ArticlePDF Available

Early oral language and later reading development in Spanish-Speaking English language learners: Evidence from a nine-year longitudinal study

Authors:

Abstract and Figures

Using nationally-representative, longitudinal data on a cohort of Spanish-speaking English language learners in the U.S., this study investigated the extent to which early oral language proficiency in Spanish and English predicts later levels and rates of growth in English reading. Latent growth models indicated that both Spanish and English proficiency in kindergarten predicted levels of English reading in third through eighth grade, but that only English proficiency was uniquely predictive. English productive vocabulary was found to be a better predictor of later English reading than more complex measures, i.e., listening comprehension and story retell, contrary to findings for native English speakers. Oral language did not predict later growth rates. Findings suggest the need for educational efforts to develop oral language during early childhood for this underserved population. Findings further suggest that such early efforts may be necessary, but insufficient to accelerate ELLs' reading trajectories as they move into adolescence.
Content may be subject to copyright.
Running head: ELLS’ EARLY ORAL LANGUAGE AND LATER READING
Published as: Kieffer, M. J. (2012). Early oral language and later reading development in
Spanish-speaking English language learners: Evidence from a nine-year longitudinal
study. Journal of Applied Developmental Psychology, 33(3), 146-157.
doi:10.1016/j.appdev.2012.02.003
Final published version available at:
http://www.sciencedirect.com/science/article/pii/S0193397312000238
Early Oral Language and Later Reading Development in Spanish-speaking English
Language Learners: Evidence from a Nine-Year Longitudinal Study
Michael J. Kieffer
Teachers College, Columbia University
Early Oral Language and Later Reading Development in Spanish-speaking English
Language Learners: Evidence from a Nine-Year Longitudinal Study
Abstract
Using nationally-representative, longitudinal data on a cohort of Spanish-speaking English
Language Learners in the U.S., this study investigated the extent to which early oral language
proficiency in Spanish and English predicts later levels and rates of growth in English reading.
Latent growth models indicated that both Spanish and English proficiency in kindergarten
predicted levels of English reading in third through eighth grade, but that only English
proficiency was uniquely predictive. English productive vocabulary was found to be a better
predictor of later English reading than more complex measures, i.e., listening comprehension and
story retell, contrary to findings for native English speakers. Oral language did not predict later
growth rates. Findings suggest the need for educational efforts to develop oral language during
early childhood for this underserved population. Findings further suggest that such early efforts
may be necessary, but insufficient to accelerate ELLs’ reading trajectories as they move into
adolescence.
Keywords: At-risk learners, oral language, reading development, longitudinal studies
ELLS’ EARLY ORAL LANGUAGE AND LATER READING
Early Oral Language and Later Reading Growth in Spanish-speaking English Language
Learners: Evidence from a Nine-Year Longitudinal Study
The precursors of successful reading develop long before students begin formal reading
instruction (Snow, Burns, & Griffin, 1998). Substantial longitudinal evidence suggests that
native English-speaking children who have developed higher levels of oral language proficiency
by kindergarten are more successful in learning to read in the primary grades than those who
enter school with underdeveloped oral language (e.g., Catts, Adlof, & Weismer, 2006; Dickinson
& Tabors, 2001; Sénéchal & LeFevre, 2002; Storch & Whitehurst, 2002). However, far less is
known about these relationships for the growing population of students who come from homes in
which a language other than English is primarily spoken, a group known as language minority
(LM) learners (August & Shanahan, 2006). In particular, English language learners (ELLs), the
subset of this larger population that enters school not yet proficient in English, demonstrate
disproportionately low English reading comprehension in the upper elementary and middle
school grades (e.g., Kieffer, 2008, 2010; NCES, 2009), yet it is unclear to what extent these
reading comprehension difficulties are predicted by ELLs’ earlier oral language development.
In the most recent comprehensive review of research on the early predictors of later
reading proficiency, the National Early Literacy Panel (NELP) (2008) found evidence for 11
early or precursor literacy skills, measured between birth and kindergarten, that had moderate to
strong correlations with later literacy abilities. Among these, oral language, defined as the
“ability to produce or comprehend spoken language, including vocabulary and grammar,” was
classified as one of the five “potentially important variables” (p. viii). These five variables
demonstrated moderate relationships with later reading that were weaker than those of the best
ELLS’ EARLY ORAL LANGUAGE AND LATER READING 2
predictors, which included alphabet knowledge, phonological awareness and memory, rapid
automatic naming, and writing. In responding to the report, some researchers have suggested
that the finding of only a “potentially important” role for oral language may not apply to ELLs,
for whom early oral language development—in both their primary and second languages—could
be particularly essential (Gutierrez, Zepeda, & Castro, 2010; Orellana & D’warte, 2010).
Leaders of the panel have responded that the relationships found may indeed depend on ELL
status, but that the extant evidence does not make it possible to determine whether this is the case
(Lonigan & Shanahan, 2010; see also Schnatschneider & Lonigan, 2010).
The present study was designed to extend the current research base on early predictors of
later reading development by investigating the role of early oral language in later reading growth
among Spanish-speaking ELLs in the U.S., using nine years of longitudinal data on a nationally
representative sample of this population. Specifically, this research focused on the
subpopulation of Spanish-speaking language minority learners who were U.S.-educated (i.e.,
born in the U.S. or immigrated before kindergarten) and who entered U.S. kindergarten
classrooms with limited English proficiency (i.e., initial English language learners at school
entry, whether or not they later were redesignated to fluent English proficient). For this
population, the roles of Spanish and English oral language in kindergarten in predicting students’
later levels and/or rates of growth in English reading between third and eighth grade were
examined. This study further investigated whether vocabulary measures or more complex
measures of oral language are best predictors of later reading in this population.
Early Oral Language and Later Reading Outcomes in Native English Speakers
In their meta-analytic review, the NELP (2008) found that oral language proficiency
measured in kindergarten had a moderate relationship with later decoding across 50 studies as
ELLS’ EARLY ORAL LANGUAGE AND LATER READING 3
well as a moderate relationship with later reading comprehension across 23 studies.
Interestingly, the NELP (2008) found large differences in the predictive relationship as a
function of the oral language measure examined, with overall composite measures (i.e., those
that combine assessments of vocabulary, syntax, and listening comprehension in the same test)
demonstrating much stronger relationships to later decoding and reading comprehension than
measures of individual skills. In addition, listening comprehension measures that require students
to integrate vocabulary and grammatical knowledge had a notably stronger relationship to later
reading comprehension, compared to measures of receptive vocabulary, which were among the
weakest predictors in the oral language domain.
While the vast majority of extant studies measured reading outcomes in first or second
grade, several studies suggest that the predictive power of early oral language on reading
comprehension persists into later grades (e.g., Catts et al., 2006; Cunningham & Stanovich,
1997; Sénéchal & LeFevre, 2002; Storch & Whitehurst, 2002; Snow, Porche, Tabors, & Harris,
2007). For instance, Snow and colleagues (2007) found that kindergarten receptive vocabulary
had a strong and stable correlation with reading comprehension scores in Grades 4, 7, and 10 for
a sample of monolingual children from low-income backgrounds. Although later reading
comprehension was also predicted by more complex language measures, including a researcher-
created measure of formal definitional skill and a researcher-created measure of narrative
production, these measures had weaker correlations than did vocabulary and they declined in
predictive power over time. Similarly, Cunningham and Stanovich (1997) found that first-grade
receptive vocabulary exhibited a moderate relationship with reading comprehension measured 10
years later. Despite the accumulation of evidence for the importance of early oral language
ELLS’ EARLY ORAL LANGUAGE AND LATER READING 4
proficiency to first-language reading development through middle school, far less is known
about the role of early oral skills for the long-term reading development of ELLs.
Language and Reading Development in Spanish-Speaking ELLs
The rapidly growing population of ELLs in the U.S., more than 70% of whom come from
Spanish-speaking homes (Capps et al., 2005), provides new challenges and opportunities to
educators who have previously relied on models of first-language reading development. In
particular, the disproportionate prevalence of English reading comprehension difficulties among
ELLs in the middle grades (e.g., Kieffer, 2010; NCES, 2009) raises the question of whether the
precursors to these difficulties can be identified much earlier. With only a few longitudinal
studies to date that have followed ELLs beyond fourth grade (e.g., Mancilla-Martinez & Lesaux,
2010; Nakamoto, Manis, & Lindsey, 2007; Reese, Gallimore, Garnier, & Goldenberg, 2000), this
question remains largely open.
Although it is likely that the conclusions of the NELP (2008) will largely hold for ELLs,
there are some reasons to suspect that early oral language proficiency may play a somewhat
different role in the reading development of Spanish-speaking ELLs than it plays in
monolinguals’ reading development. First and foremost, ELLs’ oral language skills are
distributed across two languages, so measures of English oral language proficiency alone may
not capture the full range of linguistic resources available to these students in the process of
learning to read English (e.g., Genesee, Lindholm-Leary, Saunders, & Christian, 2006; Gutierrez
et al., 2010). Indeed, evidence from language-of-instruction studies indicates that improving
first-language literacy skills supports second-language literacy outcomes (for reviews, see e.g.,
Francis, Lesaux, & August, 2006; Genesee et al., 2006; Slavin & Cheung, 2005). However, the
evidence to date on the importance of first-language oral proficiency, as opposed to first-
ELLS’ EARLY ORAL LANGUAGE AND LATER READING 5
language literacy, for second-language reading is more limited and equivocal; Geva and
Genessee’s (2006) review, as well as two recent studies (Gottardo & Mueller, 2009; Nakamoto,
Lindsey, & Manis, 2008), found more evidence for within-language effects than cross-language
effects. As a result, researchers have been skeptical that first-language oral proficiency will
predict second-language reading or compensate for underdeveloped oral proficiency in the
second language (e.g., Bialystok, 2002; Verhoeven, 1994).
A second reason to suspect that early oral language may play a unique role for ELLs is
the great heterogeneity in basic language proficiency found among these learners. Whereas
native English speakers without clinical language impairments can be assumed to have acquired
a command of commonly used vocabulary and basic grammar by kindergarten, ELLs may vary
substantially in their command of such language (e.g., August, Carlo, Dressler, & Snow, 2005).
Consider, for instance, the words that Beck, McKeown, and Kucan (2002) label Tier 1 words, or
“mostly basic words—clock, baby, happy—rarely requiring instruction in school” (p. 16); while
most native English speakers and many ELLs arrive at school knowing these words, other ELLs
do not. The substantial limitations in early oral language in English demonstrated by many ELLs
(e.g., Manis, Lindsey, & Bailey, 2004; Swanson, Rosston, Gerber, & Solari, 2008) could
seriously constrain their English reading development. At the same time, however, ELLs, as a
group defined by limited English proficiency, do not include those who have attained English
oral language skills in the upper range of the distribution, so it is also possible that this
constrained variation could lead to weaker correlations between English oral language and later
reading in this population than in native English speakers.
Third, taking a more ecological view of reading development, ELL status is frequently
confounded with low socioeconomic status (Capps et al., 2005; Cosentino de Cohen, Deterding,
ELLS’ EARLY ORAL LANGUAGE AND LATER READING 6
& Chu Clewell, 2005) as well as limited access to educational resources supporting reading
development (Gándara, Rumberger, Maxwell-Jolly, & Callahan, 2003), particularly for Spanish
speakers. As a result of this multiplicity of risk factors, Spanish-speaking ELLs may be more
vulnerable to the negative effects of lower levels of language skills than their more advantaged
counterparts. Although the effects of ELL status, first language background, and SES cannot be
easily disentangled, there is a need—at minimum—to include statistical controls for SES when
estimating the relationship between early oral language and later reading.
In addition to the possibility that the role of oral language development, in general,
differs by language background, it is also conceivable that the subcomponents of English oral
language proficiency play differentially important roles for Spanish-speaking ELLs and native
English speakers. In particular, vocabulary measures that tap more decontextualized language
knowledge may have greater specificity in identifying important English language weaknesses in
this population, as compared to global listening comprehension measures that offer more
contextual support. This would converge with a growing body of research with older ELLs that
highlights limited English vocabulary knowledge as a very common source of difficulties with
English reading comprehension (e.g., August et al., 2005; Garcia, 1991; Hutchinson, Smith,
Whitely, & Connors, 2004; Lesaux & Kieffer, 2010; Lesaux, Crosson, Kieffer, & Pierce, 2010),
including a few longitudinal studies (Mancilla-Martinez & Lesaux, 2010; Nakamoto, Manis, &
Lindsey, 2007). In addition, whereas grammatical knowledge appears to be strongly predictive
of later reading for native English speakers (National Early Literacy Panel, 2008), a few studies
have demonstrated a weaker relationship between grammatical knowledge and reading for ELLs
in the elementary grades (e.g., Lipka & Siegel, 2007; Jongejan, Verhoeven, & Siegel, 2007).
ELLS’ EARLY ORAL LANGUAGE AND LATER READING 7
Current Study
The current study investigated the role of early oral language in predicting later reading
comprehension for Spanish-speaking ELLs in the U.S. To address this aim, data on the language
and reading development of a nationally representative sample of Spanish speakers who entered
school as ELLs and were followed from kindergarten through eighth grade was used.
Specifically, the study examines the relationship of Spanish and English measures of vocabulary,
listening comprehension, and story retell (as well as a composite of these three measures) in
kindergarten with levels and rates of growth in English reading comprehension between Grades
three and eight. By providing insight into these relationships, this study aims to build the
research base on second-language reading development that can ultimately inform efforts to
improve early childhood education for linguistically diverse populations.
Based on findings from prior research, it was hypothesized that English oral language
measures in kindergarten would demonstrate moderate relationships with later levels of English
reading comprehension as well as moderate relationships with later rates of growth in reading
comprehension. It was further hypothesized that when included together in a single model, that
Spanish oral language measures would not make a unique contribution to English reading
comprehension, beyond the contribution of English oral language. Finally, it was hypothesized
that among oral language measures, that vocabulary measures would be more predictive of later
reading comprehension than listening comprehension and story retell measures.
The following specific research questions guided this study:
1) Do kindergarten levels of oral language development in English and/or Spanish predict
later levels and/or rates of growth in English reading for Spanish speakers who enter
school as English language learners? Does Spanish oral language development uniquely
ELLS’ EARLY ORAL LANGUAGE AND LATER READING 8
predict later levels and/or rates of growth in English reading, after controlling for the
effects of English oral language development?
2) Among individual indicators of early oral language development in English and Spanish,
is productive vocabulary more predictive of later levels and/or rates of growth in English
reading compared to more complex measures (i.e., listening comprehension, or story
retell measures) in this population?
Method
Dataset
This study uses data from the Early Childhood Longitudinal Study—Kindergarten cohort
(ECLS-K; Tourangeau, Lê, Nord, & Sorongon, 2009), a study conducted by the National Center
for Education Statistics (NCES) that examines U.S. school children’s academic and behavioral
development between kindergarten and eighth grade. ECLS-K was designed to capture
information on a wide range of child, home, and school characteristics through student
assessment, parent interviews, teacher interviews, and principal surveys. In conducting ECLS-K,
NCES used a multi-stage probability sampling design to select a cohort of students that was
nationally representative of students entering kindergarten in the 1998-99 school year. The
cohort of students participated in data collection on six occasions over the nine years of the
study: fall of kindergarten, spring of kindergarten, spring of first grade, spring of third grade,
spring of fifth grade, and spring of eighth grade. ECLS-K is one of the first national longitudinal
studies to incorporate ELLs fully and appropriately into cognitive assessment and other aspects
of data collection and, as such, is well suited to addressing these research questions.
ELLS’ EARLY ORAL LANGUAGE AND LATER READING 9
Analytic Sample
To address the research questions about the population of U.S.-educated, Spanish-
speaking ELLs, the current study was conducted with a sub-sample of 295 students participating
in ECLS-K who met the following criteria: 1) parental report of Spanish as the primary home
language, 2) limited English proficiency in the fall of kindergarten, i.e., scoring below the
proficiency composite cut-score on an English oral language proficiency measure (see details
below), and 3) had a valid sampling weight indicating that they had available data on English
reading achievement for Grade 3, 5, and 8.1 Together, the first two criteria also ensured the
availability of oral language development scores in students’ L1; ECLS-K did not assess students
from other language backgrounds in their L1, nor did ECLS-K assess the Spanish skills of
Spanish-speaking students who scored above the cut-score on the English oral language measure.
The third criteria ensured that the relationship between early oral language and later reading
achievement through Grade 8 could be investigated. Although this later criteria excluded
students who dropped out of the study over time, ECLS-K provides longitudinal sampling
weights that, when included appropriately in analyses, can account for attrition and non-response
over time, as well as for the purposeful over-sampling of specific groups (e.g., students attending
private schools). All analyses reported below included the appropriate longitudinal sampling
weights and thus can support generalizations to the population of Spanish-speaking students who
entered kindergarten with limited English proficiency in 1998.
It is worth noting that this definition of Spanish speakers who entered school as ELLs is
time-invariant and thus different from the definitions often used by schools and districts, in
which “ELL” or “limited English proficient” is a temporary label that children lose when they
1 In addition to the students removed for missing these scores, 23 students were excluded for missing data on the
oral language proficiency measures in kindergarten.
ELLS’ EARLY ORAL LANGUAGE AND LATER READING 10
gain English proficiency. In this way, the population of interest may be more precisely, if less
succinctly, referred to as “language minority learners with initially limited English proficiency”
(Kieffer, 2008). This longitudinal definition has the benefit of avoiding the major cohort effects
that can occur when only the subset of students with an “ELL” designation at a particular later
point in time are included. It is also worth noting that the dataset does not include students who
immigrated to the U.S. after kindergarten; for this reason, the population of Spanish-speaking
ELLs studied is further specified as “U.S.-educated ELLs,” not to be confused with newcomer
ELL populations who have been educated outside the U.S. to some extent prior to entering U.S.
schools.
The resulting analytic sample was balanced for gender (148 boys and 147 girls). On
average, the students came from low socioeconomic backgrounds; they scored 0.78 standard
deviations below the national mean on the SES composite index (see below) in kindergarten,
with similar SES levels in Grades 1, 3, 5, and 8. On average, participating students were 5 years,
7 months old in the fall of kindergarten (SD = 4 months); 9 years, 2 months old in Grade 3 (SD =
4 months); 11 years, 1 month old in Grade 5 (SD = 5 months); and 14 years, 2 months old in
Grade 8 (SD = 5 months).
Measures
English reading achievement. Children’s overall reading achievement in English was
assessed at each testing occasion using a two-stage adaptive reading test assembled by a panel of
content experts, with items from published standardized tests, from the National Assessment of
Educational Progress (NAEP), and from two earlier longitudinal studies conducted by NCES, the
National Education Longitudinal Study of 1988 and the Education Longitudinal Study of 2002
(Najarian, Pollack, & Sorongon, 2009). Drawing on and extending the NAEP framework, the
ELLS’ EARLY ORAL LANGUAGE AND LATER READING 11
test assessed six types of reading comprehension skills (initial understanding, developing
interpretation, personal reflection and response, developing a critical stance, and evaluating
complex syntax) as well as vocabulary and, in earlier grades, basic reading skills (familiarity
with print, recognition of letters, recognition of phonemes, and decoding). A greater proportion
of difficult items targeting reading comprehension was included at later grade levels than at
earlier grade levels to avoid ceiling effects and to capture age-appropriate skills, and Item-
Response Theory (IRT) methods were used by NCES to create vertically-linked scaled scores
that would be comparable over time. Najarian et al. report high estimates of IRT-based reliability
(ranging from .87 to .96, by wave) and validity evidence, including high correlations with other
standardized measures of reading and lower correlations with mathematics and science measures
during field testing. The IRT-derived theta scores were used, because they are vertically linked
to be comparable across rounds of data collection. Najarian et al. report evidence supporting the
inference that the these theta scores measure a unidimensional construct consistently over time,
including evidence that the common linking items function the same over time, strong IRT-based
fit statistics for the items in the reading measure, and appropriate correlations across waves of
data collection. The current study focused on reading achievement in the spring of Grade 3,
Grade 5, and Grade 8. It is worth noting that ECLS-K also assessed students in reading during
the kindergarten and Grade 1 testing occasions; however, ELLs were systematically excluded
from taking the reading assessment at one, two, or all of these occasions, because their oral
English proficiency was determined to be too limited for appropriate inclusion in the English
reading assessment.
English oral language. Students’ oral language development in English was assessed in
the fall of kindergarten with three of the six subtests from the English Pre- Language Assessment
ELLS’ EARLY ORAL LANGUAGE AND LATER READING 12
Scales 2000, Form C (English PreLAS; Duncan & DeAvila, 1998). These subtests were selected
in consultation with a panel of experts and outside consultants convened by ECLS-K staff;
criteria for selecting the measure included “known psychometric properties including predictive
validity and face validity among experts in the field” as well as “widespread use and acceptance
for the age group” and similarity in format to the cognitive assessments used in ECLS-K (U.S.
Department of Education, 2002, p. 2-21, 2-22).
The first subtest was a listening comprehension task referred to as “Simon Says,” in
which students were asked to respond to simple directives (e.g., touch your ear, point to the
middle of the paper, put one hand on top of the other) in English by taking an action. The
directives were designed to include those likely to be encountered in kindergarten classrooms
and to include vocabulary that is limited to parts of the body and items commonly encountered in
household and pre-school environments (e.g., pencil, floor, paper, door). This task consisted of
10 items, each scored at one point per item.
The second subtest was a productive vocabulary task referred to as “Art Show,” in which
students were asked to provide concrete nouns and basic verbs without inflectional markers to
describe pictures. The first portion of the test asks students to name a picture, while the second
portion of the test asks students to name a picture and then tell its function or purpose (e.g,
What’s this? A book. What can you do with it? Read it.) This task also consisted of 10 items,
each scored at one point per item.
The third subtest was a story retell task, referred to as “Let’s Tell Stories,” in which test
administrators read a 75-word story aloud, provided 4 pictures representing events in the story,
and asked students to provide an oral retelling of the story, which was then scored on its
coherence and elaboration as well as the accuracy and complexity of grammar used. This task
ELLS’ EARLY ORAL LANGUAGE AND LATER READING 13
consisted of two stories, each of which was scored on a scale from 0 to 5, which was then
weighted by 4 to yield a possible range of 0 to 40 points for the subtest.
The composite score was created by summing the scores from each of the three subtests
and thus possible values ranged from 0 to 60.2 The ECLS-K psychometric report indicates that
one of the authors of the test, Dr. DeAvila recommended that a cut-score of 37 or above indicate
that students’ oral English proficiency was sufficient to complete the ECLS-K cognitive
assessments in English (U.S. Department of Education, 2002). Because this cut-score was used
to define the analytic sample in the current study as initially limited English proficient, the
current sample’s scores range from 0 to 36. The ECLS-K psychometric report for the
kindergarten through first grade (U.S. Department of Education, 2002) describes good reliability
in the current sample, including a split-half reliability estimate of .97 for the composite score as
well as a high Cronbach’s alpha for the first two subtests (i.e., estimates reported to be “mostly
mid .80’s to mid .90’s” and “very high for subtests with only ten items each”; p. 7-3). These are
consistent with the reliabilities reported by the technical guide for this form of the published
PreLAS (alpha = .88 and alternate-forms reliability = .89 for “Simon Says”; alpha = .90,
alternate-forms reliability = .94 for “Art Show”; Duncan & DeAvila, 1998). Although reliability
estimates were not reported for the story retell subtest for the current sample, the technical guide
for the PreLAS reports alternate-forms reliability of .76 and inter-rater reliability of .88.
Spanish oral language. Students’ oral language development in Spanish was assessed
with the parallel subtests of the Spanish version of the PreLAS 2000, Form C: listening
2 It is worth noting that using this simple linear composite of raw scores has both advantages and disadvantages. On
one hand, this is more analogous to the composite measures used by the studies review in the NELP report. On the
other hand, forming a latent composite from the observed subscores might have had greater reliability. It also
weights the story retell measure more heavily. However, attempts to form latent composites using confirmatory
factor analysis led to convergence problems, likely due to the relatively low correlations among the Spanish
subtests. Thus, the simple linear composite of raw scores was used instead.
ELLS’ EARLY ORAL LANGUAGE AND LATER READING 14
comprehension (i.e., “Tio Simon”), productive vocabulary (i.e., “La Casita”), and story retell
(i.e., “Contando Historias”). These subtests are parallel in format and targeted constructs to
those of the English PreLAS, but utilized different stories and stimulus pictures. The test
developers reported that the Spanish PreLAS was designed for native Spanish-speaking children,
in parallel with the English PreLAS with identical “theoretical distinctions and psychometric
foundations(DeAvila & Duncan, 2000; p. 1). ECLS-K reports a split-half reliability estimate
of .92 for the composite score and “high internal consistency” in the listening comprehension
and productive vocabulary subtests (U.S. Department of Education, 2002, p. 7-4). The technical
manual for the published PreLAS report adequate reliability for the Spanish subtests (alpha = .74
for “Simon Says;” alpha = .81 for “Art Show;” correlations among stories in the story retell task
= .87-.90).
Socioeconomic status. Students’ socio-economic status (SES) was measured in
kindergarten, Grade 1, Grade 3, Grade 5, and Grade 8 using an ECLS-provided continuous
variable that draws on several survey items regarding parents’ education, occupation, and
household income. Because these five time-varying measures of SES were all highly correlated
with one another (r = .66 - .93), they were combined using confirmatory factor analysis (CFA)
into a single latent factor representing average SES over time, which was integrated into
subsequent analyses. CFA analyses indicated that this approach is reasonable, with factor
loadings all greater than .71 and statistically significant. This approach has the additional
advantage of avoiding missing data for SES in subsequent analyses; fitting the CFA models
using full-information maximum likelihood estimation accounted for missing data at any given
occasion and yielded a factor score for every student who had a reported SES value for at least
one time point, which ultimately included all students in the analytic sample. Nonetheless, it is
ELLS’ EARLY ORAL LANGUAGE AND LATER READING 15
worth noting that the substantive results reported below were robust to whether SES was
included as a latent factor or as a set of five time-varying indicators.
Data Analyses
Latent growth modeling was used to investigate the relationship between early oral
language in English and Spanish and later status and growth in English reading. Latent growth
modeling of longitudinal data has several advantages over other methods for analyzing change
over time that are relevant to the present study (Bollen & Curran, 2006). First, by modeling
between-student variability in true intercepts directly, partialing out occasion-specific
measurement error, these methods can provide more precise estimates of the relationship
between predictor variables (i.e., English and Spanish oral language in the current study) and
students’ true initial status in the outcome (i.e., third-grade levels of English reading) than
offered by observed longitudinal correlations. Second, by modeling between-student variability
in true slopes, these methods can allow researchers to determine if predictors are associated with
variation in students’ true rates of growth in the outcome. Third, by modeling growth in a
structural equation modeling framework (as opposed to a multilevel or hierarchical linear
modeling framework), latent growth modeling allows for the inclusion of other latent factors as
covariates, an advantage that was used in the present study to provide a more powerful approach
to controlling for socioeconomic status using the five available measurements of SES. Because
there was some evidence of non-normality in the oral language proficiency variables, a sandwich
estimator (i.e, the Maximum Likelihood Robust option in Mplus 4.0) was used to estimate
standard errors that are robust to deviations from normality.
Figure 1 presents a path model representing the hypothesized latent growth model to
address the first research question. The right hand side of the figure displays a three-indicator
ELLS’ EARLY ORAL LANGUAGE AND LATER READING 16
two-factor measurement model that represents the development of English reading as a linear
function of time. The model links hypothesized latent constructs (displayed as circles)
representing true initial (Grade 3) status (λ0) and true rate of change (λ1) in English reading to the
observed indicators of English reading on each of the three occasions of measurement (labeled
Spring Grade 3 Rdg, Spring Grade 5 Rdg, Spring Grade 8 Rdg). In fitting unconditional growth
models, the variance in true rate of change was found to be nonsignificant, so this was fixed to 0.
Correspondingly, the covariance between true rate of chance and true initial status was also fixed
to 0, as shown by the absence of a path between these factors in Figure 1. The bottom left side
of the figure presents five-indicator one-factor measurement model for the SES composite,
linking a hypothesized latent construct for students’ SES to the observed indicators of SES on
each of five occasions of measurement (kindergarten SES through Grade 8 SES). The effect of
this covariate on the growth parameters for English reading is taken into account with the
inclusion of structural regression paths (represented as single-headed arrows labeled γ01 and γ11).
The question predictors of interest, observed composites for English and Spanish oral language,
are represented as rectangles in the top left of the figure. The first research question was
addressed by evaluating the significance of the structural regression paths between these
predictors and the latent growth factors for English reading (represented as single-headed arrows
labeled γ02, γ12, γ03, and γ13). Residual covariances among the predictors are also included. To
address the second research question, a similar model was fitted, with the three subtests for
English oral language and three subtests for Spanish oral language replacing the oral language
composites.
Results
Preliminary Analyses
ELLS’ EARLY ORAL LANGUAGE AND LATER READING 17
Prior to fitting the latent growth curve models, preliminary analyses were conducted to
describe the observed levels and relationships among the indicators over time. Table 1 presents
the correlations, means, and standard deviations for the observed measures of early oral language
and later reading achievement, estimated using the appropriate longitudinal sampling weights.
As shown, students’ kindergarten performance on the composite measures of oral language in
Spanish and English were highly correlated. However, cross-language correlations differed
substantially across the subtests of oral language, suggesting the need to investigate these
subtests individually.3 Small but non-trivial positive correlations were found between the
composite measure of English oral language in kindergarten and later English reading, with a
larger magnitude evident for Grade 3 reading than for Grade 5 reading or Grade 8 reading. The
English vocabulary subtest demonstrated correlations with later English reading that were
slightly higher than the English listening comprehension subtest and notably higher than the
English story retell subtest. Correlations between kindergarten Spanish oral language composite
and later English reading were quite small for Grade 3 reading and Grade 5 reading and near
zero for Grade 8 reading, while the Spanish vocabulary subtest scores had slightly higher
correlations with later reading than the other two subtests. As expected, correlations among
English reading scores were highly correlated over time.
Effects of Spanish and English Oral Language Composites on Later English Reading
The first research question focused on whether English and/or Spanish oral language in
kindergarten, when measured with composite instruments, predicted levels and/or rates of
3 It also worth noting that the Spanish listening comprehension subtest appeared to demonstrate a ceiling effect (i.e.,
a mean of 9.767 close to the total possible of 10 points and a SD of less than 1 point), presumably because most
students had mastered nearly all of the items involving simple directions in their first language. Although the
maximum likelihood sandwich estimator used produces standard errors that are robust to this violation of normality,
it is worth noting that this ceiling effect may have led to an underestimation of the relationship between Spanish
listening comprehension and English reading in subsequent analyses.
ELLS’ EARLY ORAL LANGUAGE AND LATER READING 18
growth in English reading between third and eighth grade, controlling for socioeconomic status.
To address this question, a set of latent growth models was fitted to the variance-covariance
matrix estimated using the appropriate longitudinal sampling weights. In each model, the
predictors of interest were the simple linear composite scores on the Spanish and/or English oral
language measures. The fitted models are shown in Table 2.
When included individually, English and Spanish oral language in kindergarten each
significantly predicted students’ initial (third-grade) status in English reading, as shown in
Models 1 and 2 (for English oral language, standardized path coefficient = .25; z = 3.01; p < .01;
for Spanish oral language, standardized path coefficient = .14; z = 2.00; p = .04). Neither
English oral language nor Spanish oral language in kindergarten predicted students’ later rates of
growth in English reading (for English oral language, z = -1.40; p =.16; for Spanish oral
language, z = -1.30; p = .19). When English and Spanish oral language are included together in
Model 3, English oral language in kindergarten continued to be a significant unique predictor of
students’ initial (third-grade) status in English reading (standardized path coefficient = .32; z =
2.58; p = .01), but Spanish oral language was no longer a significant unique predictor of initial
status (z = -0.95; p = .34). As in Models 1 and 2, neither English nor Spanish oral language
predicted students’ later rate of growth in English reading when included together (for English
oral language, z = -0.74; p =.46; for Spanish oral language, z = -0.33; p = .74). Across the
models, SES was a statistically significant, positive predictor of initial status, but not of the rate
of growth, indicating that students from different SES backgrounds had trajectories that were
vertically separated but parallel.
In Model 4, the effect of English oral language on students’ initial (third-grade) status in
English reading is estimated, after removing the nonsignificant effects of Spanish oral language
ELLS’ EARLY ORAL LANGUAGE AND LATER READING 19
on initial status and the nonsignificant effects of Spanish and English oral language on rate of
growth. When converted into a standardized path coefficient, the effect of English oral language
in kindergarten on initial status in reading is .20, a magnitude that indicates a weak relationship
according to the conventions of the NELP (2008) and others. However, it is worth noting that
this magnitude is roughly the same as the magnitude of the estimated effect of SES (standardized
path coefficient = .22).
Given the strong correlation between the English and Spanish oral language composites,
there is reason to consider the possibility that a single composite incorporating both Spanish and
English oral language skills would be more predictive than either composite individually. Such a
composite would serve as a language-general measure that could potentially capture students’
communicative resources as they are distributed across two languages better than measures in
each language separately. To investigate this possibility, a single latent factor was created to tap
the three subtests of the Spanish PreLAS as well as the three subtests of the English PreLAS, and
this factor was used to predict initial (third-grade) status and rate of growth in English reading.
Although this language-general composite did significantly predict students’ initial status in
English reading (z = 2.07; p = .04), this relationship was weaker (standardized path coefficient =
.17) than that found for the English oral language composite alone. On closer inspection, this
latent composite had appropriately high loadings for only the three English subtests and the
Spanish story retell task, so a second language-general composite was created with these four
tasks alone; this composite had a similarly weak relationship with students’ initial status in
reading (standardized path coefficient = .17). Details on these models are available from the
author.
ELLS’ EARLY ORAL LANGUAGE AND LATER READING 20
Effects of Early Spanish and English Listening Comprehension, Vocabulary, and Story
Retell on Later English Reading
The second research question addressed whether individual measures of early English
and Spanish oral language proficiency, as distinguished by three common types of instruments,
predicted students’ levels and/or rates of growth in later English reading. Specifically, the
relative predictive strengths of three subtests of the PreLAS—listening comprehension,
productive vocabulary, and story retelling—were investigated in Spanish and English. As with
the models that addressed the first research question, a set of latent growth models was fitted, as
shown in Table 3: Model 5 included English predictors, Model 6 included Spanish predictors,
Model 7 included English and Spanish predictors together, and Model 8 represented the final
model in which nonsignificant effects have been removed. As shown in Model 5, among the
three measures of early English oral language, productive vocabulary emerged as the only
uniquely significant predictor of students’ initial (third-grade) status in English reading,
controlling for the effects of the other English measures and socioeconomic status. This effect
remained significant when the three parallel measures of Spanish oral language were also
included, as shown in Model 7. In Model 6, none of the three measures of early Spanish oral
language were uniquely significant predictors of initial status or rates of growth, controlling for
socioeconomic status. When English and Spanish predictors were included simultaneously, the
effect of Spanish story retell on initial status was negative and significant (z = -2.30; p = .02);
however, when the other nonsignificant effects were removed from the model, this effect became
nonsignificant (z = -1.89; p = .06). Thus, in the final model, Model 8, only the significant effect
of early English productive vocabulary on students’ initial (third-grade) status remained. Across
ELLS’ EARLY ORAL LANGUAGE AND LATER READING 21
models, none of the English or Spanish subtests predicted students’ rates of growth in English
reading.
When the significant effect of early English productive vocabulary on later English
reading levels in this final model was converted into a standardized path coefficient, it had a
magnitude of .29, which approached a moderate size. This indicated a substantially stronger
predictive relationship than that found for the English oral language composite (.20), for the
Spanish oral language composite (.14), and for the language-general composite (.17). As a point
of comparison, the effect of SES had a magnitude of .20.
Discussion
The current study was designed to investigate the extent to which early English and
Spanish oral language predicts later English reading development for the large, growing, and
underserved population of Spanish-speaking ELLs in the United States. Building on the
relatively robust research base on the moderate role of early oral language in reading
development for monolingual children (National Early Literacy Panel, 2008), this research
sought to investigate whether similar conclusions held for students learning to read in their
second language. Longitudinal data from a nationally representative sample followed for nine
years were used to begin to address the widely acknowledged limitations in the empirical
research base on ELLs’ reading development (e.g., August & Shanahan, 2006; Gutierrez et al.,
2010: Lonigan & Shanahan, 2010). The study yielded four major findings, each with theoretical
implications for understanding second-language reading development as well as practical
implications for designing assessment and instruction for young ELLs that can prevent later
reading difficulties. In the following four sections, each finding and its implications for research
and practice are discussed.
ELLS’ EARLY ORAL LANGUAGE AND LATER READING 22
The Importance of Early Oral Language Development
The first major finding was that the relationship between early English oral language and
later levels of English reading was statistically significant and practically meaningful in
magnitude. This finding converges with the handful of past studies that have followed Spanish-
speaking ELLs beyond fourth grade (Mancilla-Martinez & Lesaux, 2010; Nakamoto et al., 2007;
Reese et al., 2000). The effect sizes for this relationship (standard path coefficients = .20 to .29),
controlling for socioeconomic status, can be considered small to moderate in magnitude. The
larger estimate (.29), in particular, is roughly comparable to the average estimate for
monolinguals offered by the NELP (2008) (r = .33) and thus support the notion that the NELP
(2008) conclusion that oral language is a moderate predictor of later reading outcomes can be
extended to Spanish-speaking ELLs. As evidence of its practical importance, this effect size was
also found to be comparable to or larger than the well-known relationship between SES and
reading level. This finding provides valuable support for the importance of early language
development in placing ELLs on the pathway to success with reading comprehension through the
upper-elementary and middle school grades. At the same time, the similarity in magnitude
between these estimates for ELLs and native English speakers suggests that early English oral
language development may not be dramatically more important for ELLs than for their peers, as
is sometimes suggested.
One practical implication of this finding is that measuring ELLs’ oral language in early
childhood can provide valuable, if incomplete, information about students’ relative risk for
developing later reading difficulties. The moderate predictive power of oral language measures
suggests that such measures should supplement, rather than replace, measures of other early
literacy skills known to predict ELLs’ reading, such phonological awareness, letter knowledge,
ELLS’ EARLY ORAL LANGUAGE AND LATER READING 23
and working memory (for a review, see Lesaux, Koda, Siegel, & Shanahan, 2006). A second
implication is that preschool and kindergarten classrooms must enhance the opportunities for oral
language development offered to ELLs as part of regular classroom instruction, in line with the
consensus recommendations offered by the National Academy of Education (2009) among
others.
Vocabulary Tests May be More Predictive than Complex Measures for ELLs
In a more surprising finding, the English productive vocabulary measure used was a
better predictor of later levels of English reading than more complex English language measures,
contrary to the conclusion for monolinguals offered by the NELP (2008). When included
together with a listening comprehension and a story retell measure—both of which required
students to integrate grammar and vocabulary skills—the productive vocabulary measure was the
only predictor to demonstrate a significant unique contribution to later levels of English reading.
In addition, this single productive vocabulary measure was almost 50% more predictive than a
composite of the three measures, contrary to the NELP (2008) findings and to common
expectations that a composite will typically be a more reliable predictor than a single sub-test.
Three potential explanations for this divergence in findings are possible, each of which
raises questions for future research. First, it is possible that this finding is specific to the subtests
of the PreLAS used. Although the PreLAS was widely used in practice and had evidence of
solid psychometric properties when selected by a panel of experts during the planning of ECLS-
K in the late 1990’s, the content and validity of this measure has since been criticized by applied
linguistics (e.g., MacSwan & Rolstad, 2006; MacSwan, Rolstad, & Glass, 2002). In addition, the
PreLAS does not have the same level of evidence establishing the separability and validity of the
individual vocabulary, listening comprehension, and story recall subtests as do some of the more
ELLS’ EARLY ORAL LANGUAGE AND LATER READING 24
well-established measures of vocabulary, grammar, and listening comprehension, including
many of those used in the studies reviewed by the NELP (2008). Further, this measure was
originally designed for the broader purpose of informing decisions about whether ELLs have
sufficient English proficiency to benefit from mainstream English instruction, rather than to
distinguish specific aspects of oral proficiency that predict their reading development or risk for
reading difficulties. As a result, the listening comprehension and story retell tasks may have
tapped basic and general communicative competence to a greater extent than many of the clinical
and research language measures used in the NELP studies (e.g., the Clinical Evaluation of
Language Fundamentals; Wiig, Semel, & Secord, 1992), whereas many of the latter measures
tapped more explicit and metalinguistic aspects of grammatical knowledge. The PreLAS
composite also weights the story retell measure more heavily than the other subtests, which may
not have been warranted. Thus, this finding is subject to replication with different measures.
Future research is necessary to investigate not only which constructs best predict later reading
development for ELLs, but also which operational approaches to measuring those constructs
yield the most predictive power.
A second possible explanation lies in the developmental periods investigated and the
relative importance of different skills during those periods. Given that the vast majority of
studies reviewed by the NELP (2008) followed students into first or second grade, their
conclusion about the relative weak predictive power of early vocabulary may not be applicable to
predicting performance in much later grades. It is possible that the relative predictive power of
early vocabulary measures is better maintained as students grow older and leave the primary
grades, whereas the predictive power of grammar or complex measures declines more with age.
Such a changing role for these different aspects of oral language proficiency is consistent with
ELLS’ EARLY ORAL LANGUAGE AND LATER READING 25
the increasing knowledge demands of the texts students encounter as they move into the upper
elementary and middle school grades (e.g., Chall, 1996; Jetton & Alexander, 2004); to the extent
that vocabulary is a proxy for students’ range of conceptual knowledge about various domains,
its predictive power will likely grow over time as knowledge demands increase. This explanation
has some support from the few studies that have followed monolingual and Spanish-speaking
ELLs beyond the primary grades (e.g., Cunningham & Stanovich, 1997; Mancilla-Martinez &
Lesaux, 2010; Reese et al. 2000: Snow et al., 2007), which indeed suggest that early vocabulary
continues to have at least a moderate relationship with reading achievement many years later.
Although the studies conducted with ELLs did not directly compare early vocabulary with
listening comprehension or grammar in predicting post-primary outcomes, Snow et al.’s (2007)
study of monolingual children from low-income families did find that a kindergarten receptive
vocabulary measure retained the strong magnitude of its correlation with later reading
comprehension through Grade 7 and Grade 10 to a greater extent than did a narrative production
task or a definitional skill task. That said, the NELP (2008) did not find differences in the
existing literature between shorter term and longer term longitudinal studies (Lonigan &
Shanahan, 2010), suggesting that future research is needed to settle this question.
A third explanation for the divergence between the current findings and the conclusions
from the NELP (2008) is that there are meaningful differences in reading development between
Spanish-speaking ELLs and native English speakers. For Spanish-speaking ELLs, complex
measures may have less specificity in identifying individual differences that will constrain
reading development, if they provide greater contextual support and thus more opportunities to
score well on the task, obscuring important deficiencies in English proficiency. A
complementary explanation is that the errors in English grammar production that ELLs are likely
ELLS’ EARLY ORAL LANGUAGE AND LATER READING 26
to demonstrate on a complex proficiency task (such as the story retell task in the current study)
may indicate certain real limitations in English proficiency, but that such limitations have
relatively little importance to reading, which involves receptive and arguably less precise
knowledge of grammar. Indeed, there is substantial evidence from studies of second-language
acquisition in adults of persistent errors in morphosyntactic features such as inflections and
subject-verb agreement (Gass & Selinker, 2009), even among those who are academically
successful in their second language; their success, despite flawed English grammar, can be
interpreted as evidence that complete mastery of English grammar may be unnecessary for
reaching functional levels of reading competence. Such an explanation would also be consistent
with the findings from a few studies for the relatively limited predictive power of explicit
syntactic awareness tasks for explaining ELLs’ reading outcomes in the elementary grades (e.g.,
Lipka & Siegel, 2007; Jongejan, Verhoeven, & Siegel, 2007).
Pending replication and confirmation, this finding raises practical questions about the
relative emphasis in preschool and primary grade classrooms that should be placed upon
vocabulary building activities, compared with activities to promote grammatical development
and global listening comprehension. The NELP (2008) interpreted the relatively weaker
relationship between simple vocabulary measures and reading outcomes as support for an
instructional focus that goes beyond vocabulary alone. They conclude that instructional attention
to vocabulary may be important but unlikely to be sufficient for oral language development.
While the current findings do not contradict this cautious recommendation or its extension to
ELLs, they do imply that there may be a danger in under-emphasizing vocabulary as a precursor
literacy skill to be measured and targeted. Given constraints on instructional time, emphasizing
ELLS’ EARLY ORAL LANGUAGE AND LATER READING 27
the recommendation to “go beyond vocabulary” may inadvertently lead to little time devoted to
creating environments and activities that can promote vocabulary learning.
Early Spanish Predicts Later Reading, but Not Uniquely
The third major finding was that early Spanish oral language demonstrated a significant
zero-order correlation with later levels of English reading comprehension, but that this
relationship was no longer significant when controlling for early English oral language. When
combined with the finding that the early oral language composites for Spanish and English were
highly correlated with each other, this suggests that much of the variation in English reading
comprehension that is explained by Spanish skills is the same variation explained by English
skills. This finding supports a general notion of a common underlying proficiency that is not
language-specific (e.g., Cummins, 1979) but is demonstrated on both first- and second-language
measures of oral proficiency. This finding also provides some evidence to support the
conclusions of Verhoeven (1994), Bialystok (2002), and others in emphasizing the limits of first-
language vocabulary and grammar in supporting second-language reading development.
At the same time, this finding should be understood in the context of ELLs’ opportunities
to learn to read in their first and second languages. It may be that the predictive power of
Spanish oral language will be greater for those students who are receiving literacy instruction in
their primary language (e.g., Gottardo & Mueller, 2009), unlike the majority of ELLs in the U.S.
(e.g., Crawford, 2004). Indeed, the power of oral language skills in predicting later reading is
due not only to their longitudinal relationships with later oral language skills which are required
for later reading comprehension, but also due to the ways in which well-developed oral language
provides students with access to literacy instruction (i.e., allows them to understand what their
teacher is saying). For instance, in English-only classrooms, well-developed English oral
ELLS’ EARLY ORAL LANGUAGE AND LATER READING 28
proficiency may be necessary for first graders to understand their teachers’ delivery of
systematic phonics instruction in an efficient and deep way, or for third graders to understand
instruction in reading comprehension strategies. However, if such instruction is primarily in
Spanish or delivered with substantial Spanish support, Spanish oral language proficiency is much
more likely to be important to developing literacy. The relatively limited information on
language of instruction reported by ECLS-K makes it difficult to investigate these questions fully
with the current dataset, but future research investigating these questions is certainly warranted.
Early Language Development is Not Enough for Later Reading Growth
The fourth major finding was that early oral language, whether assessed in English or in
Spanish, did not predict later rates of growth in English reading between third and eighth grade.
Although students with well-developed early vocabularies had substantially higher reading levels
in third grade, compared to their counterparts with under-developed early vocabularies, their
trajectories were parallel to their counterparts through eighth grade. That is to say, early
vocabulary-related differences in reading levels were maintained through eighth grade, but did
not increase. It appears that developing early oral proficiency may provide an advantage in the
primary grades, but does not necessarily lead to a pattern of Matthews Effects in which the “rich
get richer” and those with weak oral skills fall increasingly further behind (Stanovich, 1986) over
the upper-elementary and middle school grades.
This finding contrasts with the findings of Nakamoto et al.’s (2007) study of Spanish-
speaking ELLs, which found that an English oral language composite in first grade had a
significant, negative association with initial linear rates of growth in English reading
comprehension between first and sixth grade as well as a significant, positive association with
curvature. Together, their results indicated that ELLs with well-developed early oral language
ELLS’ EARLY ORAL LANGUAGE AND LATER READING 29
had slower initial growth but less deceleration over time, compared to ELLs with lower levels of
early oral language. One explanation for this divergence of findings may be the different
developmental periods studied and specifications of growth available. Because the current study
did not examine reading growth before the spring of third grade, the negative effect of oral
language on early growth may not have been observed. In addition, the inclusion of only three
testing occasions for reading in the current study made it impossible to model curvature and thus
detect effects on this growth parameter. In addtition, this divergence may be due to differences
in the samples; the participants in Nakamoto et al were drawn from transitional bilingual
programs within a single school district in a Texas border town, as opposed to the nationally
representative sample in the current study. Finally, it is possible that this divergence is due to
differences in the oral language or reading measures used. With so few studies to date that have
investigated effects of oral language on reading growth in ELLs, there is a clear need for more
research to explore this question across a range of developmental periods using various
measures.
The findings from the current study support the developmental notion that early oral
language development may be necessary for learning to read, but that it is not sufficient to
support rapid later growth in reading achievement in the upper-elementary and middle school
grades (e.g., Chall, 1996; RAND Reading Study Group, 2002; Snow, Griffin, & Burns, 2005).
More rapid growth in these grades than was observed in this study would be necessary for
Spanish-speaking ELLs to close achievement gaps with their White, monolingual counterparts
(e.g., Kieffer, 2008, 2010; NCES, 2009). As growing research on adolescent literacy indicates
(for a review, see Carnegie Council on Adolescent Literacy, 2010), there is a wide range of
skills, knowledge, and practices—both within and outside of the domain of oral language—
ELLS’ EARLY ORAL LANGUAGE AND LATER READING 30
involved in successful reading comprehension of upper-elementary and middle school texts.
This finding upholds the general recommendation that concerted efforts to develop oral language
should not be the responsibility of early childhood educators alone, but must be sustained and
differentiated as students progress through the grades.
Limitations and Future Research
By drawing on existing data from a large-scale, long-term longitudinal research project,
this study offers many advantages over prior research, but also has several limitations. First, as
noted above, the findings from the current study may be specific to the individual language and
reading measures selected by NCES. Future studies would be well-served by including multiple
measures of each oral language skill as well as multiple measures of reading comprehension to
facilitate the investigation of relationships among these constructs, while accounting for
measurement error, using techniques such as structural equation modeling.
Second, the questions that could be investigated by this study were limited by the
collection by NCES of only quantitative assessment and survey data. The description of early
oral language would have benefitted from the collection of qualitative observational data on
students’ actual language use and literacy practices in various contexts, including both home and
school. Future large-scale longitudinal studies would benefit from more mixed-methods designs
that allow for the triangulation of findings among different data sources and types.
Third, NCES chose to limit the collection of data on English oral proficiency to language
minority learners, making it impossible to analyze directly whether the effects of early English
oral language skills differ between these learners and native English speakers. In addition,
NCES only assessed first-language oral proficiency in Spanish speakers who entered school with
initially limited English proficiency. As a result, it was not possible to determine whether
ELLS’ EARLY ORAL LANGUAGE AND LATER READING 31
different effects for first-language oral proficiency would be found for students who were
initially bilingual or for ELLs from other language backgrounds. On a more technical note, it is
also possible that this sampling decision led to a restriction of range in the English oral
proficiency score that attenuated correlations to some extent. Future comparative research on the
full range of language minority learners is needed to shed greater light on these questions.
Finally, the focus on early childhood in the ECLS-K study led NCES to collect only three
waves of reading assessment data between Grades 3 and 8, at wider intervals than during
kindergarten and Grade 1. As noted above, this made it necessary to specify a linear growth
trajectory, when it is likely that growth during this period would be better described as
curvilinear (e.g., Francis, Shaywitz, Stuebing, Shaywitz, & Fletcher, 1996; Nakamoto et al.,
2007). Although the focus of ECLS-K on early childhood is clearly warranted given its mission
and purposes, future large-scale studies that provide a more detailed look at language and
reading growth through more frequent data collection during the upper-elementary and middle
school years are also needed. The lack of effects of early oral language on students’ rates of
reading growth between third through eighth grade begs the question of which cognitive,
linguistic, social, and instructional factors do predict differences in growth during this period,
raising the need for collecting richer data during the middle grades in future longitudinal studies.
Conclusion
Although future research is clearly needed, the current study provides a step toward
building an empirical research base about how ELLs’ early oral language development supports
their later reading development. Findings from this study support the importance of early
English oral language development—particularly vocabulary development—to later English
reading performance. At the same time, findings suggest the limitations of focusing on oral
ELLS’ EARLY ORAL LANGUAGE AND LATER READING 32
language alone and on early development alone. Such an emphasis is likely to be a necessary
condition, but it is unlikely to be a sufficient condition to promoting more equal long-term
reading outcomes for the increasingly linguistically diverse population of U.S. students.
ELLS’ EARLY ORAL LANGUAGE AND LATER READING 33
References
August, D., Carlo, M., Dressler, C, & Snow, C. (2005). The critical role of vocabulary
development for English language learners. Learning disabilities: Research and practice,
20(1), 50-57.
August, D. & Shanahan, T. (Eds.), Developing literacy in second-language learners: Report of
the National Literacy Panel on Language-Minority Children and Youth. Mahwah, New
Jersey: Lawrence Erlbaum Associates.
Beck, I. L., McKeown, M.G., & Kucan, L. (2002). Bringing words to life. New York: The
Guilford Press.
Bialystok, E. (2002). Acquisition of literacy in bilingual children: A framework for research.
Language Learning, 52, 159-199.
Bollen, K.A., & Curran, P.J. (2006). Latent growth curve models: A structural equation
approach. Hoboken, NJ: Wiley.
Capps, R., Fix, M., Murray, J., Ost, J. Passel, J. S., & Herwantoro, S. (2005). The new
demography of America’s schools: Immigration and the No Child Left Behind Act
(Research report). Washington, DC: Urban Institute.
Carnegie Council on Adolescent Literacy (2010). Time to act: An agenda for advancing
adolescent literacy for college and career success. New York, NY: Carnegie
Corporation of New York.
Catts, H.W., Adlof, S.M., & Weismer, S.E. (2006). Language deficits in poor comprehenders: A
case for the simple view of reading. Journal of Speech, Language, & Hearing Research,
49, 278-293.
Chall, J. (1983). Stages of reading development. New York: McGraw Hill.
ELLS’ EARLY ORAL LANGUAGE AND LATER READING 34
Cosentino de Cohen, C., Deterding, N., & Chu Clewell, B. (2005). Who’s left behind? Immigrant
children in high and low LEP schools (Policy Report). Washington, DC: Urban Institute.
Crawford, J. (2004). Educating English learners: Language diversity in the classroom (5th ed.).
Los Angeles: Bilingual Education Services.
Cummins, J. (1979). Cognitive/academic language proficiency, linguistic interdependence, the
optimal age question and some other matters. Working Papers on Bilingualism, 19, 121-
129.
Cunningham, A. E. & Stanovich, K. E. (1997). Early reading acquisition and its relation to reading
experience and ability 10 years later. Developmental Psychology, 33, 934- 45.
Dickinson, D. K., & Tabors, P. O. (2001). Beginning literacy with language. Baltimore, MD:
Brookes.
Duncan, S.E. & DeAvila, E.A. (1998). Pre- language assessment scales [PreLAS]. Monterey,
CA: CTB/McGraw-Hill.
Francis, D.J., Lesaux,N.K., & August, D. (2006). Language of instruction. In D. August & T.
Shanahan (Eds.), Developing literacy in second-language learners: Report of the
National Literacy Panel on Language-Minority Children and Youth. Mahwah, New
Jersey: Lawrence Erlbaum Associates.
Francis, D. J., Shaywitz, S. E., Stuebing, K. K., Shaywitz, B. A., & Fletcher, J. M. (1996).
Developmental lag verse deficit models of reading disability: A longitudinal, growth
curves analysis. Journal of Educational Psychology, 88, 3-17.
ndara, P., Rumberger, R., Maxwell-Jolly, J., & Callahan, R. (2003). English learners in
California schools: Unequal resources, unequal outcomes. Education Policy Analysis
Archives, 11(36). Retrieved February 26, 2007, from http://epaa.asu.edu/epaa/v11n36/
ELLS’ EARLY ORAL LANGUAGE AND LATER READING 35
Gass, S.M. & Selinker, L. (2008). Second language acquisition: An introductory course (3rd
edition). New York: Lawrence Erlbaum Associates.
Genesee, F., Lindholm-Leary, K., Saunders, W.M., & Christian, D. (2006) Education English
language learners: A synthesis of research evidence. Cambridge, UK: Cambridge
University Press.
Geva, E. & Genessee, F. (2006). First-language oral proficiency and second-language literacy. In
D. August & T. Shanahan (Eds.), Developing literacy in second-language learners:
Report of the National Literacy Panel on Language-Minority Children and Youth.
Mahwah, New Jersey: Lawrence Erlbaum Associates.
Gottardo, A., & Mueller, J. (2009). Are first- and second-language factors related in predicting
second-language reading comprehension? A study of Spanish-speaking children
acquiring English as a second language from first to second grade. Journal of
Educational Psychology, 101, 330-344.
Gutierrez, K.D., Zepeda, M., & Castro, D.C. (2010). Advancing early literacy learning for all
children: Implications of the NELP report for dual-language learners. Educational
Researcher, 39, 334-339.
Hutchinson, J.M., Whiteley, H.E., Smith, C.D., & Connors, L. (2003). The developmental
progression of comprehension-related skills in children learning EAL. Journal of
Research in Reading, 26, 19-32.
Jetton, T.L. & Alexander, P.E. (2004). Domains, teaching, and literacy. In T. Jetton (Ed).,
Adolescent literacy research and practice (pp. 15-39). New York: Guilford Press.
Jongejan, W., Verhoeven, L., & Siegel, L.S. (2007). Predictors of reading and spelling abilities
in first- and second-language learners. Journal of Educational Psychology, 99, 835-851.
ELLS’ EARLY ORAL LANGUAGE AND LATER READING 36
Kieffer, M.J. (2008). Catching up or falling behind? Initial English proficiency, concentrated
poverty, and the reading growth of language minority learners in the United States.
Journal of Educational Psychology, 100, 851-868.
Kieffer, M.J. (2010). English proficiency, socioeconomic status, and late-emerging reading
difficulties. Educational Researcher, 39, 484-486.
Lesaux, N.K., Crosson, A.C., Kieffer, M.J., & Pierce, M. (2010). Uneven profiles: Language
minority learners’ word reading, vocabulary, and reading comprehension skills. Journal
of Applied Developmental Psychology, 31, 475-483.
Lesaux, N.K. & Kieffer, M.J. (2010). Exploring sources of reading comprehension difficulties
among language minority learners and their classmates in early adolescence. American
Educational Research Journal, 47, 596-632.
Lesaux, N. K., with Koda, K., Siegel, L. S., & Shanahan, T. (2006). Development of
literacy. In D. August & T. Shanahan (Eds.), Developing literacy in second-language
learners: Report of the National Literacy Panel on Language-Minority Children and
Youth (pp. 75-122). Mahwah, NJ: Lawrence Erlbaum Associates.
Lesaux, N. K., Rupp, A., & Siegel, L. S. (2007). Growth in reading skills of children from
diverse linguistic backgrounds: Findings from a 5-year longitudinal study. Journal of
Educational Psychology, 99, 821-834.
Lonigan, C.J. & Shanahan, T. (2010). Developing early literacy skills: Things we know we know
and things we know we don’t know. Educational Researcher, 39, 340-346.
Limbos, M., & Geva, E. (2001). Accuracy of teacher assessments of second-language students at risk for
reading disability. Journal of Learning Disabilities, 34(2), 136-151.
Lipka, O. & Siegel, L.S. (2007). The development of reading skills in children with English as a
second language. Scientific Studies of Reading, 11, 105-131.
ELLS’ EARLY ORAL LANGUAGE AND LATER READING 37
MacSwan, J. & Rolstad, K. (2006). How language proficiency tests mislead us about ability:
Implications for English language learner placement in special education. Teachers
College Record, 108, 2304-2328.
MacSwan, J., Rolstad, K., & Glass, G.V. (2002). Do some school-age children have no
language? Some problems of construct validity in the Pre-LAS Español. Bilingual
Research Journal, 26, 395-421.
Mancilla-Martinez, J., & Lesaux, N.K. (2010) Predictors of reading comprehension for
struggling readers: The case of Spanish-speaking language minority learners. Journal of
Educational Psychology, 102, 701-711.
Manis, F., Lindsey, K., & Bailey, A. (2004). Development of reading in grades K-2 in Spanish-
speaking English-language learners. Learning Disabilities Research & Practice, 19(4),
214-224.
Najarian, M. Pollack, J.M., and Sorongon, A.G. (2009). Early Childhood Longitudinal Study,
Kindergarten Class of 1998–99 (ECLS-K), Psychometric report for the eighth grade
(NCES 2009–002). National Center for Education Statistics, Institute of Education
Sciences, U.S. Department of Education. Washington, DC.
Nakamoto, J. Lindsey, K.A., & Manis, F. R. (2007). A longitudinal analysis of English language
learners’ word decoding and reading comprehension. Reading & Writing: An
Interdisciplinary Journal, 20, 691-719.
Nakamoto, J., Lindsey, K. A., & Manis, F. R. (2008). A cross-linguistic investigation of English
language learners’ reading comprehension in English and Spanish. Scientific Studies of
Reading, 12, 351-371.
ELLS’ EARLY ORAL LANGUAGE AND LATER READING 38
National Academy of Education (2009). Education policy briefing sheet: Literacy for learning.
National Academy of Education White Papers Project. Washington, DC: Authors.
National Center for Educational Statistics. (2009). Nation's Report Card: Reading, 2009.
Washington, DC: Author.
National Early Literacy Panel (2008). Developing early literacy: Report of the National Early
Literacy Panel. Washington, DC: National Institute of Literacy.
Orellana, M.F. & D’warte, J. (2010). Recognizing different kinds of “head starts.” Educational
Researcher, 39, 295-300.
RAND Reading Study Group (2002). Reading for Understanding: Toward an R&D Program in
Reading Comprehension. Santa Monica, CA: RAND Corporation.
Reese, L., Garnier, H., Gallimore, R., & Goldenberg, C. (2000). Longitudinal analysis of the
antecedents of emergent Spanish literacy and middle-school English reading achievement
of Spanish-speaking students. American Educational Research Journal, 37, 633-662.
Samson, J.F., & Lesaux, N.K. (2009). Language-minority learners in special education: Rates
and predictors of identification for services. Journal of Learning Disabilities, 42, 148-
162.
Schnatschneider, C. & Lonigan, C. J. (2010). Misunderstood statistical assumptions undermine
criticism of the National Early Literacy Panel’s report. Educational Researcher, 39, 347-
351.
Senechal, M. & LaFevre, J. (2002). Parental involvement in the development of children’s
reading skill: A five-year longitudinal study. Child Development, 73, 445-460.
Slavin, R.E., & Cheung, A. (2005). A synthesis of research on language of reading instruction
for English language learners. Review of Educational Research, 75, 247-284.
ELLS’ EARLY ORAL LANGUAGE AND LATER READING 39
Snow, C. E., Burns, M. S., & Griffin, P. (Eds.). (1998). Preventing reading difficulties in young
children. Washington, DC: National Academy Press.
Snow, C. E., Griffin, P., & Burns, M S. (Eds.). (2005). Knowledge to support the teaching of
reading. San Francisco: Jossey-Bass.
Snow, C.E., Porche, M.V., Tabors, P.O. & Harris, S.R. (2007). Is literacy enough? Pathways to
academic success for adolescents. Baltimore, MD: Paul H. Brookes, Publishing Co.
Stanovich, K. E. (1986). Matthew effects in reading: Some consequences of individual
differences in the acquisition of literacy. Reading Research Quarterly, 21, 360-407.
Storch, S.A. & Whitehurst, G.J. (2002). Oral language and code-related precursors to reading:
Evidence from a longitudinal structural model. Developmental Psychology, 38, 934-947.
Swanson, H.L., Rosston, K., & Gerber, M., & Solari, E. (2008). Influence of oral language and
phonological awareness on children’s bilingual reading. Journal of School Psychology,
46, 413-429.
Tourangeau, K., Lê, T., Nord, C., and Sorongon, A. G. (2009). Early Childhood Longitudinal
Study, Kindergarten Class of 1998–99 (ECLS-K), Eighth-grade methodology report
(NCES 2009–003). National Center for Education Statistics, Institute of Education
Sciences, U.S. Department of Education. Washington, DC.
U.S. Department of Education, National Center for Education Statistics. Early Childhood
Longitudinal Study-Kindergarten Class of 1998–99 (ECLS–K), Psychometric Report for
Kindergarten Through First Grade, NCES 2002–05, by Donald A. Rock and Judith M.
Pollack, Educational Testing Service, Elvira Germino Hausken, project officer.
Washington, DC: 2002.
ELLS’ EARLY ORAL LANGUAGE AND LATER READING 40
Verhoeven, L.T. (1994). Transfer in bilingual development: The linguistic interdependence
hypothesis revisited. Language Learning, 44, 381-415.
Wiig, E. H., Semel, E. M., & Secord, W. (1992). CELF-Preschool: Clinical evaluation of
language fundamentals—Preschool. San Antonio, Tex.: Psychological Corporation,
Harcourt Brace Jovanovich.
ELLS’ EARLY ORAL LANGUAGE AND LATER READING 41
Table 1
Longitudinal Correlations among Oral Language Skills in English and Spanish Measured in Fall of Kindergarten and English
Reading Comprehension Measured in Grade 3, 5, and 8, Estimated using Longitudinal Sampling Weights (N = 295)
1.
2.
3.
5.
6.
7.
9.
10.
1. K English Oral Language Composite
--
2. K English Listening Comprehension
.87
--
3. K English Productive Vocabulary
.81
.71
--
4. K English Story Retell
.94
.67
.62
5. K Spanish Oral Language
.72
.53
.45
--
6. K Spanish Listening Comprehension
-.12
.01
-.14
.15
--
7. K Spanish Productive Vocabulary
-.29
-.24
-.20
.13
.23
--
8. K Spanish Story Retell
.83
.60
.53
.91
-.09
-.27
9. Grade 3 English Reading
.27
.31
.33
.10
.02
.03
--
10. Grade 5 English Reading
.24
.27
.31
.13
.08
.10
.79
--
11. Grade 8 English Reading
.12
.19
.23
-0.02
.02
.03
.60
.74
Means
12.66
3.74
2.98
22.14
9.77
7.81
0.50
0.82
SD
11.95
3.68
2.75
5.72
0.92
2.12
0.30
0.24
Note. K = Kindergarten
ELLS’ EARLY ORAL LANGUAGE AND LATER READING 42
Table 2
Selected Results for Latent Growth Models for the Relationship between Initial Status and Rate of Growth in English Reading between
Third and Eighth Grade, Socioeconomic Status, Kindergarten English Oral Language Composite, and Kindergarten Spanish Oral
Language Composite (N = 295)
Symbol
Model 1
Model 2
Model 3
Model 4
Fixed Effects
Initial (Third Grade)
Status
Intercept
γ00
0.52***
0.46***
0.85*
0.53***
SES Factor
γ01
0.12*
0.14*
0.11*
0.12*
Fall, K English Oral Language Composite
γ02
0.01**
0.01**
0.01*
Fall, K Spanish Oral Language Composite
γ03
0.01*
0.00
Rate of Growth
Intercept
γ10
0.11***
0.12***
0.11***
0.10***
SES Factor
γ11
0.01
0.00
0.01
0.00
Fall, K English Oral Language Composite
γ12
0.00
0.00
Fall, K Spanish Oral Language Composite
γ13
0.00
0.00
Variance Components
Structural Model
Variance in Initial Status
σ0
2
0.05***
0.05***
0.05***
0.05***
Correlation between SES Factor and Fall, K
English Oral Language Composite
σ12
.26***
.26***
.26***
.26***
Correlation between SES Factor and Fall, K
Spanish Oral Language Composite
σ13
.12
.12
.12
.12
Correlation between Fall, K English Oral Language
and Fall, K Spanish Oral Language Composite
σ23
.72***
.72***
.72***
.71***
Measurement Model
Grade 3 Reading
σε1
2
0.04***
0.04***
0.04***
0.04***
Grade 5 Reading
σε2
2
0.01*
0.01
0.01*
0.01*
Grade 8 Reading
σε3
2
0.05***
0.05***
0.05***
0.05***
Goodness of Fit
-2LogLikelihood
4365.60
4374.22
4363.85
4368.27
AIC
4429.61
4438.22
4431.86
4430.27
BIC
4547.59
4556.20
4557.22
4544.57
χ2
103.61
105.72
100.36
106.22
CFI
.91
.91
.91
.91
RMSEA
.09
.09
.09
.09
R2 in Initial Status
.13
.09
.14
.11
ΔR2 for Initial Status1
.06
.02
.06
.04
ELLS’ EARLY ORAL LANGUAGE AND LATER READING 43
Note. Model 1 displays the effect of English oral language alone, model 2 displays the effect of Spanish oral language alone, model 3
displays the simultaneous effects of English and Spanish oral language, and model 4 displays only the uniquely significant effects of
English oral language. All Models also included the measurement model for the SES Factor. In all models, variance in rate of growth
and covariance between initial status and rate of growth are fixed to zero. Values of 0.00 are due to rounding, rather than fixing effects
to zero.
1These ΔR2 statistics are in comparison to a model including the effects of SES.
*p<.05; **p < .10; ***p < .001
ELLS’ EARLY ORAL LANGUAGE AND LATER READING 44
Table 3
Selected Results for Latent Growth Models for the Relationship between Initial Status and Rate of Growth in English Reading between
Third and Eighth Grade, Socioeconomic Status, Kindergarten English Oral Language Subtests, and Kindergarten Spanish Oral
Language Subtests (N = 295)
Symbol
Model 5
Model 6
Model 7
Model 8
Fixed Effects
Initial (Third Grade) Status
Intercept
γ00
0.49***
0.36
0.18
0.51***
SES Composite
γ01
0.10
0.15*
0.09
0.11*
Fall, K English Listening Comprehension
γ04
0.01
0.01
Fall, K English Productive Vocabulary
γ05
0.02*
0.02*
0.02**
Fall, K English Story Retell
γ06
0.00
0.01
Fall, K Spanish Listening Comprehension
γ07
0.01
0.02
Fall, K Spanish Productive Vocabulary
γ08
0.01
0.02
Fall, K Spanish Story Retell
γ09
0.01
-0.01*
Rate of Growth
Intercept
γ10
0.11***
0.12*
0.13*
0.10***
SES Composite
γ11
0.01
0.00
0.01
0.00
Fall, K English Listening Comprehension
γ14
0.00
0.00
Fall, K English Productive Vocabulary
γ15
0.00
0.00
Fall, K English Story Retell
γ16
0.00
0.00
Fall, K Spanish Listening Comprehension
γ17
0.00
0.00
Fall, K Spanish Productive Vocabulary
γ18
0.00
0.00
Fall, K Spanish Story Retell
γ19
0.00
0.00
Variance Components
Structural Model
Variance in Initial Status
σ0
2
0.05***
0.05***
0.04***
.05***
Measurement Model
Grade 3 Reading
σε1
2
0.04***
0.04***
0.04***
.04***
Grade 5 Reading
σε2
2
0.01*
0.01
0.01*
.01*
Grade 8 Reading
σε3
2
0.05***
0.05***
0.05***
.05***
Goodness of Fit
-2LogLikelihood
8415.79
8438.50
8404.00
8450.82
AIC
8539.79
8562.51
8540.00
8554.82
BIC
8768.38
8791.10
8790.72
8746.54
χ2
142.28
151.27
131.14
163.40
CFI
.92
.93
.91
.92
RMSEA
.07
.07
.07
.07
R2 in Initial Status
.18
.10
.23
.13
ΔR2 for Initial Status1
.10
.02
.15
.06
Note. Model 5 displays the effect of English oral language subtests alone, model 2 displays the effect of Spanish oral language
subtests alone, model 3 displays the simultaneous effects of English and Spanish oral language subtests, and model 4 displays only the
ELLS’ EARLY ORAL LANGUAGE AND LATER READING 45
uniquely significant effects of English picture vocabulary. All Models also included the measurement model for the SES factor as well
as structural covariances among the SES Factor, English and Spanish oral language subtests. In all models, variance in rate of growth
and covariance between initial status and rate of growth are fixed to 0. Values of 0.00 are due to rounding, rather than fixing effects to
zero.
1These ΔR2 statistics are in comparison to a model including the effects of SES.
*p<.05; **p < .10; ***p < .001
ELLS’ EARLY ORAL LANGUAGE AND LATER READING 46
Figure 1. Path diagram for hypothesized latent growth model for English reading predicted by early English and Spanish oral
language, controlling for socioeconomic status.
... Regarding the individual contribution of different language comprehension predictors to reading comprehension, most studies on L2 3rd to 7th graders have found L2 vocabulary to uniquely predict L2 reading comprehension over and above other linguistic skills (Burgoyne et al., 2011;Hutchinson et al., 2003;Kieffer, 2012;Proctor et al., 2012). There is also support for listening comprehension uniquely predicting L2 reading comprehension over other language comprehension constructs in bilingual children in this age group (Burgoyne et al., 2011;Hutchinson et al., 2003;Kieffer, 2012;Proctor et al., 2012). ...
... Regarding the individual contribution of different language comprehension predictors to reading comprehension, most studies on L2 3rd to 7th graders have found L2 vocabulary to uniquely predict L2 reading comprehension over and above other linguistic skills (Burgoyne et al., 2011;Hutchinson et al., 2003;Kieffer, 2012;Proctor et al., 2012). There is also support for listening comprehension uniquely predicting L2 reading comprehension over other language comprehension constructs in bilingual children in this age group (Burgoyne et al., 2011;Hutchinson et al., 2003;Kieffer, 2012;Proctor et al., 2012). However, the contribution from listening comprehension sometimes overlaps with vocabulary skills or is not present at all when controlling for other L2 linguistic skills (Burgoyne et al., 2011;Hutchinson et al., 2003;Kieffer, 2012). ...
... There is also support for listening comprehension uniquely predicting L2 reading comprehension over other language comprehension constructs in bilingual children in this age group (Burgoyne et al., 2011;Hutchinson et al., 2003;Kieffer, 2012;Proctor et al., 2012). However, the contribution from listening comprehension sometimes overlaps with vocabulary skills or is not present at all when controlling for other L2 linguistic skills (Burgoyne et al., 2011;Hutchinson et al., 2003;Kieffer, 2012). ...
Article
Full-text available
Bilingualism appears to be related to poor language and reading comprehension in the instructional language (L2). However, in the long term, an early age of second language acquisition (AoA) may produce higher levels of second language skills than a later onset of L2 exposure. Most studies on school-aged second language learners (L2 learners) include children with 3–7 years of second language (L2) exposure. We compared 91 early bilingual (Norwegian and a variety of different native languages) 5th graders with 196 monolingual peers on a range of linguistic skills and their relationships with reading comprehension. All bilingual learners in the sample had been exposed to the instructional language by at least the age of 2. The results, using structural equation modelling and latent variables, show that early bilingual learners’ vocabulary, listening comprehension, and reading comprehension are significantly lower than those of monolingual learners. Nevertheless, they have similar levels of knowledge of conjunctions and decoding skills and corresponding predictive patterns from linguistic skills to reading comprehension. We discuss the implications of the findings.
... Further than that, the prescribed CAPS documents require us as teachers to conduct teaching and learning processes in an English medium for most subjects underpinning the curriculum at the secondary schooling level. To curb the predicament, they found themselves in, teachers designed their strategies of translating some English words, terms, or phrases, into isiXhosa (Kieffer, 2012). These teachers reported on spending the most time on translations, code-switching, and assisting learners to ...
... Learners were observed to be experiencing difficulties communicating with their teachers in instances when code-switching was prohibited (Tsagari & Diakou, 2015;Neokleous, 2017). Makena (2020) and Kieffer (2012) had similar findings through observing that teaching an unfamiliar language needs an English language teacher to engage in a wide range of strategies that would enhance an additional language learned. As such, English language teachers argued that as they normally resort to translating in isiXhosa, this being learners' home language, they noticed conducive learning environments as learners could easily engage and debate issues without displaying attitudes relating to shyness and nervousness. ...
... Further than that, the prescribed CAPS documents require us as teachers to conduct teaching and learning processes in an English medium for most subjects underpinning the curriculum at the secondary schooling level. To curb the predicament, they found themselves in, teachers designed their strategies of translating some English words, terms, or phrases, into isiXhosa (Kieffer, 2012). These teachers reported on spending the most time on translations, code-switching, and assisting learners to ...
... Learners were observed to be experiencing difficulties communicating with their teachers in instances when code-switching was prohibited (Tsagari & Diakou, 2015;Neokleous, 2017). Makena (2020) and Kieffer (2012) had similar findings through observing that teaching an unfamiliar language needs an English language teacher to engage in a wide range of strategies that would enhance an additional language learned. As such, English language teachers argued that as they normally resort to translating in isiXhosa, this being learners' home language, they noticed conducive learning environments as learners could easily engage and debate issues without displaying attitudes relating to shyness and nervousness. ...
Conference Paper
Full-text available
This paper explores how code-switching can be meaningfully used as an empowerment approach towards improving learners' performance in the English language. In cultures with people using more than one language for communication, code-switching exists. Bilinguals as speakers of many languages, code-switch, using their languages resourcefully at conveying meaning in a variety of ways. Code-switching occurs every day during teaching and learning as most subjects in the curriculum are offered in the English language. Teaching and learning the English language in South Africa is characterized by serious challenges because the government is advocating for use of home languages for all subjects of lower grades in primary schools. However, teachers still PUPIL: International Journal of Teaching, Education and Learning ISSN 2457-0648 16 encounter challenges when using English as a medium of instruction in preceding grades because learners fail to comprehend challenging concepts and terminologies presented to them in a language besides their home language. This qualitative study revealed that using code-switching can be a worthwhile approach for use in bilingual classrooms. A possible recommendation is that English language teachers should utilize code-switching as an approach to assist language development as learners in schools investigated emanate from diverse cultures, underpinned by different linguistic backgrounds and linguistic constituencies.
... Further than that, the prescribed CAPS documents require us as teachers to conduct teaching and learning processes in an English medium for most subjects underpinning the curriculum at the secondary schooling level. To curb the predicament, they found themselves in, teachers designed their strategies of translating some English words, terms, or phrases, into isiXhosa (Kieffer, 2012). These teachers reported on spending the most time on translations, code-switching, and assisting learners to ...
... Learners were observed to be experiencing difficulties communicating with their teachers in instances when code-switching was prohibited (Tsagari & Diakou, 2015;Neokleous, 2017). Makena (2020) and Kieffer (2012) had similar findings through observing that teaching an unfamiliar language needs an English language teacher to engage in a wide range of strategies that would enhance an additional language learned. As such, English language teachers argued that as they normally resort to translating in isiXhosa, this being learners' home language, they noticed conducive learning environments as learners could easily engage and debate issues without displaying attitudes relating to shyness and nervousness. ...
Article
Full-text available
This paper explores how code-switching can be meaningfully used as an empowerment approach towards improving learners’ performance in the English language. In cultures with people using more than one language for communication, code-switching exists. Bilinguals as speakers of many languages, code-switch, using their languages resourcefully at conveying meaning in a variety of ways. Code-switching occurs every day during teaching and learning as most subjects in the curriculum are offered in the English language. Teaching and learning the English language in South Africa is characterized by serious challenges because the government is advocating for use of home languages for all subjects of lower grades in primary schools. However, teachers still encounter challenges when using English as a medium of instruction in preceding grades because learners fail to comprehend challenging concepts and terminologies presented to them in a language besides their home language. This qualitative study revealed that using code-switching can be a worthwhile approach for use in bilingual classrooms. A possible recommendation is that English language teachers should utilize code-switching as an approach to assist language development as learners in schools investigated emanate from diverse cultures, underpinned by different linguistic backgrounds and linguistic constituencies.
... 2006). Given language skills highly predict later achievement outcomes in ELs (Hoff, 2013;Kieffer, 2012), more research that seeks to understand this population to better meet their linguistic needs in educational programs are critically needed to address the well-documented achievement gap between Spanish-speaking ELs and their monolingual English-speaking peers. ...
Article
Purpose: The purpose of this study was to identify and describe latent dual language profiles in a large sample of school-age Spanish-English bilingual children designated as English learners (ELs) by their school district. Method: Data for this study include 847 Spanish-speaking ELs from kindergarten to third grade. Spanish and English narrative retell language samples were collected from all participants. Four oral language measures were calculated in Spanish and English, including the subordination index, moving average type-token ratio, narrative structure scheme (NSS), and words per minute using Systematic Analysis of Language Transcript. These indicator measures were used in a latent profile analysis to identify dual language profiles. Results: The optimal model represents a four-profile solution, including a Spanish-dominant group (average Spanish, low English), an English-dominant group (low Spanish, average English), and two balanced groups (a balanced-average group and a balanced-high group). Additionally, participants displayed uneven performance across language domains and distinct patterns of unique strength or weakness in a specific domain in one of their two languages. Conclusions: Findings from this study highlight the large variability in English and Spanish oral language abilities in school-age Spanish-speaking ELs and suggest that a dichotomous classification of ELs versus English-proficient students may not be sufficient to determine the type of educational program that best fits a specific bilingual child's need. These findings highlight the need to assess both languages across multiple language domains to paint a representative picture of a bilingual child's language abilities. The dual language profiles identified may be used to guide the educational program selection process to improve the congruence among the linguistic needs of an individual child, teachers' use of instructional language, and the goals of the educational program (i.e., improving English proficiency vs. supporting dual language development). Supplemental material: https://doi.org/10.23641/asha.20151836.
... However, once these language demands are met, dual-language learners' performance may match or even exceed academic performance of single-language students (NYCDE, 2009;Wolf et al., 2012). Therefore, a critical avenue for supporting school readiness among dual-language learners is assessing and supporting their language knowledge (Kieffer, 2012). ...
Chapter
Researchers have historically focused on characterizing vocabulary development in infants and toddlers. However, less is known about the vocabulary composition of children entering formal schooling. The authors propose that a critical next step in understanding school readiness is to characterize the academic vocabulary of children entering kindergarten. These assessments should identify knowledge of words used in general academic discourse and specific domains (e.g., science). This chapter outlines initial steps taken by the authors to identify children's science vocabulary around the age of school entry. Furthermore, the complexity of vocabulary assessment is illustrated via a discussion of vocabulary development in dual-language learners. Understanding the words that children can produce at various stages of development will help determine whether children are ready for school and inform interventions that target word knowledge. Indeed, focusing on children's vocabulary presents an exciting new opportunity to integrate developmental science with real-world educational settings.
... The emergence of potential problems in math in EL children has not been studied as extensively as in the area of reading (e.g., Baker et al., 2012;Kieffer, 2012;Lesaux et al., 2007;Rojas et al., 2019;. Studies in reading suggest that lateemerging potential problems for EL children in reading occur around grade 3 Swanson, 2017). ...
Article
Solving word problems in mathematics presents difficulties for many English learners (ELs), including those who use Spanish as a home language. In the early stage of learning to solve mathematics word problems, some children, including ELs, experience difficulties. By English status, EL refers to those children whose home language was Spanish while in the process of developing English at school. Math difficulties (MD) refers to those children with low normative mathematical problem-solving scores in both English and Spanish. The purpose of this study was to determine those measures that increase the odds of identifying EL children with emergent MD. Elementary school children (grades 1, 2, and 3) were administered a battery of math, vocabulary, reading, and cognitive measures (short-term memory [STM], inhibition, working memory [WM]) in both Spanish (L1) and English (L2) in Year 1 and again one year later. Multilevel growth modeling compared MD children identified one year later who manifested MD to children who were average math achievers or poor math achievers across the two testing waves (year 1 and year 2). The results indicated that significant growth parameters (i.e., measures of estimation, WM) increased the odds of identifying emergent MD relative to children with Persistent math deficits and average achievers. The results were discussed in terms of a multidimensional model that taps domain-specific skills and general cognitive processes that increase the odds of identifying later math difficulties.
... 2006). Given language skills highly predict later achievement outcomes in ELs (Hoff, 2013;Kieffer, 2012), more research that seeks to understand this population to better meet their linguistic needs in educational programs are critically needed to address the well-documented achievement gap between Spanish-speaking ELs and their monolingual English-speaking peers. ...
Article
Purpose: The purpose of this study was to identify and describe latent dual language profiles in a large sample of school-aged Spanish-English bilingual children designated as English Learners (ELs) by their school district. Method: Data for this study include 847 Spanish-speaking ELs from kindergarten to third grade. Spanish and English narrative retell language samples were collected from all participants. Four oral language measures were calculated in Spanish and English, including the subordination index (SI), moving average type-token ratio (MATTR), Narrative Structure Scheme (NSS), and words per minute (WPM) using Systematic Analysis of Language Transcript (SALT). These eight indicator measures were used in a latent profile analysis to identify dual language profiles. Results: The optimal model represents a four-profile solution, including a Spanish-dominant group (average Spanish, low English), an English-dominant group (low Spanish, average English), and two balanced groups (a balanced-average group and a balanced-high group). Additionally, participants displayed uneven performance across language domains and distinct patterns of unique strength or weakness in a specific domain in one of their two languages. Conclusions: Findings from this study highlight the large variability in English and Spanish oral language abilities in school-aged Spanish-speaking ELs and suggest that a dichotomous classification of ELs vs English-proficient students may not be sufficient to determine the type of educational program that best fits a specific bilingual child’s need. These findings highlight the need to assess both languages across multiple language domains to paint a representative picture of a bilingual child’s language abilities. The dual language profiles identified may be used to guide the educational program selection process to improve the congruence among the linguistic needs of an individual child, teachers’ use of instructional language, and the goals of the educational program (i.e., improving English proficiency vs. supporting dual language development).
Article
Parental level of education, instruction time, and amount of language practice that children receive have enhanced our understanding of how bilingual and multilingual children learn to comprehend text. Guided by the simple view of reading and the interdependence hypothesis, this longitudinal study conducted in Canadian French immersion programs examined the (a) within- and cross-language association between oral language skills and reading comprehension of bilingual English–French and multilingual children and (b) patterns of growth, while controlling for possible influences of parental level of education and methods of instruction on reading achievement. The sample included 150 children tested once at the beginning of Grade 4 (T1) and again at the end of Grade 4 (T2) and in Grade 6 (T3). Individual growth modeling revealed that bilingual and multilingual children showed similar development in oral language and reading skills across the timeframe. Moreover, growth in English and French reading comprehension was associated with within-language variables. English reading comprehension in Grade 4 was also associated with cross-language variables, including French listening comprehension and vocabulary knowledge. Reading development in the second and third language is enhanced in contexts where classroom instruction, as well as social, economic, and educational opportunities to learn, is equivalent for all students.
Article
Full-text available
In this chapter, we examine studies pertaining to the relationship between English oral language proficiency (vocabulary, grammar, and listening comprehension) and various English reading skills among English-language learners. Although oral language proficiency is one of the many components that influence the development of literacy, it is reported on separately in this chapter, rather than as part of chapter 4, because of the important role it is presumed to play in the development of literacy in language-minority students. The chapter is organized according to literacy outcomes. We begin by reviewing studies that examine the relationship between diverse aspects of English oral proficiency and word-level reading skills in English, including word and pseudoword reading and spelling. This is followed by an examination of research on the relationships between various aspects of oral language proficiency in English and text-or discourse-level skills (i.e., reading comprehension and writing). Each of these skill sets is addressed within a developmental framework. Quantitative studies often measure one or two aspects of oral language proficiency (e.g., vocabulary, grammatical skills). The use of such measures enables researchers to compare performance among groups in a systematic way. Qualitative studies often use elicitation techniques to obtain speech samples from a relatively small number of participants. Some researchers believe that where oral language proficiency is concerned, the whole is bigger than the sum of the parts, and that naturalistic, authentic language samples provide more valid assessments of language proficiency. Both types of studies are included in this chapter.