Professional Documents
Culture Documents
Phonological Awareness and Beginning Reading in Spanish-Speaking ESL First Graders: Research Into Practice
Phonological Awareness and Beginning Reading in Spanish-Speaking ESL First Graders: Research Into Practice
85–111, 2002
Copyright © 2002 Society for the Study of School Psychology
Printed in the USA
0022-4405/02 $–see front matter
PII S0022-4405(01)00095-4
85
86 Journal of School Psychology
RESEARCH QUESTIONS
Four research questions are addressed in the research reported in this arti-
cle. First, is phonological awareness important in the beginning stages of
learning to read English when the student’s first language is Spanish? Sec-
ond, can the prior findings of cross-language transfer from phonological
awareness in the first language to reading (Durgunoglu et al., 1993) or
phonological awareness in the second language (Ciscero & Royer, 1995)
be replicated in another sample of Spanish-speaking ESL students? If so,
what are the instructional applications of these findings? Third, in the case
of children whose first language is Spanish and second language is English,
do language skills other than phonological awareness in Spanish or En-
Quiroga et al. 91
STUDY 1
Participants
Although there is an influx of immigrants to the metropolitan region
where the study took place, no individual school had a critical mass of any
one language minority; therefore, each of the schools working with our
center had only a few Spanish-speaking ESL first graders. Thus, it was nec-
essary to draw children from different schools and classrooms. First-grade
teachers in 10 schools and 15 classrooms in three school districts (urban,
suburban, and semirural) that were participating in other studies in the re-
search center were asked to call all parents of Spanish-speaking ESL stu-
dents to tell them about the possibility of participating in a research study.
As the schools were not able to provide statistics on the total number of
Spanish-speaking ESL children in first grade in their districts, we relied on
teacher judgment for extending the invitation to participate. Parental in-
formed consent was obtained in Spanish from parents of 30 children (14 fe-
males, 16 males) who granted permission for their children to participate.
These parents had all immigrated to the United States in the last decade.
The first and second authors determined through parental interview that
three of the families met criteria discussed in the introduction for bilingual
additive (supportive) and 27 met the criteria for bilingual subtractive.
As in the Durgunoǧlu et al. (1993) and Ciscero and Royer (1995) stud-
ies, the first graders in this study were Latino and their families were immi-
grants, but in contrast to those prior studies, the children were receiving a
very different instructional program at school. Instead of receiving some
instruction or most of their instruction in Spanish, the children were re-
ceiving no instruction in Spanish. Their regular instructional program was
only in English. Their pull-out ESL services, for the purpose of teaching
them to speak English, were provided in small groups by teachers who did
not speak Spanish. Groups could include children whose first language was
any of the other 88 languages spoken in the area, depending on demo-
graphics of a particular school. Thus, the first study evaluates whether the
cross-language transfer observed in two prior studies replicates when the
92 Journal of School Psychology
Assessment Battery
As in the Ciscero and Royer (1995) study, measures of phonological aware-
ness in both English and Spanish were included. As in the Durgunoǧlu et al.
(1993) study, measures of oral language proficiency in Spanish and En-
glish, letter knowledge, and real-word and pseudoword reading were in-
cluded. However, in contrast to Durgunoǧlu et al. (1993), who used exper-
imenter-designed measures of real-word and pseudoword reading, we used
only psychometric measures of real-word and pseudoword reading with na-
tional norms. For Spanish phonological awareness, the same three mea-
sures that Durgunoǧlu et al. (1993) used and two other measures that the
present article’s second and third authors (both fluent speakers of Span-
ish) created were used. Those two measures, created for purposes of this
study, were parallel in task to the measures of English phonological pro-
cessing that were used. The same stimulus words were not used because
they would not be phonologically equivalent across languages.
Oral language proficiency. Both the Spanish and English versions of the
Pre-LAS (Duncan & DeAvila, 1986) were given and scored per instructions
in the test manual. Each has identical formats (subtests) and administra-
tive procedures, but neither test is a direct translation of the other. A total
standard score based on four subscales was used in the data analyses be-
cause the subscales were highly intercorrelated (range .99 to 1.00) within
both the Spanish and English versions. These subscales included measures
of receptive language (a “Simon Says”-type task), expressive vocabulary
(naming common objects in a drawing), listening comprehension (select-
ing a picture that illustrates what the examiner has said), and language
comprehension and production (repeating verbatim phrases with differ-
ent syntactic structures, completing sentences, and retelling stories read by
the examiner).
Verbal intelligence. The Wechsler Intelligence Scale for Children (WISC)-
III Verbal Scale (Wechsler, 1991) was given to obtain a Verbal IQ (VIQ)
in English. The Comprehension-Knowledge cluster of the Woodcock-
Muñoz Pruebas de Habilidad Cognitiva-Revisada (Woodcock & Muñoz-
Sandoval, 1996a) was given to obtain a comparable VIQ in Spanish. Both
IQ measures were administered and scored according to instructions in
the test manuals. Standard scores for age were used, with a mean of 100
and standard deviation of 15.
Real-word and pseudoword reading. The Word Identification and Word
Attack subtests of the Woodcock Reading Mastery Test-Revised (WRMT-R;
Woodcock, 1987) were given to assess real-word and pseudoword reading,
respectively, in English. The comparable subtests on the Pruebas de Aprove-
chamiento-Revisada (Woodcock & Muñoz-Sandoval, 1996b) were given to
assess the same word reading skills in Spanish. All reading measures were
administered and scored according to instructions in the test manuals.
Standard scores for age were used, with a mean of 100 and standard devia-
tion of 15.
Procedures
All tests were administered by the first two authors who are native speakers
of Spanish and English or by the third author who is a bilingual specialist
fluent in Spanish. The battery took about 2 hr to administer and, depend-
ing on when schools allowed access to children, was given in either two 1-hr
sessions on different days or all on the same day with multiple frequent
breaks. Whenever children seemed tired or to not be giving an optimal
performance, testing was discontinued and resumed when they were rested
and invested in the tasks. The battery was given in a standard order, with
Spanish measures first and English measures second, because there were
too many measures to completely counterbalance order effects with the
Quiroga et al. 95
Table 1
Descriptive Statistics for First Graders (N 30) Whose First Language
Is Spanish and Whose Parents Are Immigrants
Measures M SD
Table 2
Correlations of Phonological Measures, Letter Knowledge, Oral Language Proficiency, and
Verbal Reasoning With Beginning Word Reading of Spanish-Speaking ESL Students
a
Woodcock and Muñoz-Sandoval (1996b).
b
Woodcock (1987).
c
Durgunoǧlu et al. (1993).
d
Appendix A.
e
Berninger et al. (1987).
f
Duncan and DeAvila (1986).
g
Woodcock and Muñoz-Sandoval (1996a).
h
Wechsler (1991).
WID Word Identification, WA Word Attack, ESL English-as-a-second language.
* p .05, **p .01, ***p .000.
significant between one of the eight measures evaluated (see Table 2) and
both criterion measures of word reading (real words and pseudowords)
within a language, then the finding was considered to be reliable. By this
criterion, all correlations between predictor and criterion measures in Ta-
ble 2 were reliable within a language.
Spanish phonological awareness. As shown in Table 2, the Durgunoglu
et al. (1993) Spanish phonological awareness composite based on three
tasks was significantly correlated with real-word and pseudoword reading
in both Spanish and English. Magnitudes of the correlations were similar
across Spanish and English word reading. Likewise, the total score on the
Spanish Rosner (Appendix A) was significantly correlated with word read-
ing in Spanish and English.
English phonological awareness. As shown in Table 2, the total score on
the English measure of phonological awareness correlated significantly
with both Spanish and English word reading. Again, the magnitude of the
correlations was about the same across word reading in the two languages.
Correlation between Spanish and English phonological awareness. The
Spanish phonological awareness measure based on blending, matching, and
segmenting was significantly correlated with the English phonological mea-
sure based on syllable and phoneme segmentation, r(27) .632, p .001.
Both the Spanish and English versions of the Modified Rosner tasks were sig-
Quiroga et al. 97
VIQ. Only VIQ was given because it has been shown to be a more valid
predictor of reading achievement than Performance IQ or Full Scale IQ
(Vellutino, Scanlon, & Lyon, 2000). As shown in Table 2, the Spanish IQ
was significantly correlated with word reading in Spanish but not in En-
glish. The English IQ was significantly correlated with word reading in
both Spanish and English.
Table 3
Hierarchical Multiple Regression of Best Predictors of WRMT-R Word Identification
and Word Attack for Spanish-Speaking ESL Beginning Readers
English WAa
English phonological awarenessb .59*** .66* .62* .68*
Spanish phonological awarenessc .09d .21d .22d
Letter knowledge .27 .27
English Pre-LASd .16d
Spanish Pre-LASd .14d
R2 .589 .591 .630 .640
English WIa
English phonological awarenessb .55** .62* .55* .52*
Spanish phonological awarenessc .09d .29d .28d
Letter knowledge .42* .40*
English Pre-LASe .05
Spanish Pre-LASe .03
R2 .547 .549 .646 .648
a
Woodcock (1987).
b
Berninger et al. (1987).
c
Durgunoǧlu et al. (1993).
d
Sign reflects suppressor effect due to correlation with other predictors and pattern of correlation with
criterion.
e
Duncan and DeAvila (1986).
WRMT-R Woodcock Reading Mastery Test-Revised, WA Word Attack, WI Word Identification,
ESL English-as-a-second language.
*** p .001, ** p .01, * p .05.
Discussion
The answer to the first research question (see Research Questions) is that
phonological awareness is related to learning to read English when one’s
first language is Spanish and reading instruction is in English. Both Span-
ish and English phonological awareness predicted English word reading.
This finding is of interest because (a) Spanish has a more transparent pho-
nology (especially at the syllable level) than spoken English, and (b) Span-
ish has more regular spelling-phoneme correspondences than does En-
glish for decoding.
The answer to the second research question is that the findings of both
Durgunoǧlu et al. (1993) and Ciscero and Royer (1995) replicated and
generalized to a population that received its reading instruction in English
rather than Spanish. These findings were that phonological awareness in
Spanish, the first language, predicted reading in English, the second lan-
guage (Durgunoǧlu et al., 1993), and that phonological awareness corre-
lated across the first and second language (Ciscero & Royer, 1995). These
findings replicated even though their children read better in Spanish, the
language of their instruction, than in English; our children read better in
Quiroga et al. 99
English, the language of their instruction, than in Spanish (see Table 1).
An educational application of these findings is that if reading instruction is
only in English, it might be supplemented with phonological awareness
training in both English and Spanish. Possibly, the ESL teacher or other
Spanish-speaking professional could provide the Spanish phonological
awareness training. As second-language learners move from contextualized
language proficiency to decontextualized language proficiency (Cummins,
1984, 1991), one of the decontextualized skills they need to acquire is pho-
nological awareness. Coupling phonological awareness in the second lan-
guage with phonological awareness in the first language may be beneficial.
At the same time, Spanish phonological awareness did not contribute
uniquely in additive fashion to pseudoword or real-word reading. School
psychologists who do not speak Spanish and can only give phonological
awareness measures in English may still obtain valuable assessment infor-
mation for planning and evaluating early intervention for Spanish-speak-
ing ESL students; phonological awareness training in English only is better
than no phonological awareness training for Spanish-speaking ESL stu-
dents.
Regarding the third research question, the second major finding of Dur-
gunoǧlu et al. (1993)—that oral language proficiency in the second lan-
guage did not predict reading achievement in the second language—was
not confirmed. Oral language proficiency in the second language pre-
dicted reading achievement in the second language probably because
reading instruction was in the second language and ability to understand
oral instructional language influences reading development. Thus, the dif-
ference in findings across studies is attributable to the difference in the
language of instruction. When Spanish-speaking ESL children receive
reading instruction in English, their English oral language proficiency in
vocabulary, syntax, and discourse skills will affect how they benefit from
that instruction. However, as Durgunoǧlu et al. (1993) argued, reading in-
struction should not be withheld until children reach a certain level of oral
language proficiency. Our findings based on correlations do not support
that educational recommendation. Children will probably benefit from
concurrent instruction in both oral and written English so that their ability
to process oral instructional language develops along with their ability to
read English.
Also pertinent to the third research question are the findings that al-
though phonological awareness was correlated across the first and second
language, VIQ was not. Thus, beginning readers may draw on phonologi-
cal ability that transcends specific languages but their ability to use lan-
guage to reason may be language specific.
Future research might explore whether the findings for all three re-
search questions hold for schools having a critical mass of Spanish-speak-
ing children whose families have lived in the United States for more than
100 Journal of School Psychology
STUDY 2
Participants
The 4 lowest achieving males and 4 lowest achieving females participated
in a one-to-one tutorial. Their mean word reading achievement prior to the
tutorial was nearly two standard deviations below the mean in real-word
and pseudoword reading; see Table 4. This study used an instructional de-
sign experiment (Brown, 1992) that provided empirically supported in-
structional components to achieve the desired outcome of improved word
reading (see the introduction).
Procedures
Each child received twelve 30-min lessons that were taught by either the
second or third author (who were fluent in Spanish and English) in a one-
on-one session. The tutorials occurred twice a week over a 6-week period
and were always structured in the same way for each child: Children re-
ceived phonological awareness training in both Spanish and English, ex-
plicit instruction in alphabetic principle in which the correspondences be-
tween one- and two-letter spelling units and phonemes were modeled out
of word context and in word context (Berninger, Stage, Smith, & Hilde-
brand, 2001), practice in reading and rereading text, and comprehension
monitoring. That is, children received training in phonological awareness
in both their first and second language, and reading instruction was aimed
Table 4
Monitoring Achievement Before and After Tutorial Intervention
1 96 99 88 104
2 86 98 78 94
3 72 81 76 86
4 52 78 68 58
5 54 62 70 78
6 78 82 76 77
7 72 82 64 63
8 72 95 76 94
M 72.8 84.7 74.5 81.8
SD 13.8 11.6 6.8 14.9
at all levels of language (subword, word, and text) in the same instruc-
tional session.
At the beginning of the session, the Spanish-speaking children com-
pleted one lesson in the Spanish Phonological Awareness Training Pro-
gram. Over the course of the tutorial, they completed the first 12 lessons in
this 24-lesson program, which is an analog in terms of tasks (not words
used) of a research-validated English phonological awareness training pro-
gram (Berninger, 1998). The first six lessons focused on syllable segmenta-
tion activities and the last six lessons focused on phoneme segmentation
activities. In each lesson, they played four kinds of “sound games” in Span-
ish: detecting whether a target sound was hidden in a group of words; de-
tecting whether a target sound was missing in a partial word pronounced
by the tutor; saying a word without a designated target sound; and saying a
word with a designated sound substituted for a target sound. Appendix B
contains a sample lesson from the program. All lessons are available from
the second author.
Next, children played “sound games” in English with words selected
from the book they would read later in the session. In this way their phono-
logical awareness training was yoked close in time to specific words, first
presented in phonological form and then presented in written form. Chil-
dren counted the number of syllables in spoken words and used colored
counters to denote each phoneme in the spoken word.
Next, children learned the alphabetic principle of spelling–phoneme
correspondence in English using the Talking Letters Program (Berninger,
1998). This program teaches frequent spelling–sound correspondences in
primary-grade reading materials in written English. A student card summa-
rizes the one- and two-letter spelling units along with pictured words con-
taining the associated phonemes. As in direct instruction, the tutor mod-
eled and the child repeated, thus ensuring success on each teaching trial.
The tutor named the letter (e.g., a) or letters (e.g., au) in the spelling unit,
then named the pictured word producing the target phoneme (e.g., apple
or auto), and then made the target phoneme sound in isolation. The child
repeated the sequence of naming the letter and pictured word and making
the target phoneme in order to build automatic associations between the
spelling unit and the phoneme. The entire set of 92 spelling–sound corre-
spondences was divided into three sections that were alternated every
three lessons for a total of four exposures each. In the first section, 29
spelling–phoneme correspondences involving single consonants or short
vowels were modeled. In the second section, 27 spelling–sound correspon-
dences involving two-letter consonant or vowel combinations of moderate
predictability or difficulty were modeled. In the third section, 36 spelling–
sound correspondences involving two-letter vowel or consonant combina-
tions of moderate or low predictability were modeled. This approach to
teaching alphabetic principle fosters growth in phonological awareness be-
102 Journal of School Psychology
GENERAL DISCUSSION
Cummins (1980) found that pull-out bilingual programs focus on the
more visible manifestations of language, which do not carry over to areas
other than basic communication. He referred to this level of language pro-
ficiency as “basic interpersonal communications skills” (BICS). Within the
highly structured milieu of an elementary school, students can appear to
have more developed language ability when, in reality, they are function-
104 Journal of School Psychology
ing within the BICS level of ability. Cummins (1981) questioned the as-
sumption that a BICS level of language proficiency is adequate for reading
comprehension. Many academic tasks require verbal reasoning and a level
of language proficiency Cummins (1981) dubbed as “cognitive academic
language ability” (CALPS). He suspected that 3 to 4 years are needed to de-
velop CALPS under supportive conditions. Thus, beginning ESL readers
may appear to have sufficient proficiency with English to take advantage of
the reading instruction provided in English for word recognition but not
for reading comprehension.
Maldonado-Colon (1995) found that both parents and instructors over-
estimate the oral language abilities of ESL students as well as underesti-
mate what it takes to operate in an academic environment. Teachers may
refer ESL students at the BICS level who lack proficiency at the CALPS
level for psychological evaluation of learning disability. School psycholo-
gists can help teachers understand that the students may not have a learn-
ing or language disability because they are making progress in the acquisition
of English, demonstrating that the student can learn language (Cummins,
1984). Nevertheless, these students may also need additional instructional
experiences that help them move beyond the BICS level to the CALPS
level, and experimental controlled research as well as design experiments
are needed on this issue. We emphasize that the Spanish IQ measure used
in this study (Woodcock & Muñoz-Sandoval, 1996a) may be a better pre-
dictor of CALPS than BICS level of proficiency. Although that measure was
not correlated with beginning word reading in Spanish, it may be very valu-
able in assessing whether older students have moved beyond a BICS level
to a CALPS level of proficiency for academic learning later in schooling
(cf., Cummins, 1981), and is an important tool for school psychologists to
use when needed.
Implications for the Practice of School Psychology
Assessment. Study 1 showed that IQ and oral language proficiency in ei-
ther the first or second language did not uniquely predict beginning word
reading in English for Spanish-speaking ESL students. Thus, it may not be
necessary to include a time-consuming measure of oral proficiency or VIQ
in a screening battery for supplemental early intervention in reading for
Spanish-speaking ESL beginning readers. The differences between the En-
glish and Spanish VIQ measures may be related to differences in the norm-
ing sample, but we emphasize that neither VIQ uniquely predicted English
word reading in these Spanish-speaking ESL beginning readers. However,
Spanish and English phonological awareness measures predicted begin-
ning word reading in English, but only English phonological awareness
predicted it uniquely for Spanish-speaking ESL students instructed in En-
glish. For these children, a measure of English phonological awareness
might be included in Tier 1 assessment to identify at-risk students for early
Quiroga et al. 105
CONCLUSION
This research focused on only one ESL student group: those who speak
Spanish as a first language, are instructed in school in English, are begin-
ning readers in first grade, and live in families that have recently immi-
grated to the United States. Clearly, more research is needed with different
language-speaking ESL groups and with other groups of Spanish-speaking
ESL students. However, the two studies reported here do provide some in-
sight into how school psychologists might serve ESL students effectively by
applying research to practice. Rather than focusing on the oral language
skills a Spanish-speaking ESL student may lack in the second language,
schools should begin to work with these children as early as first grade on
learning to read. Early intervention that provides phonological awareness
training in both the first and second languages or in the second language
only, and explicit instruction in the alphabetic principle in the second lan-
guage may help these students learn to read. However, as these students
progress through school, they will also require instruction geared at im-
proving (a) oral proficiency for understanding instructional language in
English, and (b) reading comprehension for academic subjects. The most
important question a school psychologist can ask is not whether the ESL
child has a disability but, rather, what kinds of instructional support and
consultation can be put in place early in schooling to stimulate literacy at
school, at home, and in the community.
ACKNOWLEDGMENTS
This research was supported by a minority fellowship to the first author as a
supplement to a Multidisciplinary Learning Disability Center Grant from
the National Institute of Child Health and Human Development (P 50
33812-03 and -04) for which the last author is Principal Investigator and
the fourth author is the statistician. The second author, who was a gradu-
Quiroga et al. 107
ate research assistant, and third author, who was the research coordinator
for the Center, developed the Spanish version of the Modified Rosner as-
sessment instrument and the Spanish phonological awareness training pro-
gram; they also delivered the instructional interventions in this study. The
authors thank the participating children, parents, and teachers for their
cooperation, and Mardean Francis for assistance in recruiting participants.
The authors thank anonymous reviewers for helpful comments.
REFERENCES
Adams, M. (1990). Beginning to read: Thinking and learning about print. Cambridge, MA:
MIT Press.
Auguste, D., & Hakuta, K. (Eds.). (1998). Educating language-minority children (pp. 21–
32). Washington, D.C.: National Academy Press.
Berninger, V. (1998). Process assessment of the learner: Guides for reading and writing interven-
tion. San Antonio, TX: The Psychological Corporation.
Berninger, V. (in press). Best practices in reading, writing, and math assessment-inter-
vention links: A systems approach for schools, classrooms, and individuals. In A. Tho-
mas & J. Grimes (Eds), Best practices in school psychology IV . Bethesda, MD: National
Association of School Psychologists.
Berninger, V., & Abbott, R. (1994). Redefining learning disabilities: Moving beyond ap-
titude-achievement discrepancies to failure to respond to validated treatment proto-
cols. In G. R. Lyon (Ed.), Frames of reference for the assessment of learning disabilities: New
views on measurement issues (pp. 163–202). Baltimore: Brookes.
Berninger, V., Abbott, R., Brooksher, R., Lemos, Z., Ogier, S., Zook, D., & Mostafapour,
E. (2000). A connectionist approach to making the predictability of English orthog-
raphy explicit to at-risk beginning readers: Evidence for alternative, effective strate-
gies. Developmental Neuropsychology, 17, 241–271.
Berninger, V., Stage, S., Smith, D., & Hildebrand, D. (2001). A three-tier model for as-
sessment-intervention links for preventing and treating reading and writing disabil-
ity. In J. Andrews, H. Janzen, & D. Saklofske (Ed.), Ability, achievement, and behavior
assessment. A practical handbook (pp. 195–223). New York: Academic Press.
Berninger, V., Thalberg., S., DeBruyn, I., & Smith, R. (1987). Preventing reading disabili-
ties by assessing and remediating phonemic skills. School Psychology Review, 16, 554-565.
Brown, A. (1992). Design experiments: Theoretical and methodological challenges in
creating complex interventions in classroom settings. The Journal of Learning Sciences,
2, 141–178.
California State Department of Education. (1985). Case studies in bilingual education: First
year report (Federal Grant # G008303723). Sacramento, CA: California State Depart-
ment of Education.
Ciscero, C., & Royer, J. (1995). The development and cross-language transfer of phono-
logical awareness. Contemporary Educational Psychology, 20, 275–303.
Cummins, J. (1980). The construct of language proficiency in bilingual education. Paper pre-
sented at the Georgetown Roundtable of Language and Linguistics, Georgetown
University.
Cummins, J. (1981). The Cross-lingual dimensions of language proficiency: Implications
for bilingual education and the optimal age issue. TESOL Quarterly, 24, 175–187.
Cummins, J. (1984). Bilingualism and special education: Issues in assessment and pedagogy.
San Diego, CA: College-Hill Press.
108 Journal of School Psychology
APPENDIX A
Modified Rosner (Copyrighted Spanish version)
APPENDIX B
Sample Lesson From the Spanish Phonological Awareness Assessment
and Training Program (© SPAAT)a
Sílabos–Lección 1
DESCUBRE LA ESCONDIDA
Di fin. ¿Se esconde la palabra fin en la palabra final?
perfil?
finca?
fingir?
a
Permission to reproduce by Lemos-Britton and Mostafapour (1997).
APPENDIX C
Sentences Added to the Popcorn Shop (Low, 1993, Scholastic Books)