ArticlePDF Available

English Language Learners' Access to and Attainment in Postsecondary Education


Abstract and Figures

Although English language learners (ELLs) are currently the fastest-growing group among the school-age population in the United States, there is surprisingly little information on their participation in postsecondary education. Using the National Education Longitudinal Study of 1988 (NELS:88), a nationally representative sample of eighth graders who were followed for 12 years, we present one of the first national-level examinations of ELLs' access to and degree of attainment in postsecondary education. Our analyses show that ELLs lag far behind both English-proficient linguistic minority students and monolingual English-speaking students in college access and attainment. Only one in eight ELLs in the NELS:88 study earned a bachelor's degree, whereas one in four English-proficient linguistic minority students and one in three monolingual English speakers did. In addition, one in five ELLs was a high school dropout. Subsequent probit regressions reveal that a host of nonlinguistic factors, rather than the ELLs' linguistic background per se, contributed to ELLs' limited postsecondary education access and attainment.
Content may be subject to copyright.
English Language Learners’ Access to
and Attainment in Postsecondary
Temple University
Philadelphia, Pennsylvania, United States
Although English language learners (ELLs) are currently the fastest-
growing group among the school-age population in the United
States, there is surprisingly little information on their participation in
postsecondary education. Using the National Education Longitudinal
Study of 1988 (NELS:88), a nationally representative sample of
eighth graders who were followed for 12 years, we present one of the
first national-level examinations of ELLs’ access to and degree of
attainment in postsecondary education. Our analyses show that ELLs
lag far behind both English-proficient linguistic minority students
and monolingual English-speaking students in college access and
attainment. Only one in eight ELLs in the NELS:88 study earned a
bachelor’s degree, whereas one in four English-proficient linguistic
minority students and one in three monolingual English speakers
did. In addition, one in five ELLs was a high school dropout. Subse-
quent probit regressions reveal that a host of nonlinguistic factors,
rather than the ELLs’ linguistic background per se, contributed to
ELLs’ limited postsecondary education access and attainment.
doi: 10.1002/tesq.49
English language learners (ELLs), students who are in the process
of learning English and who need linguistic support in order to
learn grade-level academic content, are currently the fastest-growing
group among the school-age population in the United States (Wolf,
Herman, Bachman, Bailey, & Griffin, 2008). There are now roughly
5.3 million ELLs in K12 public schools in the United States, repre-
senting approximately 10.8% of all students (National Clearinghouse
for English Language Acquisition, n.d.). The U.S. Department of
Education estimates that this figure will increase to 25% of students
by 2025 (Spellings, 2005). If ELLs are rapidly increasing in number in
K12 schools, we can expect them to be a growing presence in postsec-
ondary education (PSE) as well. Yet little is currently known about
ELLs’ college-going patterns.
TESOL QUARTERLY Vol. 47, No. 1, March 2013
©2012 TESOL International Association
There is a large body of research on traditionally underrepresented
students’ access to and attainment in college. However, in this set of
literature, sociologists tend to focus on categories other than linguistic
background, such as race or ethnicity (Deil-Amen & Turley, 2007; Kao
& Thompson, 2003), Latinos (Arbona & Nora, 2007; Callahan, 2008;
Swail, Cabrera, Lee, & Williams, 2005), socioeconomic status (Bowen,
Kurzwell, & Tobin, 2005; McDonough, 1997; Walpole, 2007), first-gen-
eration college students (Nun
˜ez & Cuccaro-Alamin, 1998; Pascarella,
Pierson, Wolniak, & Terenzini, 2004), and undocumented immigrants
(Morales, Herrerra, & Murry, 2009). Although ELLs may be subsumed
under one or more of these categories, they are rarely studied in their
own right. On the other hand, applied linguists and composition
scholars have long studied college-level ELLs’ academic literacy and
their experiences in English as a second language (ESL) and composi-
tion classes (e.g., Harklau, Losey, & Siegal, 1999; Leki, 2007; Matsuda,
1999; Matsuda, Ortmeier-Hooper, & You, 2006; Roberge, Siegal, &
Harklau, 2009; Zamel, 1995). But these researchers focus primarily on
ELLs’ linguistic challenges, leaving unexplored the broader issues of
ELLs’ college access and success. For example, Leki’s (2007) longitudi-
nal study of the literary learning experiences of four undergraduate
ESL students casts a much wider net over their undergraduate experi-
ences as a whole than the typical second language (L2) writing study.
She argues that sociocultural relationships (i.e., the social capital that
ESL students can develop on campus) are critical to their academic
success. Even so, Leki is deeply steeped in the disciplinary tradition of
L2 writing research and has little to say about ESL students’ overall
college access and degree attainment.
A small but growing body of studies focusing specifically on ELLs’
access to and attainment in college suggests the immense challenges
that these students encounter if they want to advance to PSE.
Using the
Current Population Survey, Klein, Bugarin, Beltranena, and McArthur
(2004) found that only 13.5% of 18- to 24-year-old adults with limited
English proficiency were enrolled in PSE in 1999 compared with 37.2%
In the United States, there are essentially three types of PSE institutions from which stu-
dents can choose. Four-year colleges and universities issue bachelor’s degrees and are
considered the most prestigious type of PSE institution among the three. Admission into
a 4-year institution usually involves a selection process; some universities are more selec-
tive than others. Students can also opt for 2-year collegespublic community colleges
and private junior colleges. Community colleges, which are the most common type of 2-
year colleges, have open admissions policies (i.e., a high school diploma is not a require-
ment for admission), whereas junior colleges usually have eligibility criteria. Some stu-
dents attend community colleges to pursue an associate’s degree and/or to later transfer
into a 4-year university; others enroll in vocational programs within a community college.
The third type of PSE institution are private vocational/technical schools that offer train-
ing programs for those who want to work in specific occupational areas, such as nurse’s
aide, car mechanic, plumbing, and barbering.
of monolingual English speakers. The usual explanation for ELLs’ low
PSE participation and other academic underachievement issues has
been their limited English proficiency (Callahan, 2005; Ga´ndara &
Rumberger, 2009). However, recent research suggests that other factors
also inhibit their college access (Callahan, 2005; Callahan, Wilkinson, &
Muller, 2010; Harklau, 2000). ELLs’ placement in ESL classes in high
school steers them away from college preparatory courses, making them
less eligible for admission to 4-year institutions (Callahan et al., 2010).
Also, ELLs tend to come from lower socioeconomic status backgrounds
than do non-ELLs (Education Week, 2009); the lack of economic
resources, and therefore the need to work long hours, limits both ELLs’
college choices and engagement in collegiate activities (Kanno & Vargh-
ese, 2010). Those who enroll in community colleges are faced with the
challenge of navigating various placement tests and a vast array of
courses, often with less than helpful guidance from the college (Bunch
& Endris, 2012). Many ELLs who start their PSE career in a community
college, hoping to transfer to a 4-year college later, in reality do not
reach regular college-level courses outside of the ESL program (Razfar
& Simon, 2011). Finally, institutional labeling of ELLs tends to frame
them in deficit terms, which leads to lower teacher expectations and
academic marginalization (Harklau, 2000).
Research in other English-dominant countries shows strikingly similar
predicaments for ELLs in those countries. Scholars in the United King-
dom (Martin, 2010; Preece & Martin, 2010; Simpson & Cooke, 2010)
point out that although U.K. higher education institutions have opened
doors to more “nontraditional” studentsincluding ELLsand even
though many of these institutions highlight the importance of cultural
and linguistic diversity, in reality there is simply no recognition for the
multilingual assets that these new students bring. The situation is, there-
fore, markedly similar to the politics of linguistic diversity in U.S. higher
education (Horner & Trimbur, 2002). Similarly, Marshall (2010) reports
that in Canada many multilingual immigrant students experience the
process of “re-becoming ESL” when they enter university: They are
required to take remedial ESL courses in university and are thereby
given a “deficit remedial ESL identity” (p. 45), regardless of how many
years previously they had exited ESL programs. Again, this is a striking
parallel to the experience of many ELLs in the United States (Kanno &
Varghese, 2010). These reports show that challenges for ELLs in enter-
ing higher education and achieving success are not unique to the
United States, but are shared by other English-dominant countries,
adding to the urgency of addressing this issue.
Overall, there is still very little research available on ELLs’ college-
going patterns. In particular, apart from Klein et al.’s (2004) research,
there are no national-level statistics on ELLs’ college access and
attainment in the United States. Klein et al.’s study, moreover, pro-
vides only descriptive statistics and does not investigate the factors that
contribute to level of access and attainment. In the present study,
therefore, we use the National Education Longitudinal Study of 1988
(NELS:88; National Center for Education Statistics, n.d.), first examin-
ing ELLs’ PSE access and degree attainment at the national level and
then investigating what linguistic and nonlinguistic variables predict
PSE access and attainment levels. More specifically, this study
addresses the following research questions:
1. Are ELLs’ patterns of access to and attainment in college differ-
ent from those of monolingual English-speaking students and
those of English-proficient linguistic minority students?
2. To what extent does linguistic background predict level of
access and level of attainment compared with other, nonlinguis-
tic variables, such as family income, race or ethnicity, and high
school course-taking?
In this study, we draw on Bourdieu’s (Bourdieu, 1977, 1986; Bourdieu
& Passeron, 1990) theory of cultural reproduction to guide our inquiry.
Bourdieu argues that schools contribute to the reproduction of the
existing hierarchy in society by creating conditions in the school system
in which children of the dominant class are more likely to succeed.
Schools achieve this by privileging the knowledge, dispositions, and hab-
itscultural capitalthat children of the dominant class bring to school
and treating their cultural capital as if it were inherently more valuable
than the cultural capital of poor, working-class, and minority children.
Because children of the dominant class find a close alignment between
the values of home and school cultures, achievement is likely to come
relatively easily to them. In contrast, nondominant-class children must
bridge a much larger gulf between school and home cultures if they
are to achieve school success. Ultimately, as sons and daughters of
the dominant class achieve school success, they accumulate academic
credentialswhat Swail (2000) calls “educational capital” (p. 86)
which will translate into other forms of capital such as financial wealth
(economic capital), membership in powerful and resourceful social net-
works (social capital), and professional status (symbolic capital; Bourdieu,
1986; Swartz, 1997).
Higher education, especially 4-year college education, is the pinna-
cle of this system of unequal power distribution. It can be seen as the
final stage of cultural reproduction through education. As such, young
adults who belong to the dominant class are much more likely to have
easy access to 4-year colleges and universities compared to their peers
who come from low-income and minority backgrounds. This is not
necessarily because they are academically more meritoriousthough
the education system in which they have participated so far is likely to
have ensured that, toobut because they possess more of the
resources and knowledge, such as college-educated parents, familiarity
with the college application process, financial means, and access to
academically rigorous high schools, that are likely to have given them
an edge in the college application competition. As Swail (2000) writes,
“Access to higher education is more an issue of social and cultural cap-
ital than anything else” (p. 86).
If we view the issue of college access and success this way, we can
see that ELLs enter into this competition with inherent and multiple
disadvantages. First of all, they lack vital linguistic capital: They have
not acquired full proficiency in English, which is indisputably the most
privileged language in the U.S. college education system and often
enjoys its monopoly as the language of instruction and administration
in higher education. In addition, ELLs lack other forms of capital,
such as financial resources, the assistance of college-educated parents,
and access to a rigorous high school curriculum, that are considered
critical to college access and degree attainment.
It is important not to lose sight of the fact that ELLs possess other
forms of capital that members of the English-dominant group do not
possess, such as multilingualism, cultural sensitivity, and ability to adapt
to new environments. In certain international arenas, these experiences
and abilities would indeed be considered highly coveted forms of capi-
tal. However, in the arena of U.S. college access and success, these are
not the kinds of capital that give students an edge; ELLs tend to lack the
very forms of capital that advantage students in college access competi-
tion. The goal of this study, then, was to explore what forms of capital
contribute to different levels of college access and attainment, and
which of these types of capital ELLs do or do not possess.
NELS:88, sponsored by the National Center for Education Statistics,
started in 1988 with a nationally representative stratified random
sample of 24,599 eighth graders from 1,052 randomly selected schools,
and it followed them for the next 12 years. Students were followed
from the time they were 1314 years old in 1988 to when they were
2526 years old in 2000. After the base-year NELS survey in 1988, stu-
dents were resurveyed at 2-year intervals until 1994 and then one final
time in 2000 (i.e., four follow-ups: 1990, 1992, 1994, and 2000). Stu-
dents were followed, to the extent possible, regardless of whether they
were still enrolled in formal schooling. Information was also collected
from parents, teachers, and school administrators. Students’ achieve-
ment test scores and high school and PSE transcripts were also col-
lected. NELS intentionally drew a larger sample of Asian and Hispanic
students in proportion to their representation in the population (i.e.,
oversampled) in order to include sufficient numbers of these students
in the sample to permit statistical analyses.
Students with severely limited English proficiency (i.e., not able to
read the surveys in English) were excluded from the NELS sample.
Also, as with any large-scale longitudinal study, a substantial portion of
original participants in NELS dropped out of the study over time. Of
the 24,599 students in the initial sample, 12,140
participated in all
five waves of data collection. Of these 12,140 students, we excluded
the very small sample of Native American students due to problems
with model convergence; we also excluded students whose linguistic
background was missing as well as students whose PSE attainment data
were missing. This resulted in an analytic sample of 10,300 students.
Participant attrition is accounted for in our analyses by two methods:
(1) the NELS-calculated weights, which adjust for dropout, and (b)
our use of Full Information Maximum Likelihood (FIML) estimation.
Although the NELS data are somewhat dated by now, they nonetheless
constitute the newest data set available that allows us to track students’
educational trajectories from the beginning of high school to the end
of college.
Following the policy of the National Center for Education Statistics, which collected the
NELS data, we are rounding the numbers to the nearest 10 when reporting unweighted
subsample sizes in order to avoid the identification of individual students.
A more recent data set, the Education Longitudinal Study of 2002 (ELS:2002), tracks a
national sample of 10th graders through high school graduation and PSE participation
( Although the newness of ELS:2002 offers an
advantage over NELS:88, we decided to analyze NELS for the following reasons: ELS is
still collecting data, and currently the most recent follow-up data available are from
2006, which is 2 years after participants’ scheduled high school graduation. In other
words, ELS does not yet offer information on participants’ PSE attainment. Because we
wanted to examine not only ELLs’ PSE access but also attainment, it was crucial that we
use a data set that contains both students’ PSE access and their attainment. Once ELS
2012 follow-up data are made available, we will be able to conduct similar analyses to
compare our NELS results to those from more recently collected data.
Linguistic Background
We divided students into three categories according to their linguis-
tic backgrounds: English-monolingual students (EMs), English-proficient lin-
guistic-minority students (EPs), and English language learners (ELLs).
used a variation of Strang, Winglee, and Stunkard’s (1993) method of
identifying different linguistic background groups.
First, we categorized students into two groups: EMs and linguistic
minority students (LMs). EMs are students who self-identified on the
base-year survey that they came from monolingual English-speaking
homes and that their first language (L1) was English. LMs, on the
other hand, are students who self-identified (a) that their L1 was not
English or (b) that they primarily spoke a language other than English
at home.
We then further divided the LMs into two groups: ELLs and EPs.
ELLs are LMs who exhibited signs of limited English proficiency,
whereas EPs are LMs who showed no signs of limited English profi-
ciency. We used multiple sources of data to identify ELLs and EPs.
First, of the LMs identified as such in the first step, those who self-
rated at least one of their four skills (listening, speaking, reading, and
writing) in English in the bottom two categories of the 4-point scale in
the base-year survey were identified as ELLs.
Also, NELS surveyed two
teachers for each student in the base-year survey. If at least one of the
teachers identified a student as an ELL, we categorized the student as
an ELL.
Additionally, we used high school transcript data to further identify
ELLs, because sole reliance on the student base-year surveys for ELL
identification would result in the elimination of those ELLs who
did not answer key linguistic background questions on the survey.
Following Callahan et al. (2010), we identified any students as ELLs
whose high school transcript indicated that they took at least one ESL,
NELS:88 includes a composite variable, BYLEP, for students we call ELLs here. However,
researchers in the past expressed dissatisfaction with this variable, arguing that using it
results in false positive identification of ELLs (Guglielmi, 2008; Strang et al., 1993).
Because the accurate identification of ELLs in NELS was critical to our analysis, we
decided to identify ELLs ourselves, using a variety of variables in the data rather than
relying on a preexisting composite variable.
The four choices in the base-year survey for self-rating of the four skills were very well,
pretty well,well, and not very well. We decided to include the bottom two categories (well
and not very well) as markers of limited English proficiency, because (a) there was a clear
gap in the distribution between the top two and bottom two categories and (b) the
bottom two categories represented very small proportions of the LMs (0.4%0.7%,
unweighted). The National Center for Education Statistics, which collected the NELS
data, used anything other than the very well category as markers of limited English profi-
ciency (e.g., Aud et al., 2011). Thus, our categorization is more conservative than this
bilingual education, or sheltered English content class during high
Finally, we excluded ELLs from the entire LM group and identified
all remaining students as EPs. In all, we identified 8,450 EMs (82%),
1,500 EPs (15%), and 350 ELLs (3%), for a total of 10,300 students
(see Table 1).
Access and Attainment
Our dependent variables are access to PSE and attainment in PSE.
Access to PSE refers to the student’s first PSE institution, as reported
in the 1994 wave
(i.e., the fourth wave of data collection and 2 years
after scheduled high school graduation) and has three categories: (1)
bachelor’s institution or above, (2) vocational program or community
college, and (3) no PSE. Attainment in PSE refers to the student’s
highest degree earned by 2000 (the fifth and final wave of data collec-
tion and 8 years after scheduled high school graduation) and is
divided into five categories: (1) bachelor’s degree or higher, (2) asso-
ciate’s degree or certificate, (3) some PSE, (4) high school diploma,
and (5) high school dropout.
Data analyses were conducted in two stages. First, we examined the
distribution of EMs, EPs, and ELLs across different levels of access and
We conducted exploratory analyses to verify whether there were any differences in
important characteristics between the survey-based ELLs and the transcript-based ELLs.
We found no statistically significant differences for any of the key variables (high school
GPA, math or reading test scores, PSE access, or attainment) between the two groups,
except for one: For the highest level math course taken in high school, the mean for
survey-based ELLs was 2.0 (i.e., at the low end of middle math), whereas the mean for
transcript-based ELLs was 1.8 (i.e., at the high end of low math). Although statistically
significant, the difference is trivially small. We thus decided to combine the two groups
into one ELL group.
We are aware that minority students sometimes delay entry into college beyond a few years
after high school graduation (Goldrick-Rab, 2007). However, after the 1994 follow-up,
NELS did not collect information until 2000. Thus, the 1994 wave, 2 years after scheduled
high school graduation, is the only possible data point in NELS at which to analyze stu-
dents’ entry into PSE. Any PSE participation of those students who entered PSE after 1994
would have been accounted for with the 2000 data. However, a cross-tabulation of 1994
PSE access and 2000 PSE attainment data reveals that none of the students in this sample
who had not entered PSE by 1994 subsequently participated in PSE.
For both PSE access and attainment, vocational program and community college were
combined together due to the small number of students in each cell, especially for the
ELL population.
Descriptive Statistics by Linguistic Background (Unweighted)
Linguistic background
ELL (n~350) EP (n~1,500) EM (n~8,450)
value %
value %
value %
Entire analytic sample
PSE access 9% 9% 7%
25% 43% 46%
Vocational or
38% 38% 32%
No PSE 37% 19% 22%
PSE attainment 0% 0% 0%
degree or
21% 32% 35%
Certificate or
9% 11% 9%
Some PSE 30% 34% 30%
High school
diploma no
28% 17% 21%
High school
13% 6% 4%
55% 46% 8%
62% 44% 1%
6% 17% 2%
31% 39% 97%
Sex 24% 0% 0%
Female 50% 53% 53%
Male 50% 47% 47%
Race/ethnicity 25% 0% 1%
Asian 27% 27% 2%
Hispanic 57% 51% 3%
Black 3% 3% 11%
White 12% 20% 84%
Family income 7.91 2.91 26% 9.48 2.85 24% 10.36 2.46 14%
2.19 1.27 17% 2.86 1.47 13% 3.13 1.20 7%
7.97 3.12 33% 9.05 2.81 10% 8.65 2.72 5%
25% 2% 1%
77% 85% 81%
attainment. Once we determined that the three linguistic background
groups were unevenly distributed across different levels of access and
attainment, we then conducted probit regressions using Mplus software
Linguistic background
ELL (n~350) EP (n~1,500) EM (n~8,450)
value %
value %
value %
23% 15% 19%
Highest level
12% 18% 14%
10% 7% 8%
Low math 11% 7% 6%
Middle math 51% 46% 47%
Advanced math 27% 40% 39%
High school
2.61 0.69 36% 2.71 0.78 39% 2.72 0.74 33%
Math test score 47.53 10.35 30% 53.89 11.46 24% 54.92 10.44 22%
Reading test
44.58 13.00 30% 52.34 11.24 24% 53.76 10.63 22%
ELL (n~210) EP (n~1,150) EM (n~6,280)
Participants who attended any PSE
Part-time status 0% 0% 0%
Part-time in
42% 42% 35%
Full-time in
58% 58% 65%
Delay entering
7% 6% 5%
Some delay
entering PSE
30% 23% 19%
No delay
entering PSE
70% 77% 81%
Total credits in
first year PSE
21.25 10.55 14% 22.51 11.43 10% 23.75 10.66 7%
2.48 0.84 5% 2.63 0.82 3% 2.67 0.85 2%
Note. ELL =English language learner; EP =English-proficient linguistic minority student; EM
=English-monolingual student; PSE =postsecondary education. Family income units corre-
spond most closely to increments of $10,000 per year, where 8 =$15K to <$20K, 10 =
$25K to <$35K, and 11 =$35K to <$50K. Parental education is measured on a 6-point
scale: 2 =HS graduate, 3 =some college, and 4 =college graduate (associate or bache-
lor’s). Parental educational expectations are measured on a 12-point scale: 7 =<2 years of
college, 8 =2 or more years of college (but no degree), and 9 =completing a 2-year
degree. Math and reading test scores are IRT-scaled scores; the unweighted mean for all
NELS participants for reading is 53.03 with a standard deviation of 10.70; the unweighted
mean for math is 54.30 with a standard deviation of 10.48.
(Muthe´n & Muthe´n, 19982007) to analyze the effects of linguistic and
nonlinguistic variables that predict levels of access and attainment.
In the type of linear regression familiar to many readers, we analyze
an outcome, such as a test score. This dependent variable shows a
large range of scores and also has a mean and a standard deviation.
The results of such a linear regression show how large the contribu-
tion of each predictor is to the outcome, and the numbers expressing
these contributions are called coefficients. Results also show how well
we explained the outcome (called R
). Access and attainment, by con-
trast, are categories, not scores. Nonetheless, we can rank levels of
access and attainment (e.g., an associate’s degree can be ranked
higher than a high school diploma, but lower than a bachelor’s
degree). We therefore analyzed our data using probit regression,
which is suitable for these kinds of data and which gives us similar
coefficients and a pseudo R
For all continuous predictors (e.g., family income), we subtracted
the sample mean score from each participant’s own score to create
acentered variable in order to avoid violating a regression assump-
tion related to highly correlated predictors (called multicollinearity).
This means that results of the regression represent a 1-point or one
standard deviation increase above the mean for each centered
For all predictors, we report the unstandardized coefficients (b, indi-
cating the effect of a 1-point increase in the predictor on level of
access or attainment) and standardized coefficients (b, indicating the
effect of a one standard deviation increase in the predictor on level of
access or attainment). Standardized coefficients allow for direct com-
parison of the relative magnitudes of the impact of predictor variables.
For the probit regressions for PSE access, data for all students from
our analytical sample were included. However, for the probit regressions
for PSE attainment, only the data for students who had ever attended
PSE were entered. The probit regressions for access identified the
factors that contributed to students’ admission to PSE institutions. For
the attainment data, therefore, we wanted to tease out the factors that
influenced students’ persistence after they had entered PSE.
Because certain subgroups are oversampled in large-scale surveys
such as NELS, and because attrition from the study causes the propor-
tions of various groups to shift over time, we used a weighting variable
called F4PNLWT that accompanies the NELS:88 data set. In order to
account for missing data, we used FIML analyses in Mplus together
with a robust estimator (MLR), which is a more conservative approach
when there are missing data. All analyses were evaluated at an alpha
level of p<.05. Exact pvalues are reported in the text, except that
very small obtained pvalues are reported as p<.001.
Predictor Variables
Our theoretical assumption that various forms of capital contribute
to a student’s chances for higher education led us to consider differ-
ent facets of students’ lives that could affect their access to and attain-
ment of PSE. For the college access regression analyses, we used three
categories of predictors: (1) demographic/individual characteristics,
(2) family capital, and (3) high school factors. For the attainment
analyses, we added another layer: (4) PSE factors. For all analyses, we
entered variables in blocks in the following order: (1) linguistic back-
ground, (2) all other demographic variables, (3) family capital, (4)
high school factors, and, for attainment only, (5) PSE factors.
Demographic characteristics. We entered five demographic/individ-
ual characteristics variables: linguistic background (EM, EP, and ELL),
immigrant generation (first, second, or third), gender (M =0, F =1),
race/ethnicity, and family income. The immigrant generation variable
was created from information concerning the students’ birthplace and
the parents’ birthplace and has three categories: first generation (the
student is non-U.S.-born), second generation (the student is U.S.-born
and at least one of the parents is non-U.S-born), and third generation
(both the student and parents are U.S.-born; see the appendix). Some
previous research shows increased educational attainment with each
generation, although other studies suggest an advantage for the sec-
ond generation, which may benefit both from familiarity with the lan-
guage and culture of the adopted society and from immigrant
optimism (Hirschman, 2001; Kao & Tienda, 1995). Four categories of
race/ethnicity are included in the analysis: Asian, Hispanic, Black, and
White (reference group). The NELS questionnaire asked about the
family’s total gross annual income using 15 income ranges, from 1 (no
income) to 15 ($200,000 or more; see the appendix). Normality statis-
tics suggested this variable could be treated as a continuous variable in
all analyses.
Family capital. Family capitalcultural and social capital available
from the familyis measured by three variables: parental education,
educational expectations of parents, and family composition. Parental
education refers to the highest level of education achieved by either
parent, ranging from 1 (no high school diploma) to 6 (doctorate). In
this study, following Nun
˜ez and Kim (2012) and Walpole (2007), we
used parental education as a measure of cultural and social capital
available from the parents and family income as a measure of the stu-
dent’s economic capital. College-educated parents are more likely to
be able to provide their children with concrete advice about college
choices and applications than non-college-educated parents based on
their own experiences of college and information they gain from their
networks of college-educated friends and acquaintances. Educational
expectations of parents indicates how far the parents expect the child to
advance in his or her education, ranging from 1 (less than high school
diploma) to 12 (PhD or professional degree). An extensive body of
literature has documented strong effects of parents’ educational
expectations on children’s educational attainment (e.g., Hossler &
Stage, 1992; Morgan, 1998; Perna & Titus, 2005). As with family
income, we treated parental education and educational expectations
as continuous variables. Finally, family composition indicates whether the
student came from a two-parent (or two-guardian) home or a single-
parent (or single-guardian) home. Two-parent families on average are
likely to be able to provide more family support than single-parent
families by virtue of the fact that two individuals are sharing the
parenting tasks (Hirschman & Lee, 2005).
High school factors. High school factors tap into the forms of aca-
demic capitalacademic knowledge and credentialsthat students
can accrue in high school: highest level of high school math com-
pleted, high school GPA, and math and reading test scores in 12th
grade. Highest level of high school math completed is used as a proxy for a
student’s academic preparation in high school (Adelman, 1999, 2006;
Perna & Titus, 2005). Taking math courses beyond algebra 2, in partic-
ular, is a key predictor of students’ viability in college (Adelman, 2006;
Callahan et al., 2010). We used Burkam and Lee’s (2003) categories of
math courses to analyze the high school transcripts by NELS course
codes. We categorized students into four groups according to the
highest math course they completed in high school: nonacademic (basic
or general math), low academic (pre-algebra and algebra 1), middle
academic (algebra 2 and geometry), and advanced academic (calculus,
trigonometry, and statistics). Math and reading test scores, IRT-scaled
theta scores from the 12th grade, are included in order to examine
the effect of students’ academic abilities independent of their course-
taking and GPAs.
PSE factors. Finally, student behavior and campus experiences
have an important bearing on college attainment (Adelman, 2006).
For the regression analyses of PSE attainment, we included four mea-
sures of PSE factors: part-time status, delay in entering PSE, total aca-
demic credits earned in first year of PSE, and first-year undergraduate
GPA. Delay in entering PSE is a dichotomous variable indicating whether
there was any gap period between a student’s high school graduation
and his or her entry into the first PSE institution (1 =delay). Part-time
status and total academic credits earned in first year of PSE indicate a stu-
dents’ degree of engagement in college, whereas first-year undergraduate
GPA suggests a student’s overall academic performance in college.
Differential Levels of Access and Attainment by Linguistic
To start with our first research question regarding ELLs’ patterns of
PSE access and attainment as compared to those of EPs and EMs, we
used chi square tests and found significantly different patterns among
the three linguistic background groups for both PSE access (v
=2,424,210] =27630.1, p<.001, Φ=.107) and attainment
[8, N
=2,646,082] =43210.4, p<.001, Φ=.128). First,
with regard to PSE access, the biggest contribution to chi square was
EMs’ underrepresentation in vocational programs and community col-
leges, followed by EMs’ overrepresentation in 4-year institutions and
EPs’ overrepresentation in vocational programs and community col-
leges (see Table 2). On the other hand, ELLs were overrepresented in
the no PSE category and underrepresented in 4-year institutions. ELLs
advanced to PSE institutions at much lower rates than EPs and, espe-
cially, EMs. Only 17.8% of ELLs attended 4-year institutions, and
46.5% did not advance to any PSE at all within 2 years after their
Access to Postsecondary Education by Language Group (Weighted)
Language status
4-year institution
n9,269 101,182 906,194 1016,645
% 17.8% 38.2% 43.0% 41.9%
Contribution to chi square 12,592.0 9,811.8 22,403.8
Vocational OR community college
n18,630 111,079 686,545 816,254
% 35.7% 42.0% 32.6% 33.7%
Contribution to chi square 1,078.0 21,963.2 23,041.2
n24,229 52,406 514,676 591,311
% 46.5% 19.8% 24.4% 24.4%
Contribution to chi square 11,514.0 12,151.3 637.3
Total 52,128 264,667 2,107,415 2,424,210
Note. ELL =English language learner; EP =English-proficient linguistic minority student;
EM =English-monolingual student; PSE =postsecondary education.
scheduled high school graduation. By contrast, 43.0% of EMs moved
on to 4-year institutions within 2 years after high school graduation,
and only 24.4% did not advance to any PSE institutions. Access pat-
terns for EPs resembled those of EMs more than they resembled those
of ELLs, although a smaller percentage of EPs than EMs went to 4-year
institutions (38.2% vs. 43.0%), and a larger percentage of EPs than
EMs enrolled in vocational programs or community colleges (42.0%
vs. 32.6%).
Attainment patterns are similar to those of access (see Table 3).
The biggest contribution to chi square was EMs’ overrepresentation in
the bachelor’s degree or higher category, followed by EPs’ overrepresenta-
tion in some PSE, EMs’ underrepresentation in some PSE, EPs’ under-
representation in the high school diploma no PSE category, and ELLs’
underrepresentation in the bachelor’s degree or higher category. Within
8 years after scheduled high school graduation, 31.9% of EMs attained
bachelor’s degrees or higher, and 28.8% stopped at or before the high
school level. In contrast, only 11.7% of ELLs obtained a bachelor’s
degree or higher, and 50.8% stopped at the high school level or lower.
Roughly one in three EMs and one in four EPs had earned a
Attainment in Postsecondary Education by Language Group (Weighted)
Language status
Bachelor’s degree or higher
n7,067 72,392 730,796 810,255
% of language status group 11.7% 24.6% 31.9% 30.6%
Contribution to chi square 11,433.9 17,821.2 29,255.1
Certificate OR associate’s degree
n3,875 31,577 206,826 242,278
% of language status group 6.4% 10.7% 9.0% 9.2%
Contribution to chi square 1,657.0 4,601.9 2,944.9
Some PSE
n18,760 117,216 693,585 829,561
% of language status group 31.0% 39.8% 30.3% 31.4%
Contribution to chi square 181.7 24,853.2 24,671.6
High school diploma no PSE
n18,094 53,341 514,877 586,312
% of language status group 29.9% 18.1% 22.5% 22.2%
Contribution to chi square 4,706.5 11,938.6 7,232.1
High school dropout
n12,623 20,087 144,966 177,676
% of language status group 20.9% 6.8% 6.3% 6.7%
Contribution to chi square 8,566.1 304.7 8,870.7
Total 60,419 294,613 2,291,050 2,646,082
Note. ELL =English language learner; EP =English-proficient linguistic minority student;
EM =English-monolingual student; PSE =postsecondary education.
bachelor’s degree or higher within 8 years, whereas only one in eight
ELLs had. Conversely, approximately one in two ELLs did not earn
any college credits at all, but this was the case for fewer than one in
three EMs; furthermore, about one in five ELLs was a high school
dropout. Across linguistic background groups (more than 30% in all
groups), but especially for EPs (39.8%), a large portion of students
started college but did not finish.
Linguistic vs. Nonlinguistic Factors
Our analyses clearly demonstrates ELLs’ lower rates of both enter-
ing and finishing PSE compared with their more English-proficient
peers, especially EMs. We thus went on to ask the second research
question: To what extent does linguistic background predict level of
access and level of attainment compared with nonlinguistic variables?
Access. The final model explained 49.5% of the variance in college
access (see Table 4). Of the linguistic background factors, ELL was a
negative predictor of PSE access in model 1, but its magnitude was
lower in model 2 when other demographic factors were added to the
regression. When family capital factors were added to the regression
in model 3, ELL was no longer a significant predictor. This pattern
suggests that the disadvantages associated with being an ELL in fact
stemmed from (a) a large portion of ELLs being Hispanic and (b)
lack of family capital (e.g., non-college-educated parents, low educa-
tional expectations of parents). Descriptive statistics in Table 1 show
that 57% of ELLs in this sample were Hispanic and that ELLs’ paren-
tal education level and educational expectations were considerably
lower than those of EMs. Interestingly, EP, which was a nonsignificant
predictor in model 1, became a significant and positive predictor in
model 3. This suggests that once other disadvantages (e.g., being His-
panic, having non-college-educated parents) were accounted for, being
an EPthat is, bilingualis an advantage in college access.
With regard to the remaining demographic characteristics, neither
immigrant status nor race/ethnicity was significant. Of interest is that
Hispanic was a significant negative predictor in models 13, but was
no longer significant once high school factors were introduced in
model 4. As with being an ELL, this suggests that the disadvantages
associated with being Hispanic are in fact attributable to
factors related to high school education rather than to being Hispanic
per se. In the final model, among the demographic characteristics,
only family income was a significant, although relatively minor, predic-
tor (b=0.09).
Results of a Regression to Predict Postsecondary Education Access From a Set of Demographic, Family Capital, and High School Factors
Model 1 Model 2 Model 3 Model 4
Demographic characteristics
ELL 1.082* 0.086* 0.687* 0.038* 0.327 0.015 0.302 0.011
EP 0.032 0.005 0.173 0.020 0.366* 0.036* 0.514* 0.044*
First generation 0.416 0.034 0.216 0.015 0.230 0.014
Second generation 0.485* 0.040* 0.302 0.022 0.145 0.009
Female 0.246* 0.062* 0.252* 0.055* 0.151 0.030
Asian 0.580* 0.042* 0.280 0.018 0.405 0.025
Hispanic 0.425* 0.045* 0.460* 0.042* 0.138 0.010
Black 0.337* 0.054* 0.055 0.008 0.345 0.041
Family income 0.339* 0.414* 0.159* 0.169* 0.096* 0.085*
Family capital
Parental education
0.536* 0.278* 0.415* 0.188*
Educational expectations of parent 0.271* 0.321* 0.157* 0.157*
Two-parent family
0.047 0.008 0.094 0.013
High school factors
Nonacademic math 1.938* 0.171*
Low math 1.203* 0.106*
Middle math 0.977* 0.191*
High school GPA 0.807* 0.224*
Math test score (12th grade) 0.017 0.066
Reading test score (12th grade) 0.013 0.054
Pseudo R
0.007 0.177 0.377 0.495
Note. *p<.05
Highest level mother or father
ELL =English language learner; EP =English-proficient linguistic minority student.
Within the family capital category, parental education level and
parental expectations for children’s PSE attainment were significant
predictors, showing stronger effects on their children’s college access
(b=0.19 and 0.16, respectively) than family income (b=0.09).
This finding underscores the critical role of the cultural and social
capital, independent of financial capabilities, coming from parents.
Family composition was not a significant predictor of college access.
Academic capital that students accrued in high school had by far
the most impact on students’ access to PSE. All high school variables
except the reading and math test scores were significant predictors in
model 4. It is interesting that the reading test score, a proxy for stu-
dents’ academic English proficiency, was not a significant predictor
after controlling for other variables. High school GPA had the largest
effect size of all the significant predictors (b=0.22). An increase of
1.0 in high school GPA was predicted to boost the level of PSE
accessed by nearly four-fifths of a level (e.g., going straight to a 4-year
college as opposed to a community college, b=0.81). The highest
level of high school math courses also was a powerful predictor. That
is, taking only nonacademic- or low-level math courses in high school
effectively closed off the possibility of advancing directly to a 4-year
institution from high school (b=1.94 and 1.20, respectively), and
even staying at middle-level math decreased the level of PSE accessed
by almost a full level (b=0.98) compared with taking advanced
level math courses, holding all other predictors constant. These results
indicate the paramount importance of both having access to college
preparatory courses in high school and doing well in these courses.
Attainment. As discussed earlier, for the attainment analyses we
included only those students who had ever entered PSE. We entered
into the regressions the same set of variables as in the access analyses,
and we added several PSE variables as predictors: full-time/part-time
enrollment status, delay in entering PSE, academic credits obtained in
the first year, and first-year undergraduate GPA (see Table 5). The
final model explained 50.2% of the variance in PSE attainment.
Unlike the access regression, linguistic background variables (ELL
and EP) were no longer significant as soon as we added other demo-
graphic variables in model 2, and they remained nonsignificant across
all subsequent models. Of the remaining demographic variables, first
generation was a positive predictor of PSE attainment in all models. In
the final model, being a first-generation immigrant boosted one’s PSE
attainment by more than one level (b=1.07). Even second generation,
although no longer significant in model 4, also reflected a similarly
positive trend, suggesting the influence of immigrant optimism (Kao
& Tienda, 1995). None of the race/ethnicity variables were significant
Results of a Regression to Predict Postsecondary Education Attainment From a Set of Demographic, Family Capital, High School, and PSE Factors
Model 1 Model 2 Model 3 Model 4 Model 5
Demographic characteristics
ELL 0.875* 0.060 0.596 0.028 0.427 0.019 0.396 0.016 0.262 0.009
EP 0.458* 0.081 0.345 0.043 0.242 0.028 0.084 0.009 0.200 0.018
First generation 1.672* 0.145* 1.578* 0.129* 1.228* 0.090* 1.072* 0.068*
Second generation 0.693* 0.064* 0.615* 0.054* 0.482 0.037 0.535 0.036
Female 0.310* 0.081* 0.351* 0.088* 0.340* 0.076* 0.174 0.034
Asian 0.660* 0.054* 0.952* 0.075* 0.844* 0.063* 0.317 0.020
Hispanic 0.738* 0.076* 0.743* 0.072* 0.526* 0.044* 0.175 0.013
Black 0.392* 0.062* 0.520* 0.079* 0.342 0.044 0.258 0.028
Family income 0.217* 0.252* 0.074* 0.081* 0.084* 0.079* 0.100* 0.081*
Family capital
Parental education
0.382* 0.221* 0.291* 0.147* 0.225* 0.099*
Educational expectations of parent 0.165* 0.187* 0.063* 0.062* 0.091* 0.076*
Two-parent family (1 =two-parent) 0.166 0.030 0.108 0.017 0.271 0.036
High school factors
Nonacademic math 0.535* 0.035* 0.607 0.030
Low math 0.954* 0.081* 0.760* 0.052*
Middle math 0.594* 0.132* 0.475* 0.092*
High school GPA 1.117* 0.325* 0.563* 0.137*
Math test score (12th grade) 0.020* 0.084* 0.005 0.016
Reading test score (12th grade) 0.011 0.048 0.008 0.031
PSE factors
Part-time status (1 =part-time) 0.889* 0.164*
Delay in entering PSE 1.268* 0.173*
Credits earned in 1st year of PSE 0.052* 0.208*
First-year undergraduate GPA 0.686* 0.208*
Pseudo R
0.010 0.096 0.183 0.346 0.502
Note. *p<.05
Highest level mother or father
ELL =English language learner; EP =English-proficient linguistic minority student.
in the final model. Interestingly, however, unlike in the access regres-
sion, Asian ethnicity and Hispanic ethnicity were significant and nega-
tive predictors in model 4, even after high school factors were added,
and were nonsignificant in model 5 only after PSE factors had been
added to the regression. This pattern suggests that Asian and Hispanic
students’ persistence in college was hampered by challenges at the
PSE level, such as being a part-time student and delaying entry to
Family income remained a significant predictor in all models and
was about the same size in the attainment regression (b=0.08) as it
had been in the final access model. In contrast, parental education
level and parental educational expectations, although significant,
lost much of the impact they had had for college access (b=0.10
and b=0.08, respectively, in model 5). This would suggest that when
it comes to finishingas opposed to enteringcollege, the ability to
pay for one’s college education remains critical while the influence of
parents’ cultural and social capital diminishes in importance.
Among the high school predictors, low-level math (b=0.76),
middle-level math (b=0.48), and high school GPA (b=0.56) were
significant predictors with effects in the expected direction. Nonaca-
demic-level math had no bearing on college attainment, most likely
because of its small cell size (i.e., most nonacademic-level math takers
did not advance to PSE and therefore were eliminated from the
attainment regression analysis). As in the access regression, neither
12th-grade math test scores nor 12th-grade reading test scores were
significant predictors in the final attainment model.
All four PSE factors showed large effects on college attainment.
Both being a part-time student and delaying entry into PSE after high
school graduation had strong negative influences on finishing college,
and undergraduate GPA and the number of academic credits in the
first year of PSE were strong positive predictors of college completion.
Being a part-time student alone set back a student’s chances for
degree attainment by almost a full level (b=0.89). On the other
hand, an additional 10 academic credits completed in the first year of
PSE boosted the level of degree obtained by 50% (b=0.05), while an
increase of 1.0 in first-year GPA increased the level of degree obtained
by almost 70% (b=0.69). These findings indicate the importance of
continuing on to PSE right after high school and having a strong
and continuing engagement with academic aspects of college after
Finally, the combination of the regression analyses (Tables 4 and 5)
and the descriptive statistics (Table 1) shows that for all of the vari-
ables that contribute positively to either PSE access or attainment,
ELLs were disadvantaged compared to EMsthe only exception was
for first-generation immigrant status. EPs were more similar to EMs
than to ELLs in terms of the social, economic, and cultural capital
they possessed; however, on the whole, EPs had fewer resources than
EMs. This pattern explains why EPs’ access and attainment levels never
quite reach those of EMs, but resemble the access and attainment lev-
els of EMs more than those of ELLs.
With regard to the first research question (Are ELLs’ patterns of
PSE access and attainment different from those of EMs and EPs?), the
three groups each show a distinct pattern of college-going. EMs have
the highest levels of PSE access and attainment, and ELLs lag far
behind both EMs and EPs. EPs’ patterns of PSE access and attainment
fall somewhere between those of EMs and ELLs, but resemble the pat-
terns of EMs more than those of ELLs.
Regarding the second research question (To what extent does lin-
guistic background predict level of access and level of attainment com-
pared with other, nonlinguistic variables?), one key finding of the
study is that being an ELL was not a significant predictor in both
access and attainment after controlling for other variables. EP, on the
other hand, was a positive predictor of PSE access. In other words,
although as far as the descriptive statistics are concerned, ELLs and
EPs clearly lag behind EMs in both PSE access and attainment, what
causes these disparities is not their linguistic background per se, but
other nonlinguistic factors that are associated with them. This is fur-
ther confirmed by the fact that reading test scores, a proxy of students’
academic English proficiency, had no bearing on either college access
or attainment net of other factors.
It was particularly interesting to find EP being a positive predictor
of college access. Our finding echoes the findings of Stanton-Salazar
and Dornbusch (1995), who report the advantage of being highly pro-
ficient bilinguals among Mexican-origin high school students: “Highly
bilingual students may have an advantage over working-class English-
dominant students in gaining access to social capital” (p. 131). Follow-
ing Stanton-Salazar and Dornbusch’s theory, we can conjecture that
EPs’ proficiency in English may facilitate their access to the support of
institutional agents such as teachers and counselors; this represents
institutional social capital that their ELL counterparts may struggle to
access because of their limited English proficiency. On the other
hand, EPs may also enjoy access to the social capital available in their
ethnic communities because of their proficiency in their ethnic
language. The resources and information that EPs can obtain through
their dual (mainstream and ethnic) social networks, bolstered by their
bilingualism, may in turn give them an advantage in PSE access com-
pared to monolingual English speakers.
A variety of nonlinguistic variables contribute to differential access
and attainment levels. Rigorous academic preparation in high school
constitutes a critical form of capital for higher levels of access and
attainment (Adelman, 2006). This is of particular importance given
that ELLs are more likely to be placed in non-college-bound tracks
(Callahan, 2005; Callahan et al., 2010). Callahan et al. (2010) analyzed
the Education Longitudinal Study of 2002 (ELS:2002) and found that
linguistic-minority students who were placed in ESL programs in high
school achieved lower cumulative GPAs and had less access to college
preparatory courses than linguistic-minority students who were not
placed in ESL programs even after controlling for their English proficiency.
In other words, ESL placement itself has a direct, negative impact on
course-taking and GPAtwo forms of academic capital that have a
major impact on students’ viability in college. Consistent with Callahan
et al.’s claims, descriptive statistics in Table 1 show that a much smaller
portion of ELLs (27%) took advanced-level math in high school than
EPs (40%) and EMs (39%), and also earned lower GPAs (2.61 com-
pared to 2.71 and 2.72, respectively). These disparities are likely to
have contributed to ELLs’ lower access and attainment levels.
The relative importance of family income versus other forms of
parental capital show an interesting trend. We find it encouraging that
parental educational level and educational expectations of parents
were at least as important as, if not more important than, family
income for both access and attainment. Our results suggest that paren-
tal sources of support that can be provided by families at any income
level can significantly boost children’s chances for college education.
At the same time, parents had less of an impact on their children’s
PSE attainment than on their access. This makes sense, given that
most students become more independent of their parents once they
are in college. Although they may have received substantial parental
help in getting into college, they need to depend more on their own
agency to persist through college. On the other hand, the impact of
family income remained relatively constant through access and attain-
ment, underscoring the importance of the ability to pay for college
throughout one’s college career.
In addition, we found that all four PSE factors were significant pre-
dictors of college attainment, suggesting that what students do after
they have been accepted by a college influences their chances for their
retention (Adelman, 2006). Some choices that students make put
them at risk: enrolling part-time, delaying entering PSE, and complet-
ing only a few credits in their first year of PSE. It is likely that financial
hardship motivates some of these choices. Students enroll on a part-
time basis, delay entering college, and/or complete only a small num-
ber of courses in the first year because they need to work to support
themselves (Almon, 2010; Walpole, 2003). Although employment is
clearly necessary for some students in order to finance their college
education, working many hours while in college is a known risk factor
(Almon, 2010; Astin, 1985). We found that being a part-time student
alone reduces one’s chances for PSE attainment by almost a full level;
that is, a student who could otherwise obtain a bachelor’s degree may
in fact attain only an associate’s degree merely because he or she is a
part-time student. Similarly, delaying entry into PSE for any period of
time after high school graduation can set back one’s attainment by
more than a full level. These findings point to the importance of pro-
viding concrete financial aid information to students, so that those
who are eligible can actually receive financial aid.
Finally, we found that being a first-generation immigrant was a posi-
tive predictor for PSE attainment, but not for access. Together with
the trends for family capital predictors, we interpret these results to
mean that the hands-on help and guidance that parents can provide
to their children make a difference in seeking admission to college,
and that in this respect first-generation immigrant parents, who are
generally unfamiliar with the U.S. educational system, may not provide
their children with any advantage over their second- and third-genera-
tion immigrant counterparts. On the other hand, when it comes to
persisting through and graduating from college, it is the students’ own
determination and agency, rather than parental help, that make a dif-
ference. In this regard, two factors may help first-generation immi-
grant students prevail to complete college: the faith in the American
dream that they share with their parents, and the sense that they owe
it to their parents to obtain a college degree, given the enormous sac-
rifice that the parents made to bring them to the United States.
Before we conclude the discussion, we want to note several limita-
tions of this study. One major limitation is that the first round of data
collection in NELS was conducted 24 years ago. Since that time, we
have seen major changes in education policy (such as the No Child
Left Behind Act of 2001
), major changes in the demographics of the
No Child Left Behind (NCLB) is a U.S. federal education law that went into effect in
2002 and holds states, school districts, and schools accountable for the academic achieve-
ment of all students. ELLs are one of the subgroups of traditionally underserved
students whose standardized test scores under NCLB must be reported separately.
school-age population (e.g., growth in ELLs and non-White students)
and in educational attainment (e.g., decreasing high school dropout),
and major changes in college-going patterns (with many more PSE
institutions at all levels and more college students). Without more
recent longitudinal data, it is impossible to say how these changes have
affected the educational persistence of ELLs. We look forward to the
release of data from the postsecondary phases of ELS:2002 in order to
answer these important policy questions.
Other limitations include the exclusion from NELS of ELLs with
very low English language skills in eighth grade. Therefore, the results
of this study, sobering as they are, are in fact likely to be somewhat
optimistic, because ELLs included in the data possessed, at the very
least, minimal English proficiency. There was also a relatively large
amount of missing data, especially data for ELLs. Although the use of
FIML for our analyses compensates for missing data to a certain
extent, it is no substitute for actual data.
One final limitation is that we were not able to determine predic-
tors of PSE access and attainment for each linguistic background
group (ELLs, EPs, and EMs) separately. The predictors that we identi-
fied in this study are for the entire sample and may or may not be sig-
nificant predictors of college access and degree attainment for the
subgroups within the sample. We attempted to run a multigroup analy-
sis, but due to the small sample size of ELLs, the regressions failed to
converge. In other words, it is not possible to examine the strength
and statistical significance for each variable as a predictor of PSE
access and attainment for each group separately in a multigroup
model using our set of variables.
This study statistically confirmed what many of us who work closely
with ELLs know experientially: that ELLs’ educational opportunities
are severely limited and that ELLs who attend college and earn a
degree represent exceptions rather than the rule. Nonetheless, the
degree of the disparities between ELLs and non-ELLs was surprising
even to us. We began the project predicting that we would find ELLs
lagging behind non-ELLs in both college access and degree attain-
ment, but we did not expect the gap to be so large. Only one in eight
ELLs in this sample eventually earned a bachelor’s degree, whereas
one in four EPs and one in three EMs did. Moreover, roughly half of
ELLs never participated in PSE at all. From the point of educational
equity, such large access and attainment gaps are unacceptable.
Perhaps less intuitively obvious is the finding that it is not the lack
of linguistic capital per se that holds ELLs back; rather, it is a multi-
tude of disadvantages that tend to co-occur for ELLs that hinder their
college access and attainment. Because limited English proficiency is
the defining characteristic of ELLs, we tend to view them unidimen-
sionally, attributing their educational challenges, including lower rates
of college-going, to their inadequate academic English proficiency.
This view in turn suggests a seemingly straightforward solution: If we
help ELLs acquire sufficient English proficiency, all their academic dis-
advantages will disappear. However, the results of this study suggest
that as far as ELLs’ increased participation in higher education is con-
cerned, any policy or program that focuses solely on their academic
English developmentindisputably important though it isis likely to
be met with limited success because it would leave other, nonlinguistic
barriers unaddressed. What we found in this study prompts us to view
ELLs as multidimensional individuals whose educational opportunities
are shaped by a variety of forms of capital, not just by their English
Among the factors that can potentially impact a student’s chances
for higher education, some are nonmalleable (i.e., schools and educa-
tors have no control over them) and others are malleable (i.e., schools
and educators can alter the condition itself or at least mitigate its neg-
ative impact). The results of our study show that malleable factors
were the major predictors of college access and attainment. First, this
study found parental support to play a crucial role in students’ access
to college. It is true that we cannot change the education level of
ELLs’ parents; however, we can provide support and necessary infor-
mation to ELLs’ parents so that they can participate more actively in
their children’s college application process as equal partners with
school personnel. Yet there is clear evidence in the literature that
most schools have done very little to promote immigrant parent
involvement (Arias & Morillo-Campbell, 2008; Arzubiaga, Noguero´n, &
Sullivan, 2009). Thus, involving ELLs’ parents in their children’s
college planning is one area in which schools can make more con-
certed efforts.
Second, ELLs’ academic underpreparation is a condition for which
schools are directly responsible. The predominant focus of ELLs’ sec-
ondary education has been to ensure their high school graduation
(Callahan & Ga´ndara, 2004). ELLs tend to be assigned to less rigorous
courses so that they can pass enough courses to graduate. This policy
leaves ELLs markedly underprepared for the rigors of college-level work
(Callahan & Ga´ndara, 2004). Furthermore, ELLs respond to lower
expectations by underperforming academically (Callahan, 2005). In
other words, our education system, rather than helping ELLs reach
academic parity with their English-proficient peers, in fact widens the
gap by depriving them of the opportunity to be exposed to high-level
academic content and discourse. If we are serious about making college
education more accessible for ELLs, we “must begin to view English
learners as competitors in the academic arena” (Callahan, 2005, p. 322).
Third, although educators have no control over ELL families’ finan-
cial status, schools can do more to ensure that students who are eligi-
ble for financial aid and need-based scholarships receive all the
funding that is available. Applying for financial aid is known to be a
confusing process even for native-speaking students with U.S.-born
parents (Bettinger, Long, Oreopoulos, & Sanbonmatsu, 2009;
Roderick, Nagaoka, Coca, & Moeller, 2008). It is not hard to imagine
that many immigrant ELLs and their parents find the process over-
whelming, and these families may decide that working extra hours is a
more practical way to finance college education. Research shows that
if parents are given hands-on workshops that take them through the
process of applying for financial aid, they are far more likely to apply
for aid, whereas simply giving them information about financial aid
makes no difference in the application rate (Bettinger et al., 2009). If
more ELLs and their parents are provided with such hands-on help
preferably with L1 interpreters presentand are also informed about
the risks of working long hours and not being fully engaged in col-
lege, more students and parents may consider applying for financial
aid and secure all the funds that are available.
The context of this study was in the United States, but we believe
that many of the insights may apply to other English-dominant coun-
tries with ELL populations as well. Research suggests that in countries
such as Canada, the United Kingdom, and Australia, ELLs tend to lag
behind their English-proficient peers in academic achievement and
have limited access to higher education (e.g., Hammond, 2008a; Mar-
shall, 2010; Preece & Martin 2010; Simpson & Cooke, 2010; Watt &
Roessingh, 2001). Further, researchers in those countries are also grap-
pling with the challenge of presenting ELLs with more intellectually
demanding content and facilitating ELLs’ college participation (see,
particularly, Gibbons, 2008; Hammond, 2008b; Marshall, 2010). How-
ever, we are not aware of national-level data on ELLs’ college partici-
pation rates as compared to those of non-ELLs or comprehensive
analyses of predictors of ELL college participation in these countries.
We believe that it is critical to secure such data, because “experience
suggests that students who are not counted won’t count when deci-
sions are made and priorities are set” (Engle & Lynch, 2009, p. 7). We
eagerly await future research in this regard, which would allow cross-
national comparisons of ELLs’ college-going and the linguistic and
nonlinguistic conditions that contribute to it.
Finally, the results of this study compel us to share a sobering
message with our international audience. One of the most common
reasons for adults to immigrate to the United States is to give their
children a better education and a better future (Sua´rez-Orozco &
Sua´rez-Orozco, 2001). For many, the United States represents the land
of opportunity, where hard work and determination will be rewarded
(Johnson, 2006). Sua´rez-Orozco and Sua´rez-Orozco (2001) point out
that parents’ hopes for their children’s educational achievement, in
fact, intensify after immigration, once reality sets in and they realize
that their own lack of English proficiency, relevant degrees, and
U.S.-specific experiences severely limit their own social mobility in the
United States. The results of this study suggest that, unfortunately,
opportunities for children of immigrants in the United States are also
limited unless appropriate education and support are provided. Only a
small percentage of immigrant youth with limited English proficiency
in our study completed a college education. College education is
increasingly becoming a prerequisite for obtaining middle-class status
in the United States. According to the U.S. Department of Labor
(2006), two thirds of the 18.9 million new jobs created between 2004
and 2014 will be filled by workers with at least some level of postsec-
ondary education. If 50% of ELLs receive no postsecondary education,
let alone a college degree, their chances of achieving stable employ-
ment and middle-class status in the United States will be severely
This finding echoes the declassing phenomenon in the United King-
dom described by Simpson and Cooke (2010). They argue that many
immigrant youth experience downward mobility as they immigrate to
the United Kingdom, despite family aspirations for an upward trajec-
tory. Our study indicates that with regard to providing newcomers with
opportunities to receive higher education, the United States, too, is
failing to live up to its own dream. We realize that immigration is a
complex process motivated by a variety of push factors (i.e., reasons to
leave the country of origin) and pull factors (i.e., attractions of the
adopted country). However, on the basis of our findings, we believe
that local agencies, teachers, and would-be immigrants need to
become better acquainted with the realities of educational opportuni-
ties in the United States for immigrant children who come from non-
English-speaking backgrounds.
This study was supported by the Association for Institutional Research, the
National Center for Education Statistics, the National Science Foundation, and
the National Postsecondary Education Cooperative under Association for Institu-
tional Research Grant Number RG 09-141. We would like to thank Dr. Judith Stull
for facilitating our access to NELS:88 data, Dr. Linda Harklau for providing feed-
back on an earlier manuscript, and Ms. Sarah Grosik and Ms. Sara Kangas for
assisting with data analysis and copyediting.
Yasuko Kanno is associate professor in the Department of Teaching and Learning
at Temple University. She is interested in linguistic minority students’ educational
opportunities and recently coedited Linguistic Minority Students Go to College: Prepa-
ration, Access, and Persistence (Routledge, 2012). She coordinates the Master’s
Program in TESOL at Temple.
Jennifer G. Cromley is associate professor in the Department of Psychological,
Organizational, and Leadership Studies in Education at Temple University. Her
research focuses on comprehension of illustrated scientific text and reading com-
prehension with students in middle school and older. She teaches basic through
advanced graduate statistics with a focus on online learning.
Adelman, C. (1999). Answers in the toolbox: Academic intensity, attendance patterns, and
Bachelor’s degree attainment. Washington, DC: U.S. Department of Education,
Office of Educational Research and Involvement.
Adelman, C. (2006). The toolbox revisited: Paths to degree completion from high school
through college. Washington, DC: U.S. Department of Education, Office of Edu-
cational Research and Involvement.
Almon, P. C. (2010). English language learner engagement and retention in a
community college setting (Unpublished doctoral dissertation). Temple Univer-
sity, Philadelphia, PA.
Arbona, C., & Nora, A. (2007). The influence of academic and environmental fac-
tors on Hispanic degree attainment. Review of Higher Education,30, 247269.
Arias, M. B., & Morillo-Campbell, M. (2008). Promising ELL parental involvement:
Challenges in contested times. Retrieved from
Arzubiaga, A. E., Noguero´ n, S. C., & Sullivan, A. L. (2009). The education of chil-
dren in im/migrant families. Review of Research in Education,33, 246271.
Astin, A. W. (1985). Achieving educational excellence. San Francisco, CA: Jossey-
Aud, S., Hussar, W., Kena, G., Bianco, K., Frohlich, L., Kemp, J., & Tahan, K.
(2011). The condition of education 2011. Washington, DC: U.S. Government Print-
ing Office. Retrieved from
Bettinger, E. P., Long, B. T., Oreopoulos, P., & Sanbonmatsu, L. (2009). The role of
simplification and information in college decisions: Results from the H&R Block
FAFSA experiment. Retrieved from
Bourdieu, P. (1977). Cultural reproduction and social reproduction. In J. Karabel
& A. H. Halsey (Eds.), Power and ideology in education (pp. 487511). New York,
NY: Oxford University Press.
Bourdieu, P. (1986). The forms of capital. In J. G. Richardson (Ed.), Handbook of
theory and research for the sociology of education (pp. 241258). Westport, CT:
Greenwood Press.
Bourdieu, P., & Passeron, J.-C. (1990). Reproduction in education, society and culture
(R. Nice, Trans., 2nd ed.). London, England: Sage.
Bowen, W. G., Kurzwell, M. A., & Tobin, E. M. (2005). Equity and excellence in Amer-
ican higher education. Charlottesville: University of Virginia Press.
Bunch, G. C., & Endris, A. K. (2012). Navigating “open access” community
colleges: Matriculation policies and practices for US-educated linguistic
minority students. In Y. Kanno & L. Harklau (Eds.), Linguistic minority stu-
dents go to college: Preparation, access, and persistence (pp. 165183). New York,
NY: Routledge.
Burkam, D. T., & Lee, V. (2003). Mathematics, foreign language, and science coursetak-
ing and the NELS: 88 transcript data. Washington, DC: National Center for Edu-
cational Statistics.
Callahan, R. M. (2005). Tracking and high school English learners: Limiting
opportunity to learn. American Educational Research Journal,42, 305328.
Callahan, R. M. (2008). Latino language-minority college going: Adolescent boys’
language use and girls’ social integration. Bilingual Research Journal,31, 175
200. doi:10.1080/15235880802640664
Callahan, R. M., & Ga´ ndara, P. (2004). On nobody’s agenda: Improving English-
language learners’ access to higher education. In S. Michael (Ed.), Teaching
immigrant and second-language students: Strategies for success (pp. 107127). Cam-
bridge, MA: Harvard Education Press.
Callahan, R. M., Wilkinson, L., & Muller, C. (2010). Academic achievement and
course taking among language minority youth in U.S. schools: Effects of ESL
placement. Educational Evaluation and Policy Analysis,32,84117. doi:10.3102/
Deil-Amen, R., & Turley, R. L. (2007). A review of the transition to college litera-
ture in sociology. Teachers College Record,109, 23242366.
Education Week. (2009). A distinctive population. Quality Counts 2009, 15.
Engle, J., & Lynch, M. (2009). Charting a necessary path: The baseline report of Access
to Success Initiative. Washington, DC: NASH/Education Trust.
Ga´ ndara, P., & Rumberger, R. W. (2009). Immigration, language, and education:
How does language policy structure opportunity? Teachers College Record,111,
Gibbons, P. (2008). “It was taught good and I learned a lot”: Intellectual practices
and ESL learners in the middle years. Australian Journal of Language and Literacy,
31, 155173.
Goldrick-Rab, S. (2007). What higher education has to say about the transition to
college. Teachers College Record,109, 24442481.
Guglielmi, R. S. (2008). Native language proficiency, English literacy, academic
achievement, and occupational attainment in limited-English-proficient stu-
dents: A latent growth modeling perspective. Journal of Educational Psychology,
100, 322342. doi:10.1037/0022-0663.100.2.322
Hammond, J. (2008a). Challenging pedagogies: Engaging ESL students in intellec-
tual quality. Australian Journal of Language and Literacy,31, 101105.
Hammond, J. (2008b). Intellectual challenge and ESL students: Implications of
quality teaching initiatives. Australian Journal of Language and Literacy,31, 128
Harklau, L. (2000). From the “good kids” to the “worst”: Representations of Eng-
lish language learners across educational settings. TESOL Quarterly,34,3567.
Harklau, L., Losey, K. M., & Siegal, M. (Eds.). (1999). Generation 1.5 meets college
composition: Issues in the teaching of writing to U.S. learners of ESL. Mahwah, NJ:
Lawrence Erlbaum.
Hirschman, C. (2001). The educational enrollment of immigrant youth: A test of
the segmented-assimilation hypothesis. Demography,38, 317336. doi:10.1353/
Hirschman, C., & Lee, J. (2005). Race and ethnic inequality in educational attain-
ment in the United States. In M. Rutter & M. Tienda (Eds.), Ethnicity and causal
mechanisms (pp. 107138). Cambridge, England: Cambridge University Press.
Horner, B., & Trimbur, J. (2002). English only and U.S. college composition. Col-
lege Composition and Communication,53, 594630. doi:10.2307/1512118
Hossler, D., & Stage, F. (1992). Family and high school experience influences on
the postsecondary plans of ninth-grade students: A structural model of predis-
position to college. American Educational Research Journal,29, 425451. doi:
Johnson, H. B. (2006). The American dream and the power of wealth: Choosing schools
and inheriting inequality in the land of opportunity. New York, NY: Routledge.
Kanno, Y., & Varghese, M. M. (2010). Immigrant and refugee ESL students’
challenges to accessing four-year college education: From language policy to
educational policy. Journal of Language, Identity, and Education,9, 310328.
Kao, G., & Thompson, J. S. (2003). Racial and ethnic stratification in educational
achievement and attainment. Annual Review of Sociology,29, 417442.
Kao, G., & Tienda, M. (1995). Optimism and achievement: The educational per-
formance of immigrant youth. Social Science Quarterly,76,119.
Klein, S., Bugarin, R., Beltranena, R., & McArthur, E. (2004). Language minorities
and their educational and labor market indicatorsRecent trends. Retrieved
Leki, I. (2007). Undergraduates in a second language: Challenges and complexities of
academic literacy development. Mahwah, NJ: Lawrence Erlbaum.
Marshall, S. (2010). Re-becoming ESL: Multilingual university students and a defi-
cit identity. Language and Education,24,4156. doi:10.1080/09500780903194044
Martin, P. (2010). “They have lost their identity but not gained a British one”:
Non-traditional multilingual students in higher education in the United
Kingdom. Language and Education,24,920. doi:10.1080/09500780903194028
Matsuda, P. K. (1999). Composition studies and ESL writing: A disciplinary
division of labor. College Composition and Communication,50, 699721.
Matsuda, P. K., Ortmeier-Hooper, C., & You, X. (Eds.). (2006). The politics of second
language writing: In search of the promised land. West Lafayette, IN: Parlor Press.
McDonough, P. M. (1997). Choosing colleges: How social class and schools structure
opportunity. Albany: State University of New York Press.
Morales, A., Herrerra, S., & Murry, K. (2009). Navigating the waves of social and
political capriciousness: Inspiring perspectives from DREAM-eligible immigrant
students. Journal of Hispanic Higher Education. Advance online publication.
Morgan, S. L. (1998). Adolescent educational expectations: Rationalized, fanta-
sized, or both? Rationality and Society,10, 131162. doi:10.1177/1043463980
Muthe´ n, L. K., & Muthe´n, B. O. (1998-2007). Mplus user’s guide: The comprehensive
modeling program for applied researchers (5th ed.). Los Angeles, CA: Muthe´n &
National Center for Education Statistics. (n.d.). National Education Longitudi-
nal Study of 1988 (NELS:88). Retrieved from
National Clearinghouse for English Language Acquisition. (n.d.). NCELA FAQ:
How many school-aged Limited English Proficient (LEP) students are there in
the U.S.? Retrieved from
˜ez, A.-M., & Cuccaro-Alamin, S. (1998). First-generation students: Undergradu-
ates whose parents never enrolled in postsecondary education. Retrieved from
˜ez, A.-M., & Kim, D. (2012). Building a multicontextual model of Latino stu-
dents’ college enrollment: Student, school, and state-level effects. Review of
Higher Education,35, 237263. doi:10.1353/rhe.2012.0004
Pascarella, E. T., Pierson, C. T., Wolniak, G. C., & Terenzini, P. T. (2004). First-
generation college students: Additional evidence on college experiences and
outcomes. Journal of Higher Education,75, 249284. doi:10.1353/jhe.2004.0016
Perna, L. W., & Titus, M. A. (2005). The relationship between parental involve-
ment as social capital and college enrollment: An examination of racial/ethnic
group differences. Journal of Higher Education,76, 485518. doi:10.1353/
Preece, S., & Martin, P. (2010). Imagining higher education as a multilingual
space. Language and Education,24,38. doi:10.1080/09500780903343070
Razfar, A., & Simon, J. (2011). Course-taking patterns of Latino ESL students:
Mobility and mainstreaming in urban community colleges in the United States.
TESOL Quarterly,45, 595627. doi:10.5054/tq.2011.268060
Roberge, M., Siegal, M., & Harklau, L. (Eds.). (2009). Generation 1.5 in college
composition: Teaching academic writing to U.S.-educated learners of ESL. New York,
NY: Routledge.
Roderick, M., Nagaoka, J., Coca, B., & Moeller, E. (with Roddie, K., Gilliam, J., &
Patton, D). (2008). From high school to the future: Potholes on the road
to college. Retrieved from
Simpson, J., & Cooke, M. (2010). Movement and loss: Progression in tertiary
education for migrant students. Language and Education,24,5773.
Spellings, M. (2005). Academic gains of English language learners prove high stan-
dards, accountability paying off. Retrieved from
Stanton-Salazar, R. D., & Dornbusch, S. M. (1995). Social capital and the repro-
duction of inequality: Information networks among Mexican-origin high school
students. Sociology of Education,68, 116135. doi:10.2307/2112778
Strang, W. E., Winglee, M., & Stunkard, J. (1993). Characteristics of secondary-school-
age language minority and limited English proficient youth. Arlington, VA: Develop-
ment Associates.
Sua´rez-Orozco, C., & Sua´rez-Orozco, M. M. (2001). Children of immigration.
Cambridge, MA: Harvard University Press.
Swail, W. S. (2000). Preparing America’s disadvantaged for college: Programs that
increase college opportunity. New Directions for Institutional Research,107,85
101. doi:10.1002/ir.10706
Swail, W. S., Cabrera, A. F., Lee, C., & Williams, A. (2005). Pathways to the bachelor’s
degree for Latino students: Latino students and the educational pipeline, part III.
Stafford, VA: Educational Policy Institute. Retrieved from http://www.eric.ed.
Swartz, D. (1997). Culture and power: The sociology of Pierre Bourdieu. Chicago, IL:
University of Chicago Press.
U.S. Department of Labor. (2006). America’s dynamic workforce: 2006. Retrieved
Walpole, M. (2003). Socioeconomic status and college: How SES affects college
experiences and outcomes. Review of Higher Education,27,4573. doi:10.1353/
Walpole, M. (2007). Economically and educationally challenged students in higher
education: Access to outcomes. ASHE Higher Education Report,33(3), 1113.
Watt, D., & Roessingh, H. (2001). The dynamics of ESL drop-out: “Plus c¸a
changeCanadian Modern Language Review,58, 203222. doi:10.3138/cmlr.
Wolf, M. K., Herman, J. L., Bachman, L. F., Bailey, A. L., & Griffin, N. (2008). Rec-
ommendations for assessing English language learners: English language proficiency mea-
sures and accommodation uses (CRESST Report 737). Los Angeles, CA: National
Center for Research on Evaluation, Standards, and Student Testing. Retrieved
Zamel, V. (1995). Strangers in academia: The experiences of faculty and ESL stu-
dents across the curriculum. College Composition and Communication,46, 506521.
Variable Type Source
Dependent variable
PSE access Categorical (13) Derived from REFTYPE
PSE attainment Categorical (15) Derived from HDEG and F4HSDIPL
Independent variable
EM Dichotomous Derived from BYS17 and BYS21
EP Dichotomous Derived from BYS17, BYS22, and
ELL Dichotomous Derived from BYS17, BYS22, BYS27A-D,
BYT1_12, and high school transcripts
Immigrant generation Categorical (13) Derived from BYP11, BYP14, and BYP17
Female Dichotomous SEX
Appendix (Continued)
Variable Type Source
Race/ethnicity Categorical (14) Derived from RACE
Family income Continuous F2P74
Parental education Continuous F2PARED
Educational expectations
of parent
Continuous BYP76
Two-parent family Dichotomous Derived from BYFCOMP
Highest level of high
school math completed
Categorical (14) Derived from high school transcripts
High school GPA Continuous HSGPA
Math test score
(12th grade)
Continuous F22XMTH
Reading test score
(12th grade)
Continuous F22XRTH
Part-time status Dichotomous STUPTANY
Delay entering PSE Dichotomous Derived from DELAY
Credits earned in
first year of PSE
Continuous TCREDG
First-year undergraduate
Continuous GPA1
Panel weight F4PNLWT
Note. PSE =postsecondary education; EM =English-monolingual student; EP =English-profi-
cient linguistic minority student; ELL =English language learner.
... In the largest study of L2 student access to and achievement in U.S. higher education at the time of this study, Kanno and Cromley (2013) used data from the National Education Longitudinal Study (NELS) of 1988 to articulate differences in access and achievement between L1 and L2 students. The authors explained about 20% of L2 students were high school dropouts, rendering it nearly impossible to pursue postsecondary education, whereas only 6% of L1 students dropped out of high school. ...
... The authors explained about 20% of L2 students were high school dropouts, rendering it nearly impossible to pursue postsecondary education, whereas only 6% of L1 students dropped out of high school. Ultimately, Kanno and Cromley (2013) found that 12.5% of L2 students earned a bachelor's degree, compared to 33% of L1 students who earned a bachelor's degree from the same NELS 1988 cohort. ...
Full-text available
Colleges continue to use technology to connect students to information, but a research gap exists regarding how colleges use a ubiquitous technology in the business world: chatbots. Moreover, no work has addressed whether chatbots address Spanish-speaking students seeking higher education in the form of automated (AI) chatbot responses in Spanish or Spanish-programmed chatbots. This study randomly sampled 331 United States institutions of higher education to learn if these institutions embed chatbots on their undergraduate admissions websites and if these chatbots have been programmed to speak Spanish. Results suggest 21% of institutions (n=71) embed chatbots into their admissions websites and only 28% of those chatbots (n=20) were programmed to provide Spanish-language admissions information. Implications for college access and equity for English learners and L1 Spanish speakers are addressed.
... This present study included a sub-sample of 8,790 students, who responded to the base year, first follow up, second follow up, and third follow up surveys and had non-missing responses to items related to math attitudes and math self-efficacy. Using a similar classification system as described by Kanno and Cromley (2013), respondents were classified as English Language Learner if they met any one of the following criteria: (1) respondents indicated that English was not their first language and their complete high school transcripts indicated they took a course with the following labels: "English as a Second Language (ESL)," "English language (EL)," "English Language Learner (ELL)," "English language development (ELD)," "Limited English Proficiency (LEP)," "Sheltered (integration of native language and content instruction)," and "Specially Designed Academic Instruction in English (SDAIE)" (Callahan et al., 2010;Finkelstein, Huang, & Fong, 2009); (2) respondents indicated that English was not their first language and they reported they did not read, speak, write, and/or understand English very well; (3) respondents indicated that English was not their first language and their teacher reported the student was behind in math due to limited English proficiency (LEP). ...
... First, this study only used data from respondents who participated in all four data collection waves (i.e., base year, first follow up, second follow up, and third follow up). Students who respond to multiple data collection waves and persist through a longitudinal study are likely to persist through their educational and career goals (Kanno & Cromley, 2013). Thus, limiting the sample to those who participated in all data collection waves may show a more optimistic perspective than if all the respondents in the base year completed all waves of data collection. ...
Background For many years now, there have been many job vacancies in science, technology, engineering, and mathematics (STEM), but not enough workers to fill these vacancies. Much attention has been given to understanding and changing this situation in our country. Purpose The purpose of this study is to address this dilemma by understanding what may be gained by investigating student's attitudes towards STEM in high school. Specifically, we study the relationship between students’ math attitudes and math self-efficacy beliefs and their career outcomes in STEM. Further, we do this across different English proficiency levels to see if any understanding may be gained by studying these groups differently. Research Design This study implemented secondary analysis by using a nationally representative sample of U.S. 10th graders from the Education Longitudinal Study. A latent class analysis was used to classify students’ math attitudes and self-efficacy. Results The results from this study provide empirical support suggesting that across all three English proficiency groups, students with high math attitudes and high math self-efficacy were more likely to have a career in STEM. When examining demographic characteristics, female students were more likely to have lower math attitude and lower math self-efficacy, which helps to explain why there is an underrepresentation of female students in STEM fields. We also found that race/ethnicity and socioeconomic status operated differently for each of the English proficiency groups. Conclusions/Recommendations This study directly links student math attitudes and self-efficacy to later career choice. This study has implications for researchers and policymakers who are developing interventions, suggesting that fostering positive math attitudes and self-efficacy would help encourage more students to pursue careers in STEM, particularly for non-native English speakers and female students.
Using data from the Wabash National Study of Liberal Arts Education, this study examined non-native English-speaking (NNES) students on selected college experiences and outcomes. The results suggested that NNES students were less likely to have close relationships with faculty and satisfying relationships with peers. They also experienced greater difficulties in making friends and worked for longer hours compared to other students. However, NNES students tended to have higher college GPA than their peers. Additionally, there was no significant difference between NNES students and their peers in bachelor’s degree attainment and graduate degree plans.
Students who have remained classified as English Learners (ELs) for more than six years are often labeled “Long-term English Learners” (LTELs). The present study examined the English Language Development (ELD) test scores and demographic information in a group of 560 students identified as LTELs. Despite assumptions that these students are still learning English, results showed many students who are labeled LTELs exhibited advanced English skills, especially on measures of expressive and receptive oral language (i.e., speaking and listening subtests). At the same time, ELD assessments showed many of these students struggled with literacy skills, especially reading. Perhaps due to these overlapping circumstances, we found many LTELs were also identified with learning disabilities. Based on these findings, we explored the impact of restricting domains needed for reclassification as English proficient on reclassification rates. Compared with existing decision rules in the students’ state, proposed models allow many more LTELs to reclassify as English proficient, and most LTELs not reclassifying are students in special education. Discussion focuses on interpreting ELD scores for students who have remained classified as ELs for more than a few years.
With the rapid increase of English Learners (ELs) in K–12 schools, school districts are struggling to find ways to meet the needs for EL teachers. One approach to address the shortage is to build teacher capacity by collaborating with higher education institutions where English as a Second Language (ESL) teacher preparation programs are offered. However, such collaborations are expensive to local schools due to the credit hours that those programs require. In this paper, comparing the contexts in the State of Michigan and the State of New York, we describe a partnership experience between a university in Michigan and its neighboring K–12 partner school districts. In 2016, the collaboration secured a five-year, 2.53 million, grant to support districts’ efforts to address such teacher shortage. Using Richardson’s (1994) crystallization method, we identified the unique features of three evolving stages of the school district’s capacity-building process. We conceptualized these stages into a two-layered model, based on the partners’ discourse patterns, role played, ownership, and information flow. We argue that the model can be used by other K–12 higher-education collaborations, particularly in the States like New York and Michigan. Specific recommendations are offered to maximize such collaborative efforts.
Although current and former English Learner (EL) or “ever-EL” students comprise one of the fastest-growing K-12 populations, we still know relatively little about the factors that influence their college-going. Using Perna’s seminal college-going model as a launching point, we propose a policy-driven empirical approach to explore how state and federal policy uniquely inform ever-EL students’ academic trajectories. This model considers how EL education policy is largely defined at the federal level but interpreted and implemented by state and local actors (i.e., the Lau and Castañeda cases). In addition, largely of immigrant origin, ever-EL students are directly affected by federal immigration policy as well as state immigrant policies. We suggest that the unique status of EL education in K-12 schools and the framing of immigrant-origin communities in federal and state policies make it necessary to consider both federal and state policy contexts in ever-EL college-going research.
Full-text available
Making decisions and constructing arguments with scientific evidence and reasoning are essential skills for all members of society, especially in a world facing complex socioscientific issues (climate change, global pandemics, etc.). Argumentation is a complex linguistic practice, but little is known about how students from diverse language backgrounds engage in argumentation. The goal of this study was to identify how students’ English language proficiency/history was associated with the reasoning demonstrated in their written arguments. We found that students with lower English proficiency and less English history produced fewer causal responses compared to students with higher English language proficiency and history. Follow-up interviews with fifteen participants revealed that students’ comfort communicating in English on assessments depended on a combination of general and academic language experiences. Findings suggest a need to identify what barriers students from diverse language backgrounds encounter during argumentation to ensure students from all language backgrounds have equitable opportunities to demonstrate their abilities.
Background/Context Language-minoritized and emergent-bilingual (EB) students have historically and frequently been underexamined in the context of research on minoritized students’ pathways in higher education. Understanding the school to college pipeline for emergent bilinguals (EBs) is becoming a critical area of study to help identify and address the barriers that they experience as they attempt to transition to and navigate postsecondary education. Despite there being a greater knowledge of the barriers experienced by EBs in getting to college, less is known about the resources they bring and their agency, the way they actually mobilize the resources that they possess in negotiating their success to get to and complete college. Purpose/Research Question This study examines why and how some EB students can successfully navigate their environments in order to apply for, get into and complete a selective four-year college. It is guided by two overarching questions: (1) What forms of capital do first generation immigrant EBs draw on to apply for and navigate selective four-year college? (2) How do first generation immigrant EBs navigate and complete selective four-year college? Research Design We examined the pathways of EBs through a conceptual framework which frames their college success as being a result of the relationship between what we refer to as their college capital which they have access to and that they draw on, and their constraint agency. Through interviews, this study analyzes 33 first generation undergraduate immigrant EBs’ transition to and completion of tertiary education, with further analysis being supplemented with in-depth case studies of five out of the 33 EBs. Additionally, we interviewed 14 university administrators and instructors involved in the admission and instruction of EB students on campus. Conclusions/Recommendations EB immigrant students drew on different forms of college capital, which included traditional and non-traditional. Students who drew more on traditional kinds of capital participated more in high participatory agentive ways while students who drew more on non-traditional forms of college capital participated more in low participatory agentive ways. Both forms of participating (low and high) lead to students navigating and completing four-year college. We suggest that more differential forms of help, resources and EB-student–focused partnerships between high school, community colleges, and four-year college which include working on their agentive selves are needed as well as challenging the racism and linguicism that holds White monolingual students as the norm to configure policies and services that will help EBs’ postsecondary pathways.
Context/Background Currently, chances for English learners (ELs) to reach higher education in the United States are slim. Almost half of ELs do not attend postsecondary education (PSE), and access to four-year college is particularly limited, but we do not exactly know why. Purpose To examine what inhibits ELs’ four-year-college access in the United States, Bourdieu's notion of habitus and a related concept of institutional habitus were used as the theoretical framework. Research Design A longitudinal, ethnographic investigation. The study tracked the college choice experiences of two high-performing ELs who nonetheless elected to attend a local community college without applying to a single four-year institution. Data consist of interviews with the students and key staff members, classroom observations, and relevant documents. Data Analysis The data on each EL were first qualitatively analyzed to create an overall picture of her college trajectory (within-case analysis); the cases were then compared with one another to identify common barriers to their college access (cross-case analysis). Data segments related to the school's institutional habitus and the students’ individual habitus were extracted and coded, and patterns of the interplay between the two were identified. Results Three factors inhibiting ELs’ four-year college access were identified: (a) limited access to advanced-level college preparatory courses; (b) underdeveloped college knowledge to effectively navigate college planning and application; and (c) linguistic insecurity about their English proficiency. The school's institutional habitus highlighted ELs’ linguistic deficits and inclined educators to view high-performing ELs as community-college-bound. The students themselves internalized the deficit orientation and came to view community college as the only possible college choice for them. Conclusions A fundamental reexamination of the deficit orientation to ELs’ linguistic and academic capabilities is necessary. ELs need to be placed in advanced college preparatory courses commensurate with their abilities and provided with regular, frequent, and accessible college guidance.
Full-text available
With some 460 languages spoken across the land, the United States has a deep reservoir of linguistic diversity. But our nation's inconsistent language-learning policies and practices present a variety of obstacles for learning English. Understanding and then addressing student needs during the critical transition phase for newcomer students is an important area for intervention. The shared fortunes of immigrant and native citizens alike will be tied to successfully linking our youngest new Americans to the educational and economic opportunity structure, to civic belonging, and full democratic participation.
The abstract for this document is available on CSA Illumina.To view the Abstract, click the Abstract button above the document title.
Background/Context This review focuses on the transition to college literature in sociology published since 1983 with an emphasis on revealing the contribution that sociology has made to our understanding of under-represented U.S. populations and their transition into and completion of postsecondary education. Purpose The review is organized around four main themes: 1) college preparation, 2) college access, 3) financing college, and 4) college completion and/or retention. Five dimensions that cut across these themes are emphasized: 1) disadvantaged or underrepresented students, 2) parents, families, and social networks of these students, 3) institutions, 4) federal, regional, state, local, or other policies, and 5) systemwide or interactive factors. Research Design This is an analytic essay of prior analyses. These prior analyses include but are not limited to a range of methods, such as qualitative case study and secondary analysis of national, regional, and institutional data. Findings/Results This review finds that while most sociological research has focused on college preparation, with disadvantaged students at the center of this work, very little research has studied college financing. Conclusions/Recommendations Sociological studies relevant to the transition to college continue to strive toward that end, but the field still remains underdeveloped with regard to an emphasis on how the wider societal system of stratification and opportunity interact with individuals, social groups, and educational institutions in a dynamic interplay that affects opportunities for quality educational advancement. In some respects, the prominence of the status attainment framework has limited progress in the field of sociology. Although multilevel modeling affords the opportunity to consider not just the individual, but the individual embedded in particular educational contexts and other contexts, the role of institutional and systemwide factors requires further development among sociologists of education.