0.
18
4
-0
.1
29
0.
18
4
-0
.1
24
0.
18
6
N
u
m
b
er
o
f
si
b
li
n
gs
-0
.1
64
**
*
0.
04
2
-0
.1
53
**
*
0.
04
1
-0
.1
03
*
0.
04
1
-0
.1
07
**
0.
04
1
-0
.1
03
*
0.
04
1
So
ci
oe
co
n
om
ic
s
ta
tu
s
0.
69
7*
**
0.
09
5
0.
63
4*
**
0.
10
0
0.
59
1*
**
0.
10
4
0.
55
6*
**
0.
10
8
Sc
ho
ol
t
yp
e
P
u
b
li
c
sc
h
oo
l
(o
m
it
te
d
)
C
at
h
ol
ic
s
ch
oo
l
2.
14
9*
**
0.
26
9
1.
71
8*
**
0.
29
9
1.
65
5*
**
0.
29
8
P
ri
va
te
s
ch
oo
l
-0
.5
5
7
0.
74
6
-1
.3
21
0.
96
2
-1
.5
09
1.
04
3
N
ei
gh
bo
rh
oo
d
an
d
ge
og
ra
ph
y
va
ri
ab
le
s
Pe
rc
en
t
u
rb
an
0.
00
5*
*
0.
00
2
0.
00
1
0.
00
2
Pe
rc
en
t
H
is
p
an
ic
-0
.0
02
0.
00
2
0.
00
4
0.
00
3
N
ei
gh
b
or
h
oo
d
m
ed
ia
n
h
ou
se
h
ol
d
i
n
co
m
e
0.
02
8*
**
0.
00
7
0.
01
4
0.
00
7
C
on
st
an
t
-2
.2
31
**
*
0.
28
5
-1
.7
42
**
*
0.
33
0
-1
.5
14
**
*
0.
33
7
-1
.4
96
**
*
0.
33
1
-2
.0
52
**
*
0.
40
7
L
og
L
ik
el
ih
oo
d
-1
0
0
6
.5
8
0
-9
9
3
.9
0
0
-9
6
8
.3
3
4
-9
4
3
.4
9
7
-9
4
0
.3
1
0
So
ur
ce
: N
at
io
na
l E
du
ca
tio
na
l L
on
gi
tu
di
na
l S
ur
ve
y (
N
=
1
61
6)
.
**
*p
<
.0
01
, *
*p
<
.0
1,
*p
<
.0
5
(s
ta
nd
ar
d
er
ro
rs
ar
e r
ob
us
t)
.
112 MIGRACIONES INTERNACIONALES
much or more to do with their family background status than with the
languages they speak” (2001:70; see also Bean and Tienda, 1987;
Fernández, Paulsen, and Hirano-Nakanishi, 1989; Portes and Truelove,
1987; Valenzuela and Dornbusch, 1994; Vélez, 1989; Warren, 1996).
The strong and significant positive effect of biliteracy on high school
completion, net of the effects of the control variables, persists when
school type and neighborhood characteristics are included in Models 4
and 5. Thus, in terms of high school completion, when controlling for
socioeconomic status, biliterate students appear to have an advantage
over students who are proficient in English only.
The effect of dual language proficiency on college entry (defined here as
entry into an associate’s degree or bachelor’s degree program) appears to
be even stronger. The results of five logistic-regression models address
this (table 4). At the bivariate level, dual language proficiency is signifi-
cantly associated with college entry. In the five models, all Spanish-main-
taining groups—including English-dominant, oral bilingual, and
biliterate students—have a significantly greater likelihood of entering
college than do those students who speak only English, net of the effects
of the control variables. Similar to the previous case, the coefficients for
all the Spanish-maintaining groups increased when socioeconomic status
was entered as a control variable. Net of the effect of the control variables
in Model 3—including socioeconomic status—compared to those who
speak only English, English-dominant students are 2.01 times as likely
to enter college (2.01=exp(.700)), oral bilingual students are 2.19 times
as likely to enter college (2.19=exp(.786)), and biliterate students are
2.21 times as likely to enter college (2.21=exp(.795)). When all the con-
trol variables are included in the model, as in Model 5, English-domi-
nant Latinos are 1.91 times as likely to enter college (1.91=exp(.645)),
oral bilinguals are 2.07 times as likely to enter college (2.07=exp(.727)),
and biliterates are 2.18 times as likely to enter college, compared to those
who speak only English (2.18=exp(.781)). Thus, in terms of college en-
try, biliteracy and oral proficiency in Spanish (including limited oral pro-
ficiency) in addition to English appear to give Latino students an advan-
tage relative to English monolingualism.
When we examine entry into a bachelor’s degree program as opposed
to entry into any college, a somewhat different story emerges (table 5).
At the bivariate level, biliteracy is significantly associated with entry
into a bachelor’s degree program, but neither of the Spanish oral-profi-
ciency categories, English-dominant or oral bilingual, is significant at
the bivariate level. When individual-level controls are included (Model
1), biliterate students are significantly more likely to enter bachelor’s
degree programs compared to their peers who speak only English. The
effect of biliteracy remains positive and significant, net of the effects of
Ta
bl
e
5.
E
ff
ec
ts
o
f D
u
al
L
an
gu
ag
e
P
ro
fi
ci
en
ci
es
on
E
n
tr
y
in
to
a
B
ac
h
el
or
’s
D
eg
re
e
P
ro
gr
am
.
B
iv
ar
ia
te
M
od
el
1
M
od
el
2
M
od
el
3
M
od
el
4
M
od
el
5
E
st
i-
E
st
i-
E
st
i-
E
st
i-
E
st
i-
E
st
i-
m
at
e
S
E
m
at
e
S
E
m
at
e
S
E
m
at
e
S
E
m
at
e
S
E
m
at
e
S
E
D
ua
l l
an
gu
ag
e
pr
of
ic
ie
nc
y
E
n
gl
is
h
o
n
ly
(
om
it
te
d
)
E
n
gl
is
h
d
om
in
an
t
0.
20
8
0.
24
4
0.
40
7
0.
28
3
0.
42
4
0.
28
1
0.
49
4
0.
29
0
0.
44
1
0.
28
4
0.
39
5
0.
28
4
B
il
in
gu
al
0.
09
4
0.
23
5
0.
32
2
0.
29
2
0.
36
6
0.
30
5
0.
57
2
0.
30
7
0.
49
1
0.
29
8
0.
43
6
0.
30
0
B
il
it
er
at
e
0.
45
9*
0.
21
2
0.
74
2*
*
0.
25
9
0.
7
5
5*
*
0.
26
1
0.
99
4*
**
0.
26
8
0.
98
9*
**
0.
26
7
0.
92
3*
**
0.
27
2
E
th
ni
ci
ty
M
ex
ic
an
(
om
it
te
d
)
C
ub
an
0.
67
7*
0.
31
6
0.
56
9*
0.
28
8
0.
37
5
0.
30
3
-0
.1
60
0.
27
8
-0
.3
61
0.
31
8
-0
.2
95
0.
33
3
P
u
er
to
R
ic
an
0.
24
2
0.
31
7
0.
31
0
0.
28
3
0.
29
4
0.
28
8
0.
01
8
0.
28
9
-0
.1
95
0.
28
0
0.
02
1
0.
28
3
O
th
er
H
is
p
an
ic
0.
73
6*
**
0.
19
5
0.
95
0*
**
0.
20
9
0.
8
94
**
*
0.
20
9
0.
54
3*
0.
23
7
0.
48
4*
0.
23
0
0.
54
6*
0.
23
3
R
ac
e
B
la
ck
-0
.1
78
0.
47
0
-0
.1
27
0.
45
8
-0
.0
93
0.
46
0
0.
02
7
0.
46
9
0.
01
6
0.
43
0
0.
04
1
0.
43
2
W
h
it
e
0.
38
7*
0.
18
0
0.
3
5
2
**
0.
19
5
0.
28
6
0.
19
9
0.
09
9
0.
21
6
0.
12
3
0.
21
4
0.
12
6
0.
21
5
O
th
er
r
ac
e
(o
m
it
te
d
)
G
en
er
at
io
n
T
h
ir
d
g
en
er
at
io
n
(
om
it
te
d
)
Se
co
n
d
g
en
er
at
io
n
0.
14
3
0.
18
9
-0
.1
40
0.
18
3
-0
.0
97
0.
18
1
0.
13
6
0.
18
2
0.
13
9
0.
18
1
0.
17
3
0.
18
8
F
ir
st
g
en
er
at
io
n
-0
.0
90
0
.2
4
0
-0
.3
2
3
0.
28
0
-0
.1
95
0.
27
8
0.
07
6
0.
30
8
0.
11
2
0.
31
4
0.
14
2
0.
30
6
U
n
kn
ow
n
g
en
er
at
io
n
-0
.3
9
0
0
.2
4
6
-0
.2
4
8
0.
29
4
-0
.2
69
0.
29
5
-0
.0
51
0.
29
3
-0
.1
39
0.
28
9
-0
.1
21
0.
30
2
Se
x
M
al
e
0.
07
7
0.
15
6
0.
12
4
0.
16
2
0.
09
9
0.
16
2
0.
06
2
0.
16
5
0.
03
2
0.
16
4
0.
01
0
0.
16
7
Ta
bl
e
5.
C
on
ti
n
u
at
io
n
.
B
iv
ar
ia
te
M
od
el
1
M
od
el
2
M
od
el
3
M
od
el
4
M
od
el
5
E
st
i-
E
st
i-
E
st
i-
E
st
i-
E
st
i-
E
st
i-
m
at
e
S
E
m
at
e
S
E
m
at
e
S
E
m
at
e
S
E
m
at
e
S
E
m
at
e
S
E
R
el
ig
io
n
C
at
h
ol
ic
0.
34
7
0.
19
1
0.
33
7
0.
24
1
0.
32
8
0.
24
3
0.
23
7
0.
24
8
0.
13
6
0.
24
8
0.
13
7
0.
25
1
T
im
e
sp
en
t
on
h
om
ew
or
k
0.
06
9*
**
0.
01
2
0.
07
3*
**
0.
01
2
0.
07
2*
**
0.
01
2
0.
05
8*
**
0.
01
3
0.
05
8*
**
0.
01
3
0.
05
8*
**
0.
01
2
F
am
il
y
va
ri
ab
le
s
Si
n
gl
e
p
ar
en
t
-0
.3
9
4
0
.2
2
7
-0
.4
2
4
0.
25
9
-0
.1
89
0.
27
2
-0
.2
03
0.
28
4
-0
.1
95
0.
28
4
N
u
m
b
er
o
f
si
b
li
n
gs
-0
.1
8
0
**
*
0.
04
8
-0
.1
56
**
*
0.
04
6
-0
.0
87
0.
04
5
-0
.0
90
0.
04
6
-0
.0
90
0.
04
6
So
ci
oe
co
n
om
ic
s
ta
tu
s
0.
86
5*
**
0.
11
7
0.
80
7*
**
0.
13
2
0.
75
1*
**
0.
13
9
0.
80
3*
**
0.
14
3
Sc
ho
ol
t
yp
e
P
u
b
li
c
sc
h
oo
l
(o
m
it
te
d
)
C
at
h
ol
ic
s
ch
oo
l
1
.5
5
1*
**
0.
40
6
1.
05
7*
*
0.
38
8
1.
12
0*
*
0.
37
7
P
ri
va
te
s
ch
oo
l
0.
17
0
0.
70
4
-0
.4
42
0.
89
3
-0
.3
92
0.
86
5
N
ei
gh
bo
rh
oo
d
an
d
ge
og
ra
ph
y
va
ri
ab
le
s
P
er
ce
n
t
u
rb
an
0.
00
1
0.
00
2
-0
.0
02
0.
00
2
P
er
ce
n
t
H
is
p
an
ic
-0
.0
03
0.
00
3
0.
00
3
0.
00
3
N
ei
gh
b
or
h
oo
d
m
ed
ia
n
h
ou
se
h
ol
d
i
n
co
m
e
0.
01
9
*
0.
00
7
-0
.0
03
0.
01
0
C
on
st
an
t
-2
.2
9
2*
**
0.
34
4
-2
.4
81
**
*
0.
37
2
-2
.1
84
**
*
0.
38
1
-2
.1
33
**
*
0.
37
7
-2
.0
33
**
*
0.
45
7
L
og
L
ik
el
ih
oo
d
-7
5
5
.4
5
3
-7
4
5
.9
7
4
-7
1
5
.6
9
6
-7
0
6
.1
5
1
-7
0
4
.2
6
9
So
ur
ce
: N
at
io
na
l E
du
ca
tio
na
l L
on
gi
tu
di
na
l S
ur
ve
y (
N
=
1
61
6)
.
**
*p
<
.0
01
, *
*p
<
.0
1,
*p
<
.0
5
(s
ta
nd
ar
d
er
ro
rs
ar
e r
ob
us
t)
.
LANGUAGE AND EDUCATION OF BILITERATE LATINOS 115
family structure, socioeconomic status, school type, and neighborhood
characteristics. Net of the effects of the other variables (Model 5),
biliterate students are more than two-and-a-half times as likely to enter
a bachelor’s degree program than are students who are proficient in
English only (2.52=exp(.923)), whereas students who are English-domi-
nant or oral bilinguals (both categories indicating oral proficiency in
Spanish, but not literacy) are not significantly different in their likeli-
hood of entering a bachelor’s degree program than are those who speak
only English.
In additional analysis, not shown here, multiplicative terms were added
to the models in tables 3, 4, and 5 to determine whether dual language
proficiency interacts with socioeconomic status, ethnicity, and urban-
ism to produce significant differences in the likelihood of high school
completion, college entry, or entry into a bachelor’s degree program, as
suggested by the segmented assimilation perspective. However, these
interaction terms did not add significantly to the models.
Discussion
Although Latino students face barriers to educational attainment com-
pared to the majority group, the relative success of biliterate students,
compared to those who speak only English, does not support the hy-
pothesis suggested by the assimilation perspective, which would pre-
dict that “only English” offers the greatest educational advantage. In-
stead, in terms of educational attainment, the ability to read and write
in both English and Spanish appears to offer biliterate students a sig-
nificant advantage over their peers who are proficient only in English.
Net of the effects of the controls, particularly socioeconomic status,
biliterate students are significantly more likely to complete high school
compared to their monolingual peers, whereas those with oral profi-
ciency in Spanish (whether limited or substantial) in addition to En-
glish did not have significantly different outcomes from those who spoke
only English. Biliterate students are also significantly more likely to
enter college, and particularly, a bachelor’s degree program, than are
their peers who speak only English.
In terms of entry into any college program, oral proficiency in Span-
ish (whether limited or substantial) provides students with a greater
likelihood of college entry compared to their English monolingual peers.
Perhaps social networks, maintained in Spanish within families or neigh-
borhoods, are conducive to communicating the importance of college
entry to students. However, in terms of entering a bachelor’s degree
program, as opposed to any college program, biliteracy, but not oral
116 MIGRACIONES INTERNACIONALES
proficiency, offers Latino students a significant advantage relative to
their English-monolingual co-ethnics.
Overall, this research indicates that bilingualism, particularly in the
form of biliteracy, offers Latino students an advantage in terms of educa-
tional attainment. These findings suggest that compared to the assimila-
tion perspective, greater explanatory power exists in theoretical mod-
els—such as the cognitive, multicultural, or segmented assimilation
perspectives—that anticipate that maintenance of an ethnic language in
addition to English enhances educational attainment. The findings in
this article cannot definitively reveal the ways in which language operates
to enhance educational attainment (that is, through social control of co-
ethnics, cultural capital that places students on the college track, or cogni-
tive advantages related to learning two language systems). However, the
effect of oral proficiency on college entry hints that communication among
co-ethnics may be relevant to promoting attainment aspirations among
students, as suggested by the segmented assimilation theory. Neverthe-
less, oral proficiency alone is not significant in terms of students entering
a bachelor’s degree program, which suggests that efforts among co-ethnics
to communicate attainment aspirations are not enough to significantly
enhance Latino students’ participation in those programs. The signifi-
cance of biliteracy for entrance into bachelor’s degree programs may sug-
gest that biliterate students gain greater attainment than do students
who do not maintain Spanish through higher levels of achievement, as
suggested by the cognitive perspective, or because of a greater level of
participation in college-track curricula, as suggested by the multicultural
perspective.
Future Research
To better distinguish among the aspects of language emphasized in
each of the theoretical perspectives, further research might look at not
only whether bilingual proficiency offers students an advantage in terms
of educational attainment but also the nature of that advantage. Al-
though the research reported here indicates that biliterate students,
proficient in both English and Spanish, have a significant educational
advantage over those who are proficient in English only, the nature of
this advantage requires more research.
Additional research might determine whether this advantage is a re-
sult of cognitive benefits associated with early language learning or of
cultural or social capital. Given that other studies have shown higher
levels of achievement associated with bilingualism (including, but not
specifying biliteracy), some scholars question the causal effect of lan-
guage proficiency on educational outcomes. For example, some wonder
LANGUAGE AND EDUCATION OF BILITERATE LATINOS 117
whether early bilingualism/biliteracy produces enhanced cognitive ca-
pabilities, or whether those with higher cognitive capabilities are more
likely to have strong dual language abilities (see for example MacNamara,
1966). It may be the case that biliteracy is maintained only among the
most diligent students, or students with high aptitudes for language.
Such students would also be likely to do well in academic pursuits.
Similarly, in other work, I have found that the maintenance of biliteracy
requires significant investments of time and effort by both children and
their parents, and therefore, students may benefit from unusually high
levels of parental involvement (Lutz, 2003).
The role of coursework in establishing dual language proficiency is an
element that might be included in future biliteracy research (see
Alexander and Cook, 1982). Perhaps students with early biliteracy gain
an advantage in their pursuit of college entry in that they may be more
likely than other students to take foreign-language coursework, a pre-
requisite for admission to many colleges and universities.
For research to progress in this field, however, additional data collec-
tion is warranted. The greatest obstacle to research on the impact of
language on educational outcomes is the paucity of survey data, par-
ticularly nationally representative survey data that can address these
issues. It is critical that new sources of survey data include measures of
proficiency in both English and Spanish (and/or other languages).
References
Alba, Richard, “The Twilight of Ethnicity among Americans of Euro-
pean Ancestry: The Case of Italians”, Ethnic and Racial Studies, 8,
1985, pp. 134-158.
Alexander, Karl L., and Martha A. Cook, “Curricula and Coursework: A
Surprise Finding to a Familiar Story”, American Sociological Review,
47, 1982, pp. 626-640.
Bain, B., “Bilingualism and Cognition: Toward a General Theory”, in
Stephen T. Carey (ed.), Bilingualism, Biculturalism, and Education,
Edmonton, University of Alberta, 1974, pp. 119-128.
Bankston, Carl L. III, and Min Zhou, “Effects of Minority Language
Literacy on the Academic Achievement of Vietnamese Youths in New
Orleans”, Sociology of Education, 68, 1995, pp. 1-17.
Baral, D., “Academic Achievement of Recent Immigrants from Mexico”,
Journal for the National Association for Bilingual Education, 3, 1979,
pp. 1-13.
Bean, Frank D., and Marta Tienda, The Hispanic Population in the United
States, New York, Russell Sage Foundation, 1987.
118 MIGRACIONES INTERNACIONALES
Bialystok, Ellen, “Levels of Bilingualism and Levels of Linguistic Aware-
ness”, Developmental Psychology, 24, 1988, pp. 560-567.
Bourdieu, Pierre, “Cultural Reproduction and Social Reproduction”,
in Richard Arum and Irenee R. Beattie (eds.),The Structure of School-
ing: Readings in the Sociology of Education, Mountain View (CA),
Mayfield, 2000, pp. 56-68.
Byun, Myung-Sup, “Bilingualism and Bilingual Education: The Case
of the Korean Immigrants in the United States”, International Journal
of the Sociology of Language, 82, 1990, pp. 109-128.
Clifton, Rodney A., Raymond P. Perry, Karen Parsonson, and Stella
Hryniuk, “Effect of Ethnicity and Sex on Teachers’ Expectations of
Junior High School Students”, Sociology of Education, 59, 1986, pp.
58-67.
Cook, Virginia, “Evidence for Multicompetence”, Language Learning,
42, 1992, pp. 557-591.
Cummins, J., “The Influence of Bilingualism on Cognitive Growth: A
Synthesis of Research Findings and Explanatory Hypotheses”, Work-
ing Papers on Bilingualism, 19, 1976, pp. 121-129.
Donato, Rubén, Martha Menchaca, and Richard R. Valencia, “Segre-
gation, Desegregation, and Integration of Chicano Students: Prob-
lems and Prospects”, in Richard R. Valencia (ed.), Chicano School
Failure and Success: Research and Policy Agendas for the 1990s, London:
Falmer Press, 1991, pp. 27-63.
Duncan, Sharon E., and Edward A. De Avila, “Bilingualism and Cog-
nition: Some Recent Findings”, NABE Journal (National Association
for Bilingual Education), 4, 1979, pp. 15-50.
Ekstrom, Ruth B., Margaret E. Goertze, and Judith M. Pollack, “Who
Will Drop Out of High School and Why?”, Teacher’s College Record,
87, 1986, pp. 356-373.
Evans, Francis B., and James G. Anderson, “The Psychocultural Ori-
gins of Achievement and Achievement Motivation: The Mexican-
American Family”, Sociology of Education, 46, 1973, pp. 396-416.
Fernández, Roberto, and François Nielson, “Bilingualism and Hispanic
Scholastic Achievement: Some Baseline Results”, Social Science Re-
search, 15, 1986, pp. 43-70.
Fernández, Roberto, R. Paulsen, and M. Hirano-Nakanishi, “Drop-
ping Out among Hispanic Youth”, Social Science Research, 18, 1989,
pp. 21-52.
Fernández-Kelly, M. Patricia, “Social and Cultural Capital in the Ur-
ban Ghetto: Implications for the Economic Sociology of Immigra-
tion”, in Alejandro Portes (ed.), The Economic Sociology of Immigra-
tion: Essays on Networks, Ethnicity, and Entrepreneurship, New York,
Russell Sage Foundation, 1995, pp. 213-247.
LANGUAGE AND EDUCATION OF BILITERATE LATINOS 119
Fernández-Kelly, M. Patricia, “From Estrangement to Affinity: Dilem-
mas of Identity among Hispanic Children”, in F. Bonilla, E. Meléndez,
R. Morales, and M. de los Ángeles Torres (eds.), Borderless Borders:
U.S. Latinos, Latin Americans, and the Paradox of Interdependence, Phila-
delphia, Temple University Press, 1998, pp. 83-104.
Fillmore, Lily Wong, “Language and Cultural Issues in the Early Educa-
tion of Language Minority Children”, in S. L. Kagan (ed.), The Care
and Education of America’s Young Children: Obstacles and Opportunities,
vol. 1, Chicago, University of Chicago Press, 1991, pp. 30-49.
Fishman, Joshua A., The Sociology of Language: An Interdisciplinary So-
cial Science Approach to Language in Society, Rowley (MA), Newbury
House Publishers, 1972.
Fligstein, Neil, and R. M. Fernández, “Educational Transitions of Whites
and Mexican-Americans”, in George J. Borjas and Marta Tienda (eds.),
Hispanics in the U.S. Economy, Orlando (FL), Academic Press, 1985,
pp. 161-192.
Fordham, Signithia, and John U. Ogbu, “Black Students’ School Suc-
cess: Coping with the Burden of ‘Acting White’”, Urban Review, 18,
1986, pp. 176-206.
Gans, Herbert J., “Second-Generation Decline: Scenarios for the Eco-
nomic and Ethnic Futures of the Post-1965 American Immigrants”,
Ethnic and Racial Studies, 15, 1992, pp. 173-192.
———, “Symbolic Ethnicity: The Future of Ethnic Groups and Cul-
tures in America”, in Werner Sollors (ed.), Theories of Ethnicity, New
York, New York University Press, 1979/1996, pp. 425-459.
García-Coll, Cynthia, and Katherine Magnuson, “The Psychological Ex-
perience of Immigration: A Developmental Perspective”, in Alan Booth,
Ann C. Crouter, and Nancy Landale (eds.), Immigration and the Family,
Mahwah (NJ), Lawrence Erlbaum Associates, Inc., 1997, pp. 91-132.
Glick, Jennifer E., and Michael J. White, “The Academic Trajectories of
Immigrant Youths: Analysis within and across Cohorts,”, Demogra-
phy, 40, 2003, pp. 759-783.
Goldberg, David Theo, Multiculturalism: A Critical Reader, Cambridge
(MA), Blackwell Publishers, 1994.
Hirschman, Charles, “Studying Immigrant Adaptation from the 1990
Population Census: From Generational Comparisons to the Process of
‘Becoming American’”, in Alejandro Portes (ed.), The New Second Gen-
eration, New York, Russell Sage Foundation, 1996, pp. 54-81.
———, “The Educational Enrollment of Immigrant Youth: A Test of
the Segmented Assimilation Hypothesis”, Demography, 38, 2001, pp.
317-336.
Hollinger, David A., Postethnic America: Beyond Multiculturalism, New
York, Basic Books, 1995.
120 MIGRACIONES INTERNACIONALES
Kallen, Horace M., “Democracy versus the Melting Pot: A Study of
American Nationality”, in Werner Sollors (ed.), Theories of Ethnicity:
A Classical Reader, New York, New York University Press, 1915/1996,
pp. 67-92.
Kecskés, István, “The State of L1 Knowledge in Foreign Language Learn-
ers”, WORD, 49, 1998, pp. 320-340.
———, and Tünde Papp, Foreign Language and Mother Tongue, Mahwah
(NJ), Lawrence Erlbaum Associates, Inc., 2000.
Kennedy, Eugene, and Hae-Seong Park, “Home Language as a Predic-
tor of Academic Achievement: A Comparative Study of Mexican- and
Asian-American Youth”, The Journal of Research and Development in
Education, 27, 1994, pp. 188-194.
Logan, John R., and Richard D. Alba, “Locational Returns to Human
Capital: Minority Access to Suburban Community Resources”, De-
mography, 30, 1993, pp. 243-268.
Lutz, Amy, “The Role of Family and Community Contexts in Deci-
sions about Home Language in Latino Families”, presented at the
annual meeting of the American Sociological Association, Atlanta,
Georgia, 2003.
MacNamara, John, Bilingualism and Primary Education: A Study of Irish
Experience, Edinburgh, University of Edinburgh Press, 1966.
Matute-Bianchi, María Eugenia, “Ethnic Identities and Patterns of
School Success and Failure among Mexican-Descent and Japanese-
American Students in a California High School: An Ethnographic
Analysis”, American Journal of Education, 95, 1986, pp. 233-255.
Mouw, Ted, and Yu Xie, “Bilingualism and the Academic Achievement
of First- and Second-generation Asian Americans: Accommodation
with or without Assimilation?”, American Sociological Review, 64, 1999,
pp. 232-252.
National Center for Education Statistics (NCES), National Education Lon-
gitudinal Study 1988-2000 data files and electronic codebook system,
Washington, D.C., National Center for Education Statistics, U.S.
Dept. of Education, Office of Educational Research and Improve-
ment, 2002.
Nielson, François, and Steven J. Lerner, “Language Skills and School
Achievement of Bilingual Hispanics”, Social Science Research, 15, 1986,
pp. 209-240.
Oboler, Suzanne, “The Politics of Labeling: Latino/a Cultural Identities
of Self and Others”, Latin American Perspectives, 19, 1992, pp. 18-36.
Oropesa, R. S., and Nancy Landale, “In Search of the New Second
Generation: Alternative Strategies for Identifying Second Generation
Children and Understanding Their Acquisition of English”, Sociologi-
cal Perspectives, 40, 1997, pp. 429-455.
LANGUAGE AND EDUCATION OF BILITERATE LATINOS 121
Peal, Elizabeth, and Wallace E. Lambert, “The Relation of Bilingual-
ism to Intelligence”, Psychological Monographs, 76, 1962, pp. 1-23.
Portes, Alejandro, “Children of Immigrants: Segmented Assimilation
and Its Determinants”, in Alejandro Portes (ed.), The Economic Soci-
ology of Immigration: Essays on Networks, Ethnicity, and Entrepreneur-
ship, New York (NY), Russell Sage Foundation, 1995, pp. 248-280.
———, and Lingxin Hao, “The Price of Uniformity: Language, Family
and Personality Adjustment in the Immigrant Second Generation”,
Ethnic & Racial Studies, 25, 2002, pp. 889-912.
Portes, Alejandro, and Robert D. Manning, “The Immigrant Enclave:
Theory and Empirical Examples”, in J. Nagel and S. Olzak (eds.), Com-
petitive Ethnic Relations, Orlando, Academic Press, 1986, pp. 47-68.
Portes, Alejandro, and Rubén G. Rumbaut, Immigrant America: A Por-
trait, Berkeley, University of California Press, 1990.
———, Immigrant America: A Portrait, Berkeley, University of Califor-
nia Press, 1996.
———, Legacies: The Story of the Immigrant Second Generation, Berke-
ley, University of California Press/Russell Sage Foundation, 2001.
Portes, Alejandro, and Cynthia Truelove, “Making Sense of Diversity:
Recent Research on Hispanic Minorities in the United States”, An-
nual Review of Sociology, 13, 1987, pp. 359-385.
Portes, Alejandro, and Min Zhou, “The New Second Generation: Seg-
mented Assimilation and Its Variants”, The Annals, 503, 1993, pp.
74-96.
Ready, Timothy, Latino Immigrant Youth: Passages from Adolescence to
Adulthood, New York, Garland, 1991.
Roscigno, Vincent, María B. Vélez, and James W. Ainsworth-Darnell,
“Language Minority Achievement, Family Inequality, and the Im-
pact of Bilingual education”, Race and Society, 4, 2001, pp. 69-88.
Rumbaut, Rubén G., “The Crucible Within: Ethnic Identity, Self-Es-
teem, and Segmented Assimilation among Children of Immigrants”,
in Alejandro Portes (ed.),The New Second Generation, New York,
Russell Sage Foundation, 1996, pp. 119-170.
———, “Assimilation and Its Discontents: Between Rhetoric and Real-
ity”, International Migration Review, 31, 1997, pp. 923-960.
Rumberger, Russell, and Katherine Larson, “Toward Explaining Differ-
ences in Educational Achievement among Mexican American Language-
Minority Students”, Sociology of Education, 71, 1998, pp. 69-93.
Stevens, Gillian, “Sex Differences in Language Shift in the United States”,
Social Science Review, 71, 1986, pp. 31-36.
Valenzuela, A., and S. M. Dornbusch, “Familism and Social Capital in
the Academic Achievement of Mexican Origin and Anglo Adolescents”,
Social Science Quarterly, 75, 1994, pp. 18-36.
122 MIGRACIONES INTERNACIONALES
Valverde, S. A., “A Comparative Study of Hispanic High School Drop-
outs and Graduates: Why Do Some Leave School Early and Some
Finish?”, Education and Urban Society, 19, 1987, pp. 320-329.
Vélez, William, “High School Attrition among Hispanic and Non-His-
panic White Youths”, Sociology of Education, 62, 1989, pp. 119-133.
Vygotsky, L. S., Thought and Language, Cambridge (MA), MIT Press,
1962.
Warren, J. R., “Educational Inequality among White and Mexican-ori-
gin Adolescents in the American Southwest: 1990”, Sociology of Edu-
cation, 69, 1996, pp. 142-158.
White, Michael J., and Jennifer E. Glick, “Generation Status, Social
Capital, and the Routes Out of High School”, Sociological Forum, 15,
2000, pp. 671-692.
Willig, A. C., “A Meta-Analysis of Selected Studies on the Effectiveness
of Bilingual Education”, Review of Educational Research, 55, 1985,
pp. 269-317.
Zhou, Min, Chinatown: The Socioeconomic Potential of an Urban En-
clave, Philadelphia, Temple University Press, 1992.
———, “Segmented Assimilation: Issues, Controversies, and Recent
Research on the New Second Generation”, International Migration
Review, 31, 1997, pp. 975-1008.
———, and Carl L. Bankston, “Social Capital and the Adaptation of
the Second Generation: The Case of Vietnamese in New Orleans”, in
Alejandro Portes, The New Second Generation, New York City, Russell
Sage Foundation, 1996, pp. 197-220.
———, Growing Up American: How Vietnamese Children Adapt to Life
in the United States, New York, Russell Sage Foundation, 1998.
Fecha de recepción: 6 de julio de 2004
Fecha de aceptación: 3 de septiembre de 2004