You are on page 1of 14

Developmental Psychology

2002, Vol. 38, No. 6, 934 947

Copyright 2002 by the American Psychological Association, Inc.


0012-1649/02/$5.00 DOI: 10.1037//0012-1649.38.6.934

Oral Language and Code-Related Precursors to Reading:


Evidence From a Longitudinal Structural Model
Stacey A. Storch and Grover J. Whitehurst

This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

State University of New York at Stony Brook

This study examined code-related and oral language precursors to reading in a longitudinal study of 626
children from preschool through 4th grade. Code-related precursors, including print concepts and
phonological awareness, and oral language were assessed in preschool and kindergarten. Reading
accuracy and reading comprehension skills were examined in 1st through 4th grades. Results demonstrated that (a) the relationship between code-related precursors and oral language is strong during
preschool; (b) there is a high degree of continuity over time of both code-related and oral language
abilities; (c) during early elementary school, reading ability is predominantly determined by the level of
print knowledge and phonological awareness a child brings from kindergarten; and (d) in later elementary
school, reading accuracy and reading comprehension appear to be 2 separate abilities that are influenced
by different sets of skills.

a childs reading ability, including semantic (word knowledge,


expressive and receptive vocabulary), syntactic (knowledge of
word order and grammatical rules), and conceptual knowledge
(e.g., Bishop & Adams, 1990; Bowey, 1986; Demont & Gombert,
1996; Gillon & Dodd, 1994; Vellutino & Scanlon, 1985; Vellutino,
Scanlon, & Spearing, 1995; Vellutino, Scanlon, & Tanzman,
1991), as well as narrative discourse (the ability to construct an
original story and retell a recently heard story; e.g., Klecan-Aker &
Caraway, 1997; Perfetti, 1985; Snyder & Downey, 1991).
To understand the role of these code-related and oral language
skills as precursors to eventual literacy, it is important to consider
the reading process and the development of reading ability. Reading requires the coordination and interaction of multiple skills,
including recognition of individual letters, translation of letters
into sounds, determination of the meaning of a word, and interpretation and understanding of the text as a whole (Adams, 1990).
Although these processes may be inseparable in the mature, fluent
reader, these processes are not integrated initially; if this were the
case, children would instantly be fluent readers (Adams, 1990;
Van Kleeck, 1998).
Several researchers have advanced the idea that different emergent literacy skills make their most significant contributions to
reading achievement at different points in development (Speece,
Roth, Cooper, & de la Paz, 1999; Whitehurst & Lonigan, 1998).
Research demonstrates that print knowledge and phonological
awareness play a critical role in the early elementary school years
when children have just begun reading instruction and are in the
code-cracking stage. Not only have these skills been shown to
make the task of decoding printed words easier (Beck & Juel,
1999; Bradley & Bryant, 1985; Byrne & Fielding-Barnsley, 1993,
1995; Tunmer, Herriman, & Nesdale, 1988), but there also appears
to be a reciprocal relationship between phonological awareness
and rudimentary reading skill (i.e., letter knowledge) in preschoolage children (Burgess & Lonigan, 1998).
It is commonly assumed that early difficulties in oral language
lay the foundation for later reading problems (Roth, Speece, Cooper, & de la Paz, 1996). Specific language impairment during the

Research has demonstrated that a range of skills developed by


the preliterate child lays the foundation for later reading and
writing ability (e.g., Whitehurst & Lonigan, 1998). Whitehurst and
Lonigan defined emergent literacy as the skills, knowledge, and
attitudes that are developmental precursors to reading and writing
(Sulzby & Teale, 1991; Teale & Sulzby, 1986), as well as the
environments that support these developments (e.g., shared book
reading; Whitehurst et al., 1988, 1994). The notion of emergent
literacy implies a continuum between prereading and reading, in
which literacy-related behaviors and activities taking place during
the preschool period are essential aspects of the course of literacy
development. The code-related skill base of emergent literacy
includes the conventions of print (e.g., knowing that writing goes
from left to right and top to bottom across a page), beginning forms
of writing (e.g., writing ones name), knowledge of graphemes
(e.g., naming letters of the alphabet) and graphemephoneme
correspondence (e.g., that the letter b makes the /b/ sound), and
phonological awareness (e.g., that the word bat begins with the /b/
sound). Research has shown positive correlations and longitudinal
continuity between individual differences in these code-related
skills and later differences in reading ability.
In addition to code-related skills, a variety of oral language
skills during the preschool period have been shown to contribute to
Stacey A. Storch and Grover J. Whitehurst, Department of Psychology,
State University of New York at Stony Brook.
This research was supported by Grants 90CD0957, 90YD0026, and
90YD0079 from the U.S. Administration for Children and Families. Views
herein are those of the authors, and these views have not been cleared by
the grantors.
We are deeply appreciative of the support of the administration, staff,
and families of Long Island Head Start, Childrens Community Head Start,
and numerous school districts in Suffolk County, New York, without
whom this work could not have been accomplished.
Correspondence concerning this article should be addressed to Stacey
A. Storch, Department of Psychology, State University of New York at
Stony Brook, Stony Brook, New York 11794-2500. E-mail: stacey.storch@
sunysb.edu
934

This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

ORAL LANGUAGE AND CODE-RELATED SKILLS

preschool period predicts later reading difficulties (Bishop & Adams, 1990), and oral language problems represent a significant risk
factor for reading disability during the elementary school years and
beyond (Catts, 1993; Roth & Spekman, 1994). Indeed, several
studies have demonstrated correlations between oral language skill
and reading ability within typically developing, reading-delayed,
and language-delayed children (e.g., Bishop & Adams, 1990;
Butler, Marsh, Sheppard, & Sheppard, 1985; Pikulski & Tobin,
1989; Scarborough, 1989; Share, Jorm, MacLean, & Mathews,
1984).
This literature is complicated, however, by the recognition of
the correlation between oral language abilities, such as vocabulary
size, and phonological skills. One set of reviewers found little
support for the influence of language abilities on reading beyond
the influence of phonological awareness skills (Roth et al., 1996).
Indeed, many researchers have contended that oral language skills,
such as semantic and syntactic knowledge, do not play a central
role in the development of early reading skills (e.g., Bryant,
MacLean, & Bradley, 1990; Speece et al., 1999; Vellutino et al.,
1996), an argument based largely on the assertion that the relationship between oral language and reading may be driven by
general intelligence, an often unmeasured variable. When controlling for IQ and socioeconomic status, for example, Bryant et al.
(1990) found that whereas phonological awareness accounted for
unique variance in first grade reading ability, semantic, syntactic,
and metasyntactic knowledge did not.
On the other hand, many researchers have warned that the role
of phonological abilities has been overemphasized, and argued that
other components of oral language, such as semantic and syntactic
ability, play a key role in reading development (e.g., Bishop,
1991). Empirical evidence supporting this argument exists on a
number of fronts. A basic piece of evidence comes from Tunmer
and Nesdale (1985), in which they reported that the relationship
between phonological awareness and decoding ability is nonlinear;
that is, phonological awareness was necessary, but not sufficient,
for the acquisition of decoding ability. Catts, Fey, Zhang, and
Tomblin (1999) found that oral language skills in kindergarteners
are partially independent of phonological abilities, and that both
sets of skills are significant independent predictors of word recognition. Bishop and Adams (1990) found that in a sample of
children with language disorders, reading problems were best
correlated with severity of the childrens syntactic difficulties.
Menyuk et al. (1991) found that reading problems were best
predicted by deficiencies in awareness of semantic relations.
Although the results of Catts et al. (1999) caution against
emphasizing phonological abilities to the exclusion of oral language abilities in early readers, Catts et al. contended that the
majority of studies presented as support for the claim that phonological abilities act as a mediator of the relationship between oral
language abilities and reading achievement are generally aimed at
explaining word recognition, rather than reading comprehension.
Thus, it appears that the role of oral language abilities is called into
question primarily in the early stages of reading acquisition, as
opposed to the later stages. Indeed, the relationship between language and reading comprehension is stronger than that between
language and reading accuracy; moreover, the association between
language and reading performance is greater for older children
than for children in the early stages of learning to read (Gillon &
Dodd, 1994; Share & Silva, 1987; Vellutino et al., 1991; White-

935

hurst & Lonigan, 1998). As children progress and attempt to


comprehend units of text larger than individual words, oral language skills, particularly semantic knowledge, become increasingly important (Mason, 1992; Nation & Snowling, 1998; Snow,
Barnes, Chandler, Hemphill, & Goodman, 1991; Whitehurst,
1997). A link between other oral language skills, such as narrative
discourse and syntactic awareness, and later reading achievement
has also been shown (e.g., Bowey, 1986; Demont & Gombert,
1996; Feagans & Applebaum, 1986; Tunmer et al., 1988).
The research described above highlights the varied and continually changing nature of the domain of literacy development. Yet,
despite evidence for some well-established connections (e.g., between phonological awareness and decoding skills), the relationship between other skills, especially oral language ability, and
reading achievement is unclear and subject to debate. One reason
for the lack of a consensus in the literature is that studies have
tended to focus on only a few emergent literacy skills at a time and
no two studies have assessed exactly the same skills. The variation
in outcome measures used by researchers to assess reading
achievement, which have included word and nonword decoding
tasks, passage comprehension, and reading fluency tasks, presents
an additional problem in determining the relationship between
emergent literacy skills and reading achievement. For example,
results of the Gillon and Dodd (1994) study demonstrated that the
relationship between phonological abilities, syntactic and semantic
skills, and reading performance depends on whether the outcome
measure is reading accuracy or reading comprehension.
A problem specific to studies investigating the relationship
between oral language skills and reading outcomes is the sampling
procedure used. Studies of this kind generally follow one of two
strategies: follow-up of children diagnosed with language impairment at a young age, or retrospective study of children with
reading difficulties in elementary school. In many of these studies,
inclusion in a language-delayed sample occurred without consideration for a childs nonverbal ability (e.g., Richman, Stevenson, &
Graham, 1982; Silva, Williams, & McGee, 1987). As a result, many
children included in these studies had general developmental delays,
with language impairment being only one symptom (Bishop &
Adams, 1990). Often, those children whose early language delay
was associated with reading problems tended to show low ability
on a range of academic measures. Moreover, as Bishop and Adams
(1990) pointed out, when specific language delays are considered,
the relationship between oral language skill and reading outcomes
often disappears. For instance, in a study conducted by Silva,
McGee, and Williams (1985), when looking at only those children
whose reading performance was significantly below what would
be predicted by their nonverbal IQ at age 9, the researchers found
that they did not differ from control children on scores on language
tests given at various times between ages 3 and 9. Thus, it is clear
that language disorders cannot be considered a unitary category.
Although some children have deficits in all aspects of language,
others may only have specific problems, for instance, phonological
awareness deficits. The type of language problem being studied
impacts the nature and strength of the relationship between oral
language abilities and reading performance.
In addition to the varying and sometimes conflicting evidence
regarding the role of various domains of literacy in the development of reading ability, the literature also suffers from a paucity of
studies examining the longitudinal relationship between the oral

This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

936

STORCH AND WHITEHURST

language and code-related domains. Studies either have tended to


take a cross-sectional approach (e.g., Chaney, 1992; Nation &
Snowling, 1998; Shankweiler et al., 1999) or have involved longitudinal frames that spanned only small portions of the developmental course of reading acquisition, with a large majority of
longitudinal studies focused on the period between kindergarten
and second grade (e.g., Bryant et al., 1990; Catts et al., 1999;
Lonigan, Burgess, & Anthony, 2000; Share et al., 1984; Speece et
al., 1999). Though the importance of early literacy skills for later
reading achievement is well accepted, the role of emergent literacy
skills from the time of their development in preschool throughout
late elementary school has yet to be systematically investigated.
This is especially problematic when considering that some relationships, for instance, those between semantic and syntactic abilities and reading comprehension, do not appear until the later
elementary school grades.
In the present study, we aimed to provide a more conceptually
coherent and empirically powerful examination of the role of both
code-related and oral language precursors in the development of
reading ability than has heretofore been reported. We used structural equation modeling to map the relationship between the coderelated skills, language ability, and later reading skills of children
from low-income families who have been followed from the time
they were in preschool through the fourth grade. Structural modeling allowed us to evaluate the effects of multiple emergent
literacy skills on reading achievement throughout elementary
school. Furthermore, structural modeling allowed us to address
questions about the changing nature of reading from the beginning
of formal reading instruction through the fourth grade.
The connections between code-related skills, oral language, and
reading achievement that were hypothesized in the model reported
herein were drawn from prior research and theory. Three main
findings from the literature or conceptual distinctions were utilized. First, we divided our battery of emergent literacy measures
into two domains code-related and oral language skills on the
basis of research demonstrating that skills comprising these two
domains may be most strongly related to reading development at
different points in the process of reading acquisition. Measures of
print knowledge, emergent writing, and phonological awareness
comprised the code-related domain, as all of these skills have been
shown to relate to the development of the ability to crack the
alphabetic code relating graphemes to phonemes. Measures of
semantic, syntactic, and conceptual knowledge comprised the oral
language domain. It is important to note that phonological awareness, though clearly a skill involving sensitivity to and awareness
of oral language, was included with the code-related skills. This
decision was based on two threads of research. The first, by
Whitehurst and colleagues (Storch & Whitehurst, 2001; Whitehurst & Fischel, 2000; Whitehurst & Lonigan, 1998), demonstrated that semantic, syntactic, and conceptual measures of oral
language could be statistically separated from phonological awareness, print knowledge, and emergent writing. The second, by
Speece and colleagues (e.g., Roth et al., 1996; Speece et al., 1999),
argued that the domain of oral language does not have a homogenous effect on reading acquisition; rather, factors such as phonological awareness have their greatest impact early in the development of reading ability, whereas other linguistic factors, such as
understanding of narrative discourse, have their impact later in the
reading sequence.

Second, we modeled code-related skills and oral language skills


as being linked during the preschool and kindergarten periods.
This decision was based on research demonstrating a high degree
of association between these two sets of skills during this time,
in particular, language, letter knowledge, emergent writing, and
knowledge of lettersound correspondence (Whitehurst & Lonigan, 1998). The direction of influence between these two sets of
skills was modeled as flowing from the oral language skills to
code-related skills, on the basis of our knowledge that language
skills, such as oral vocabulary, develop much earlier in a childs
life than do code-related skills, such as knowledge of letters and
their sounds.
Last, we hypothesized that reading achievement in the early
elementary school grades would be best conceptualized as a single
factor incorporating both reading accuracy and reading comprehension tasks, whereas in later elementary school grades we conceptualized reading as consisting of two factors. Research has
demonstrated that in the early stages of reading development,
reading accuracy and reading comprehension skills appear to be
closely intertwined, and the relationship between word recognition
and reading comprehension ability is strong (e.g., Hoover &
Gough, 1990; Shankweiler et al., 1995, 1999). However, in older
children there are discrepancies of varying degrees between reading accuracy and reading comprehension ability (Shankweiler et
al., 1999), suggesting that these skills, though clearly related, are
not one in the same. Furthermore, reading accuracy and reading
comprehension are differentially affected by other literacy skills,
particularly phonological awareness (e.g., Bowey, 1994; Wagner,
Torgesen, & Rashotte, 1994; Wagner et al., 1997) and semantic
and syntactic skills (e.g., Gillon & Dodd, 1994; Vellutino, Scanlon, & Tanzman, 1991; Whitehurst & Lonigan, 1998), respectively, which highlights the need to evaluate these domains of
reading achievement separately.

Method
Participants
The sample consisted of 626 four-year-olds who attended classrooms in
one of eight selected Head Start centers in Suffolk County, New York.
These children belonged to one of three cohorts who attended Head Start
during the 19911992, 19921993, or 19931994 school year, and whose
birthdates made them eligible to enroll in public kindergarten the following
year. After Head Start, the children dispersed into 22 school districts,
with 9 of those districts holding substantial numbers of participants and
each of the remaining districts having 8 or fewer children from the sample.
The sample of children was 52% male. The ethnic breakdown of the
sample was 39% African American, 34% Caucasian, 16% Latin American,
and 11% other or not identified. This closely mirrors the ethnic composition of Head Start on a national level in 1999 (U.S. Department of Health
and Human Services, 1999). All participants families qualified for admission into Head Start. The median reported family income of children
attending Head Start during the 19921993 year was $10,500, as determined by Head Start records.

Measures
Childrens language and literacy skills were assessed six timesin the
spring of Head Start, kindergarten, and the first through fourth grades on
the measures described below. Examiners for all assessments were doctoral
students in clinical psychology who had extensive experience in assessing

937

This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

ORAL LANGUAGE AND CODE-RELATED SKILLS


children. Each year, testing for each child generally occurred across two
sessions and took place in a quiet area away from other children. Each
assessment device was standardized, normed on a national sample of
children, and had internal reliability of .80 or higher, as determined from
the standardization sample.
Code-related skills. The children were tested in the spring of their
Head Start year and followed in the spring of kindergarten on 12 subtests
from the Developing Skills Checklist (DSC; CTB, 1990), an extensive
battery of standardized school readiness tasks. These subtests were drawn
from four subscales of the DSC that pertain to literacy development:
Memory (naming pictured letters, identifying sounds and letters), Auditory
(segmenting sentences, segmenting compound and regular words, rhyming), Print Concepts (identifying people engaged in reading, differentiating
print from pictures and letters from numerals, identifying functions of
print, identifying components of written communication), and Writing and
Drawing Concepts (printing first name, drawing a person, message-writing
mechanics). Each of the 12 subtests consists of multiple questions that are
administered individually in the typical style of a standardized test of
language or intelligence.
We have engaged in extensive prior analyses to determine the structure
and arrangement of these 12 code-related skills into statistically and
conceptually meaningful factors: These factors are print principles, emergent writing, and phonological awareness (Storch & Whitehurst, 2000).
Results of this prior work demonstrate that specific code-related skills have
differential ability to predict later reading achievement depending on the
age at which the skills are assessed. For instance, knowledge of letter
sound correspondence as assessed in kindergarten has a strong relationship
to reading ability in Grade 2; however, this same skill as assessed in Head
Start does not have significant predictive ability (Storch & Whitehurst,
2000). Thus, the DSC subtests used in the present model represent the best,
most predictive profile of code-related skills for each of the Head Start and
kindergarten time points. As a result, the model includes 7 of the 12 DSC
subtests in Head Start and 6 in kindergarten.
Oral language skills. The childrens oral language skills were assessed
in the spring of Head Start and kindergarten using the Peabody Picture
Vocabulary TestRevised (PPVTR; Dunn & Dunn, 1981), a test of
receptive vocabulary, and the One Word Picture Vocabulary Test (One
Word; Gardner, 1990), a test of expressive vocabulary. Preschool assessment also included the Renfrew Bus Story (BUS; Glasgow & Cowley,
1994), a test of narrative recall in which children are required to use
pictures to retell a recently heard story. The Information subtest of the BUS
was used for the present analysis and reflects the extent of a childs
memory for the content of the story. Kindergarten assessment also included
the Basic Concepts subtest from the Clinical Evaluation of Language
FundamentalsPreschool (CELFP; Wiig, Secord, & Semel, 1992), a test
of conceptual knowledge.
Additionally, the PPVTR was administered to the children at the end of
first through fourth grades. Assessment in first grade also included the
Word Structure subtest from the Clinical Evaluation of Language FundamentalsRevised (CELFR; Semel, Wiig, & Secord, 1987), a test of
syntactic ability.
Reading measures. Standardized tests were used to assess the reading
achievement of the children in our sample once they began formal reading
instruction. The children were assessed at the end of first grade on the
Word Reading subscale of the Stanford Achievement TestEighth Edition
(SAT, hereafter referred to as Word Reading; Psychological Corporation,
1989), a test of the ability to match printed words with pictures, and on the
Reading subscale of the Wide Range Achievement TestRevised
(WRATR, hereafter referred to as Reading; Jastak & Wilkinson, 1984), a
test of the ability to read aloud printed words of increasing difficulty. In
addition, the children were assessed on the Word Attack subscale of the
Woodcock Reading Mastery TestsRevised (WRMTR; hereafter referred to as Word Attack; Woodcock, 1987), a test of the ability to sound
out printed pseudowords. At the end of second grade, the Word Reading

subtest from the SAT and the Word Attack subtest from the WRMTR
were once again administered. Additionally, to reflect the changing abilities of the children, the Reading subtest of the WRATR was replaced with
two more complex tasks from the SAT: the Reading Comprehension
subtest, a test of the ability to extract meaning from and answer questions
about short paragraphs, and the Word Study Skills subtest, a test of reading
accuracy, including the ability to recognize relationships between letters
and sounds within the context of words, to identify spoken words in print,
and to divide words into syllables.
Administration of the Word Attack, Word Study Skills, and Reading
Comprehension subtests was repeated in the third and fourth grade. Additionally, a new test was added, the Reading Vocabulary subtest of the SAT,
which measures childrens ability to associate printed words with their
meanings.
The above reading measures can be divided conceptually into two
domains: reading accuracy and reading comprehension. The reading accuracy domain includes those subtests that focus on a childs ability to sound
out individual words (both actual and nonsense words), whereas the reading comprehension domain includes those tests that focus on a childs
ability to determine the meaning of words and text. Thus, WRAT Word
Reading, SAT Word Reading, Word Attack, and SAT Word Study Skills
are considered reading accuracy tasks, whereas SAT Reading Comprehension and SAT Reading Vocabulary are considered reading comprehension
tasks.

Results
Descriptive Statistics
In our analyses we used raw scores or age equivalents as results
from all child assessments rather than standard scores because
some of our subtests did not generate standard scores. In addition,
age standardized scores may be misleading when considering the
performance of children in educational settings in which all children in the same grade or classroom are expected to achieve
equivalent outcomes regardless of differences in their birthdates.
Standard scores are useful, however, as a means of understanding
how the performance of this sample of low-income children compares to a national sample. Thus, Table 1 shows the means and
standard deviations for all assessments, reporting both raw or age
equivalent scores and standard scores where available. It is important to note that the standardized means of our sample on the One
Word, PPVTR, Word Attack, CELFP, CELFR, and WRATR
measures were all below the national average, suggesting that the
overall performance of our sample was lower than that of the
national standardization sample. However, the standard deviation
of our sample on each of these measures was comparable to that of
the national standardization sample, suggesting that the range of
performance in our sample was normal and unrestricted.

Structural Equation Modeling


To map the relations between language skills, code-related
skills, and reading achievement, we tested the explanatory model
depicted in Figure 1 using the Amos 4.0 program (SmallWaters
Corporation; Arbuckle, 1999). The input to the Amos path model
was in the form of raw data. The method of estimation was
maximum likelihood.
The fit of the model in Figure 1 was estimated using various
goodness-of-fit measures: Bentlers comparative fit index (CFI) !
.985 (values of .90 or greater indicate close fit), root mean square
error of approximation (RMSEA) ! .040 (values of .050 or less

938

STORCH AND WHITEHURST

Table 1
Means and Standard Deviations of Yearly Assessment Measures
Preschool

This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

Measure
Oral language
One word
Age equivalent
Standard
PPVT
Age equivalent
Standard
BUS Information
CELFP Basic Concepts
Raw
Standard
CELFR Word Structure
Raw
Standard
Code-related
Naming Letters
Segmenting Sentences
Segmenting Words
Differentiating Print
Knowing Print Functions
Drawing a Person
Printing Name
Knowing Print Components
Identifying Letters/Sounds
Reading
Word Attack
Raw
Standard
SAT Word Reading
WRAT Word Reading
Raw
Standard
SAT Word Skills
SAT Reading Comprehension
Reading accuracy
Word Attack
Raw
Standard
SAT Word Skills
Reading comprehension
SAT Reading Vocabulary
SAT Reading Comprehension

Kindergarten

SD

SD

45.48
82.68

12.89
15.04

64.26
92.51

15.03
14.78

48.10
83.29
16.01

9.50
15.60
7.85

61.20
90.04

12.16
13.69

16.76
9.47

4.02
0.63
2.71
3.87
2.38
2.28
1.85

3.94
1.07
1.39
1.21
1.41
0.86
1.11

Grade 1

Grade 2

Grade 3

Grade 4

SD

SD

SD

SD

72.60
88.64

13.94
15.06

87.63
91.82

16.44
15.73

97.20
90.69

17.30
15.58

114.34
96.82

121.61
15.76

26.22
7.91

5.21
2.16

8.94
94.33
16.62

8.55
14.63
6.20

18.80
97.64
24.53

10.76
15.65
5.91

40.20
94.88

10.82
16.30
25.67
30.75

6.29
8.98
24.03
96.69
33.49

10.74
18.42
7.90

27.56
97.56
31.18

9.95
16.43
8.74

26.65
22.28

9.50
9.33

24.37
30.42

8.32
11.21

1.61
2.95a

8.69
2.04
3.69
5.09

3.45
1.45
1.26
1.07

3.00
3.11

1.16
2.59

Note. Unless otherwise noted, raw scores are presented for all assessments. Where standard scores are available, both raw and standard scores are
presented. For the PPVT and One Word tests, age equivalents (in months) are presented in place of raw scores. PPVT ! Peabody Picture Vocabulary
TestRevised (Dunn & Dunn, 1981); BUS ! The Renfrew Bus StoryAmerican edition (Glasgow & Cowley, 1994); CELFP ! Clinical Evaluation
of Language FundamentalsPreschool (Wiig, Secord, & Semel, 1992); CELFR ! Clinical Evaluation of Language FundamentalsRevised (Semel,
Wiig, & Secord, 1987); Word Attack ! Word Attack subscale of the Woodcock Reading Mastery TestsRevised (Woodcock, 1987); SAT ! Stanford
Achievement TestEighth Edition (The Psychological Corporation, 1989); WRAT ! Wide Range Achievement TestRevised (Jastak & Wilkinson,
1984). Empty cells indicate data not applicable.
a
For the CELF, nationally standardized mean ! 10 (SD ! 3).

indicate close fit), and the discrepancy/degrees of freedom ratio ! 1.978 (ratios of 2 to 1 or less indicate close fit). The CFI of
.985 reported here was somewhat inflated from the result that
would have been obtained had our data set been free of missing
values. The Full Information Maximum Likelihood (FIML)
method Amos uses in the presence of missing data requires the
estimation of means and intercepts, which affects the degrees of
freedom available in the independence or baseline model. Thus, fit

indices that involve the comparison of a hypothetical model to a


baseline model, such as Bentlers CFI, are generally higher when
the sample contains missing data. However, the other fit measures
reported above were not affected by the presence of missing data
and all indicated a very good fit between the model and the sample
data.
Figure 1 illustrates the structural equation model with standardized beta weights. All paths present in the model are significant at

This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

ORAL LANGUAGE AND CODE-RELATED SKILLS

Figure 1. Structural model of literacy development from preschool through fourth grade. PPVT ! Peabody
Picture Vocabulary TestRevised (Dunn & Dunn, 1981); One Word ! One Word Picture Vocabulary Test
(Gardner, 1990); BUS ! The Renfrew Bus StoryAmerican edition (Glasgow & Cowley, 1994); CELF !
Clinical Evaluation of Language Fundamentals (Semel et al., 1987; Wiig et al., 1992); SAT ! Stanford
Achievement TestEighth Edition (The Psychological Corporation, 1989); WRAT ! Wide Range Achievement TestRevised (Jastak & Wilkinson, 1984). Gr. ! Grade; Comprehen. ! Comprehension.

939

This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

940

STORCH AND WHITEHURST

the p " .05 level. To improve the fit of the model, the residual
(error) variances of six pairs of variables were allowed to covary:
preschool DSC Naming Letters kindergarten DSC Naming Letters; preschool phonological awareness kindergarten phonological awareness; preschool One Word kindergarten One Word;
Grade 1 ReadingGrade 2 Word Attack; Grade 1 ReadingGrade
3 4 Reading Accuracy; Grade 2 Word AttackGrade 3 4 Reading
Accuracy. These covariances were suggested by Amoss modification indices and represent measurement error held in common by
some of the assessments we used. Error terms, as well as these six
error covariances, are not present in the model in order to keep the
graphic representation as simple as possible.
Inspection of the direction and relative magnitude of the path
weights in Figure 1 reveals a number of important findings. First,
the relationship between oral language and code-related skills was
quite strong in the preschool years, but it weakened over time. In
preschool, oral language skills predicted 48% of the variance in
code-related skills. In kindergarten, however, the predictive power
of oral language skills was weakened, accounting for less than
10% of the variance in code-related skills. By first and second
grade, the path between oral language skills and reading ability
was not statistically significant (standardized beta to Grade 1
Reading ! .06, to Grade 2 Reading ! .04) and therefore was not
modeled in Figure 1. However, the importance of language skills
re-emerged in Grades 3 and 4, accounting for 7% of the variance
in reading comprehension.
Second, there was longitudinal continuity within both the oral
language and code-related skill domains. For example, within the
oral language domain, approximately 90% of the variance in a
childs oral language ability in kindergarten was accounted for by
preschool ability, 96% of the variance in Grade 12 oral language
was accounted for by kindergarten ability, and 88% of the variance
in Grade 3 4 oral language skills was accounted for by Grade 12
ability. Code-related skills showed similar, though not as strong,
continuity: 38% of the variance in kindergarten code-related skill
ability was accounted for by preschool code-related skill.
Third, reading ability in early elementary school, which was
measured using tasks that reflect and require both reading accuracy
and reading comprehension skills, was directly related to a childs
kindergarten code-related skills. Fifty-eight percent of the variance
in Grade 1 reading ability was accounted for by kindergarten
code-related ability. Second grade reading ability was accounted
for by both Grade 1 reading ability (12% of the variance) and
kindergarten code-related ability (30% of the variance), reinforcing the importance of code-related skills for reading achievement.
Fourth, reading comprehension skills in Grades 3 and 4 were
significantly influenced by three sources: a childs prior reading
achievement (accounting for 18% of the variance), a childs concurrent reading accuracy (accounting for 16% of the variance),
and, importantly, a childs concurrent language skill (accounting
for 7% of the variance). Reading accuracy in Grades 3 and 4 was
heavily determined by prior word reading ability; Grade 2 reading
ability accounted for 56% of Grade 3 and 4 Reading Accuracy.
Missing data. Some children were present at some testing
points but missed assessments at one or more time points, resulting
in a total of 340 children with complete data. Although this may
appear to be a large amount of missing data, it is not unlikely that
an individual child would be unavailable for assessment on 1 of
the 39 variables at some point during the 6 years of the study. As

mentioned above, Amos uses the FIML approach, in which all


observed data values are used to estimate the model, even in the
presence of missing data. Using such an approach to handle
missing data is preferable to both casewise or pairwise deletion,
which can result in large amounts of data being discarded and may
introduce bias to the extent that the completely observed cases
differ from those cases with missing data, and to means imputation, which generally yields biased variance and covariance
estimates.
As a check on the validity of using our entire data set of 626
children, we tested the model in Figure 1 using only those children
who had complete data. The path weights in the model estimated
using the casewise-deleted data set were virtually identical to those
estimated using the full data set, with the standardized betas for the
paths connecting latent variables differing by less than .10. Moreover, the model fit for the casewise-deleted data set (CFI ! .982)
was also virtually identical to that obtained with the full data set
(CFI ! .985).
Indirect effects. Though an independent variable may not directly affect a specific dependent variable (e.g., there is no significant direct path from kindergarten oral language skills to Grade 1
Reading), an independent variable may indirectly influence a dependent variable through an intermediary or mediating variable
(e.g., kindergarten oral language skills indirectly affect Grade 1
Reading through the mediating effects of kindergarten coderelated skills). Estimates of indirect effects are provided by Amos
and represent the sum of the products of all possible direct paths
from the variable of interest to a second variable. In the aforementioned example, the indirect effect of kindergarten oral language
skills on Grade 1 Reading was equal to the direct effect of
kindergarten oral language skills on kindergarten code-related
skills (.31) multiplied by the direct effect of kindergarten coderelated skills on Grade 1 Reading (.76), producing a standardized
indirect effect of .24.
Although there were a number of indirect effects in this model,
our focus was on the indirect influence of early oral language skills
on reading achievement. Oral language skills in Head Start had a
significant indirect effect on kindergarten code-related skills (standardized coefficient for indirect effect ! .72), Grade 1 reading
ability (.55), Grade 2 reading ability (.58), Grade 3 4 Reading
Accuracy (.43), and Grade 3 4 Reading Comprehension (.65).
Oral language skills in kindergarten had a significant, though
somewhat weaker, indirect effect on Grade 1 reading ability (.24),
Grade 2 reading ability (.25), Grade 3 4 Reading Accuracy (.19),
and Grade 3 4 Reading Comprehension (.43). In each case, this
indirect effect was a combination of the relationship between oral
language skills and code-related skills and between code-related
skills and later reading achievement. Thus, these results demonstrate that code-related skills mediate the relationship between
preschool and kindergarten oral language skills and elementary
school reading ability.
Testing alternative models. Though the goodness-of-fit measures we obtained suggest that our model fits the data well, they do
not prove our model. The correlations to which a structural model
is fit may support other models that are also conceptually meaningful. Importantly, there are no conceptually meaningful paths
that could be added to the model in Figure 1 that would be
statistically significant, for example, a path from Grade 12 oral
language skills to Grade 1 Reading, and no conceptually mean-

941

This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

ORAL LANGUAGE AND CODE-RELATED SKILLS

ingful paths that could be removed from the model without worsening fit, for example, the path from Grade 3 4 Reading Accuracy
to Grade 3 4 Reading Comprehension. In this regard, Figure 1
implicitly disconfirms several other models, such as a model in
which oral language skills are directly linked to reading achievement consistently from preschool through fourth grade.
In addition to the implicit testing of alternative models through
the use of Amoss modification indices, we explicitly tested two
alternative models that focus on the changing nature of the reading
process from early to later elementary school. These two alternative models were compared to the model presented in Figure 1
using the traditional fit measures, discrepancy/degrees of freedom
ratio and RMSEA, as well as the Akaike information criterion
(AIC). AIC is a goodness-of-fit measure that can be used to
compare models, where the lower AIC reflects the model with the
better fit. The AIC of the model in Figure 1 was 1,623.108. One
alternative to our model would be to split reading ability in
Grade 2 into a reading accuracy factor and a reading comprehension factor, as we did in the later grades. When doing so, however,
not only did the overall fit of the model decrease (discrepancy/
degrees of freedom ratio ! 2.489; RMSEA ! .066; AIC !
1,969.592), but also Amoss modification indices suggested that
linking the reading comprehension task to the Grade 2 reading
accuracy tasks would improve the fit of the model. Thus, reading
ability at Grade 2 is best conceptualized as a single factor.
A second conceptually plausible alternative to Figure 1 would be to
model Grade 3 4 reading ability as a single factor, as opposed to
splitting the reading tasks into a reading accuracy and a reading
comprehension factor. When Grade 3 4 reading was conceptualized globally in this manner, the overall fit of the model decreased
(discrepancy/degrees of freedom ratio ! 2.339; RMSEA ! .046;
AIC ! 1,868.226). More important was the decrease in the
strength of the path from oral language skills to reading. As shown
in Figure 1, the weight of the path between oral language skills and
reading comprehension in Grades 3 and 4 was .27; in this alternative model the weight of the path, which now leads to a single
Grade 3 4 reading factor, was reduced to .20. Thus, the role of
oral language skills in later elementary school reading achievement was stronger when reading ability was conceptualized as
consisting of two distinct sets of skills, reading accuracy and
reading comprehension, as opposed to when reading ability was
conceptualized globally as a single factor. Furthermore, the effect
of oral language skills on reading ability in later elementary school
largely appeared to be a function of the specific relationship
between oral language skills and reading comprehension.
We also explicitly tested a series of models that focused on the
usefulness and appropriateness of the code-related/oral language
domain distinction. The series of models depicted in Figure 2
represents a systematic investigation of the consequences of conceptualizing the abilities of prereaders as consisting of two distinct
domains. This model series began with the simplest conception of
literacy development (see Model A), a model in which skills were
grouped solely on the basis of educational level; that is, all skills
assessed at each age (preschool, kindergarten, Grade 1, etc.) were
considered to represent a single emergent literacy factor. This
model supposed no distinction between the various oral language
and code-related skills; rather, it assumed that the range of skills
assessed at any particular grade makes a global contribution to a
childs skills at the next grade. Such a conceptualization of literacy

development does not fit the data from our sample of low-income
children well, however, as reflected in the fit measures (discrepancy/degrees of freedom ratio ! 3.884; RMSEA ! .068; AIC
! 2,924.580).
Models B, C, and D represented a systematic departure from the
global notion of literacy development as portrayed in Model A to
one in which literacy development was considered to be a function
of two domains: the oral language and code-related domains.
Model B began this departure by separating out the oral language
skills and by dividing reading ability in Grades 3 and 4 into a
reading comprehension and a reading accuracy factor (a decision
based on results cited above regarding the appropriateness of a
distinction between these two skills in more mature readers).
Model C followed suit by extending this distinction to Grades 1
and 2, in which measures tapping reading ability (which have been
kept together on the basis of the results cited above regarding the
high degree of association between reading accuracy and reading
comprehension tasks at this time) were separated from oral language skills. Importantly, Model B (discrepancy/degrees of freedom ratio ! 3.623; RMSEA ! .062; AIC ! 2,743.121) represented an improvement over Model A. Similarly, Model C
(discrepancy/degrees of freedom ratio ! 2.397; RMSEA ! .047;
AIC ! 1,895.504) represented an improvement over Model B.
Thus, as we moved toward conceptualizing reading ability as
consisting of two distinct domains, we improved the fit of our
model. In Model D we extended this distinction to emergent
literacy skills in the kindergarten year, in which we were no longer
assessing reading ability but precursors to reading ability. Thus,
we divided the many skills assessed at this age into a code-related
(print principles and phonological awareness) and an oral language
(vocabulary and conceptual knowledge) factor. The result was a
likewise improvement in model fit (discrepancy/degrees of freedom ratio ! 2.181; RMSEA ! .043; AIC ! 1,760.732).
Finally, Model E represented the last step in this series in which
the two-domain notion of literacy skill was extended back to our
earliest time point, the preschool period. Not only did the modifications made in Model E indicate further improvement in fit over
Models A through D (discrepancy/degrees of freedom ratio ! 1.978; RMSEA ! .040; AIC ! 1,623.108), but also the fit
measures obtained in this final model met the strictest requirements for fit (i.e., discrepancy/degrees of freedom ratio was below 2 to 1, RMSEA was below .050, as were the upper and lower
bounds of RMSEA).

Discussion
Our structural model of language and literacy development
provides a basis for both predicting and understanding the influence of the oral language and code-related domains on later
reading achievement. Importantly, our testing of alternative models provides empirical support for the appropriateness and value of
conceptualizing emergent and conventional literacy as consisting
of two distinct domains. Our results demonstrate the strong relationship between the two domains of emergent literacy skills
during the preschool period, which is consistent with studies that
have found significant correlations between oral language skills
(i.e., vocabulary) and code-related skills (i.e., phonological awareness) in very young children (e.g., Burgess & Lonigan, 1998;
Chaney, 1992; Lonigan, Burgess, Anthony, & Barker, 1998). As

Figure 2. Model series evaluating the effects of a two-domain conceptualization of literacy development. Note that the fit measures improve steadily from
Model A, a single domain conceptualization, to Model E, our complete, two-domain conceptualization of emergent and conventional literacy.

This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

942
STORCH AND WHITEHURST

This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

ORAL LANGUAGE AND CODE-RELATED SKILLS

children begin formal schooling, the relationship between the oral


language and code-related domains diminishes. In Grades 1 and 2
the relationship between oral language and reading ability is nonsignificant, a finding consistent with recent work by Evans, Shaw,
and Bell (2000) and Senechal and LeFevre (2002). During this
time, code-related skills maintain a strong and direct influence on
reading achievement. Kindergarten print knowledge and phonological awareness play a critical role in the success or failure
children have in beginning to learn to read. In Grade 2, though the
tasks become more complex, reading is still heavily determined by
the code-related skills children acquired by the end of their kindergarten year. Furthermore, reading comprehension at this stage
is highly correlated with word and nonword reading tasks, reinforcing the position that at least during the early stages of reading
development, reading comprehension is primarily a function of
word reading abilities (e.g., Hoover & Gough, 1990; Shankweiler
et al., 1995). By Grades 3 and 4, the pattern of influence changes
significantly as reading accuracy and reading comprehension tasks
can be reliably separated
Our findings provide additional support for previous research
demonstrating that in the early stages of reading development,
reading accuracy and reading comprehension skills appear to be
closely intertwined, and the relationship between word recognition
and reading comprehension ability is strong (e.g., Hoover &
Gough, 1990; Juel, Griffith, & Gough, 1986; Shankweiler et al.,
1995, 1999; Tunmer & Nesdale, 1985). In early readers, deficiency
in mapping the relationship between the alphabetic or orthographic
representation of words and their spoken counterparts presents the
chief barrier to comprehension (Shankweiler et al., 1999). According to a review conducted by Biemiller (1994), until children
become efficient at translating print into spoken language, inefficiency in decoding print interferes with the ability to understand
print. This research on decoding efficiency, as measured by reading speed, finds that for rapid readers the gap between general
comprehension and reading comprehension is minimal. However,
for children reading at a slower rate, the gap between general and
reading comprehension in much larger (Biemiller, 1994).
In older children there are discrepancies of varying degrees
between decoding and comprehension ability (Shankweiler et al.,
1999), suggesting that these skills, though clearly related, are not
one in the same. Deciphering all of the words in a passage does not
automatically lead to understanding of the meaning of that passage. Researchers have identified subsets of readers with ageappropriate decoding skills, yet poor reading comprehension, as
well as readers with adequate reading comprehension, despite poor
decoding skills (e.g., Nation & Snowling, 1998; Shankweiler et al.,
1999; Stothard & Hulme, 1995). Shankweiler et al. (1999) found
that these two patterns were not equally likely; readers with good
comprehension, despite difficulties in decoding words out of context and nonwords, were twice as likely in their sample as the
reverse pattern. Often, children with normally developed decoding
skills but poor reading comprehension tend to exhibit comprehension difficulties in a wide variety of areas, not strictly related to
reading. Many suggest that this deficiency in reading comprehension is indicative of a more general difficulty with language
comprehension (e.g., Stothard & Hulme, 1992). For example,
Bishop and Adams (1990) found that among children identified as
having specific language impairment at age 512, the poor reading
skill they showed at age 812 was not an isolated impairment, but

943

occurred along with pervasive verbal deficits in the understanding


and expression of spoken language. Importantly, many of the
children in this group were able to decode words accurately, yet
had poor comprehension of what they read.
According to our model, reading accuracy in this later stage is
heavily influenced by prior word recognition and decoding abilities. Reading comprehension, on the other hand, is determined by
multiple sources: prior reading ability and both concurrent reading
accuracy and concurrent language ability. This model supports the
view that sentence and text comprehension are affected by a
childs general verbal ability and oral language skills (Snow et al.,
1991; Whitehurst, 1997). Furthermore, the model demonstrates
that oral language abilities re-emerge as a strong, direct force later
in the sequence of learning to read, a relationship that may have
gone unnoticed in previous studies in which reading accuracy and
reading comprehension tasks were considered together or studies
in which reading achievement beyond Grade 2 was not investigated.
Importantly, our model demonstrates that the relationship between oral language and reading skill in the early stages of reading
development is mediated by code-related skills, such as phonological processing and print concepts. Thus, although oral language
abilities do not appear to make a direct contribution to reading
during Grades 1 and 2, a childs skill with spoken language does
play an essential, albeit an indirect role in reading achievement
during the early stages of reading acquisition. This is a significant
finding in light of the many research studies that have not found
support for a connection between oral language ability and early
reading achievement, beyond the influence of phonemic awareness
skills. The results of our model suggest that there may be a danger
in emphasizing phonological processing skills to the extent that the
role of other language skills is underestimated. Although phonological processing skills play a more visible, direct role in early
reading achievement, these skills are determined, in part, by a
childs oral language ability. This finding expands upon existing
correlational research demonstrating a significant relationship between childrens vocabulary skills and their phonological sensitivity in preschool (e.g., Burgess & Lonigan, 1998; Chaney, 1992;
Lonigan et al., 1998) and early elementary school (e.g., Bowey,
1994; Wagner, Torgesen, Laughon, Simmons, & Rashotte, 1993;
Wagner et al., 1997) by mapping the nature of this relationship
over time.
Taken together, these findings from the model provide empirical
support for the view that code-related skills and oral-language
skills play their most significant role at different points during the
development of reading ability. Heretofore, the theory that different literacy-related abilities make their most important contributions to reading achievement at different times has been based on
the results of separate studies assessing different sets of literacy
abilities, often representing only one emergent literacy domain, at
different times during the acquisition of reading skill. Importantly,
this study addressed the influence of skills falling into both the
code-related and oral language domains over a continuous 6-year
period. Moreover, this study demonstrates that the distinction
between code-related and oral language skills is not merely an
arbitrary, or even theoretical distinction, but rather one that is
supported empirically and allows for the recognition of the differential effects of these two domains on later reading achievement.

944

STORCH AND WHITEHURST

This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

Implications and Recommendations


Findings from the model hold powerful implications for children from low-income backgrounds. A number of studies have
documented social class differences in childrens exposure to the
particular literacy experiences shown to be important for the
development of language and emergent literacy skills. Research
has demonstrated social class differences in both book ownership
(e.g., Feitelson & Goldstein, 1986; McCormick & Mason, 1986;
Raz & Bryant, 1990; Teale, 1986) and the frequency of shared
book reading (e.g., Adams, 1990; Feitelson & Goldstein, 1986;
Harris & Smith, 1987; Raz & Bryant, 1990).
Children whose home literacy environments are lacking with
regard to shared reading activities and print materials are likely to
have poor oral language skills (Storch & Whitehurst, 2001). The
strong relationship between preschool oral language and coderelated skills and the strong continuity between code-related skills
and later reading achievement places children with early deficits at
risk for later reading difficulties. Indeed, research has demonstrated that children from low-income backgrounds begin formal
schooling behind their peers with regard to language ability
(Whitehurst, 1997), phonological sensitivity (e.g., Bowey, 1995;
Dickinson & Snow, 1987; Raz & Bryant, 1990), and knowledge of
print concepts (Smith & Dixon, 1995). Not surprisingly, children
from low-income backgrounds are also at risk for later reading
difficulties (Dubow & Ippolito, 1994; National Center for Education Statistics, 2001; White, 1982).
Thus, this model holds particular implications for the role of
early school experiences in childrens later reading achievement.
Research suggests that more formal interactions with print, such as
teaching a child about letters, influence those skills falling into the
code-related domain (e.g., letter knowledge, phonological awareness; Evans et al., 2000; Senechal & LeFevre, 2002; Senechal,
LeFevre, Thomas, & Daley, 1998; Whitehurst et al., 1994). Importantly, preschool code-related skills account for 38% of the
variance in kindergarten code-related skills, which are significantly predictive of both Grade 1 and Grade 2 reading outcomes.
These findings point to the necessity of engaging in teaching
activities during the preschool period that will support the development of code-related skills. Furthermore, attention to problems
in both domains by preschool and kindergarten teachers should
help to identify those children most likely to have later reading
problems before they begin formal reading instruction.
This model also holds implications for the remediation of older
children who are already demonstrating reading problems. To help
children falling behind in the early elementary school grades, when
reading is primarily a decoding task, we should concentrate on
boosting word recognition abilities through phonological training.
As reading tasks begin to move beyond decoding to involve
reading comprehension, the potential for other factors, particularly
oral language skill, to influence reading ability increases. Remediation studies demonstrate, for example, that whereas phonological training results in growth in phonological skills it does not lead
to unique growth in reading comprehension over reading training
encompassing similar amounts of time yet less explicit instruction
in phonological skills (e.g., Torgesen, Wagner, & Rashotte, 1997;
Wise, Ring, & Olson, 2000). Moreover, similar studies have
shown that younger, lower level readers benefit more from phonological training than do older, higher level readers (e.g., Wise et

al., 2000). Thus, attempts to remediate children in the later elementary school grades who are performing poorly at reading must
be sensitive to the various factors that may be contributing to their
difficulties.
Importantly, we must be careful not to focus on promoting
decoding skills to the exclusion of comprehension skills, even in
those readers with decoding difficulties. Though improving coderelated skills, such as phonological awareness and print knowledge, may necessarily be a focus of intervention in those children
who have not yet acquired sufficient skill in reading words, we
must not wait until children have solved the decoding puzzle to
begin instruction in oral language skills, such as vocabulary and
syntax. These oral language skills should be an integral part of
reading instruction beginning in preschool and throughout elementary school. Not only are oral language abilities linked to the
code-related skills that promote word-reading abilities, but early
oral language abilities also provide the foundation for development
of the advanced oral language skills necessary for successful
comprehension in more skilled readers.
Results of research on children with early language impairments
fit well with the findings of the present study. For example, recent
follow-up studies of children with a preschool history of language
impairment show that the nature of their reading problems changes
over time to encompass difficulties in both decoding and comprehension (e.g., Snowling, Bishop, & Stothard, 2000; Stothard,
Snowling, Bishop, Chipchase, & Kaplan, 1998). Such findings
support the far-reaching impact of oral language ability on reading
achievement modeled in the present study. Moreover, Snowling et
al. (2000) argued that the development of word recognition in
children with early history of language impairment is dependent
upon normal levels of semantic knowledge. Snowling et al.
pointed out that good decoding skills alone will not assure normal
reading progress in children with a history of language impairment; oral language skills, such as semantic and syntactic skills,
make an important contribution to literacy development as reading
tasks become more complex.
Despite the significant findings, limitations of this study must be
mentioned. Not all code-related and oral language skills were
measured. In addition to phonological sensitivity, measured in this
study by the ability to segment words and sentences and rhyming
skill, measures of phonological processing including phonological
memory and phonological naming have also been identified as
significant correlates of reading skill (e.g., Bowers & Wolfe, 1993;
Wagner et al., 1994, 1997; Wolfe, 1991). Our oral language
measures focused mainly on receptive and expressive vocabulary,
with measures of narrative skills, basic concepts, and word structure each assessed at a single time point. It is possible that the
inclusion of other oral language measures, such as tests for auditory comprehension of language and grammatical sensitivity, may
impact the strength of influence of the oral language domain on
reading achievement.
Furthermore, although our results emphasize the continuities
and discontinuities in childrens progression from prereaders to
readers, they do not address the question of the origins of coderelated and oral language skills. Future studies should address
questions concerning the developmental origins of skills such as
phonological sensitivity, letter knowledge, and oral language. Such
knowledge would provide a basis for the development of interven-

This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

ORAL LANGUAGE AND CODE-RELATED SKILLS

tions aimed at preventing later reading difficulties in populations


of children at risk.
It is also essential to emphasize that the model presented in this
article is based on data from low-income children. The performance of our sample on nationally standardized measures such as
the PPVTR and Word Attack suggest that these children are
consistently performing below the national average. Thus, we must
be cautious in generalizing the results of this model to middle- and
upper-class children. Although it is possible that the precursors to
reading achievement differ in children from varying socioeconomic backgrounds, research has demonstrated similar predictive
patterns of influence in middle-class and mixed income populations (e.g., Burgess & Lonigan, 1998; Torgesen & Burgess, 1998;
Wagner et al., 1993, 1994, 1997).
In summary, individual differences in code-related and oral
language skills are causally and powerfully connected to individual differences in reading achievement in elementary school. The
implication of this developmental model for social and educational
policy is that investing resources to improve both code-related and
oral language skills may enhance reading achievement in children.
Moreover, we must not wait until children begin formal reading
instruction to identify and help those who are at risk for reading
difficulties. Future efforts to prevent reading problems need to be
sensitive to the developmental relationships between code-related
skills, oral language, and reading achievement, and the changing
nature of reading during the course of the elementary school years.

References
Adams, M. J. (1990). Learning to read: Thinking and learning about print.
Cambridge, MA: MIT Press.
Arbuckle, J. L. (1999). Amos 4.0 [Computer software]. Small Waters
Corporation.
Beck, I., & Juel, C. (1999). The role of decoding in learning to read. In
Consortium on Reading Excellence (Eds.), Reading research anthology:
The why of reading instruction (pp. 78 87). Novato, CA: Arena Press.
Biemiller, A. (1994). Some observations on acquiring and using reading
skill in elementary schools. In C. K. Kinzer & D. J. Leu (Eds.),
Multidimensional aspects of literacy research, theory, and practice:
Forty-third yearbook of the National Reading Conference (pp. 209
216). Chicago: National Reading Conference.
Bishop, D. V. M. (1991). Developmental reading disabilities: The role of
phonological processing has been overemphasized. Mind & Language, 6, 97101.
Bishop, D. V. M., & Adams, C. (1990). A prospective study of the
relationship between specific language impairment, phonological disorders and reading retardation. Journal of Child Psychology and Psychiatry and Allied Disciplines, 31, 10271050.
Bowers, P. G., & Wolfe, M. (1993). Theoretical links among naming
speed, precise timing mechanisms and orthographic skill in dyslexia.
Reading and Writing, 5, 69 85.
Bowey, J. A. (1986). Syntactic awareness in relation to reading skills and
on-going reading comprehension monitoring. Journal of Experimental
Child Psychology, 41, 282299.
Bowey, J. A. (1994). Phonological sensitivity in novice readers and nonreaders. Journal of Experimental Child Psychology, 58, 134 159.
Bowey, J. A. (1995). Socioeconomic status differences in preschool phonological sensitivity and first-grade reading achievement. Journal of
Educational Psychology, 87, 476 487.
Bradley, L., & Bryant, P. (1985). Rhyme and reason in reading and
spelling. Ann Arbor, MI: University of Michigan Press.

945

Bryant, P., MacLean, M., & Bradley, L. (1990). Rhyme, language, and
childrens reading. Applied Psycholinguistics, 11, 237252.
Burgess, S. R., & Lonigan, C. J. (1998). Bidirectional relations of phonological sensitivity and prereading abilities: Evidence from a preschool
sample. Journal of Experimental Child Psychology, 70, 117141.
Butler, S. R., Marsh, H. W., Sheppard, M. J., & Sheppard, J. L. (1985).
Seven-year longitudinal study of the early prediction of reading achievement. Journal of Educational Psychology, 77, 349 361.
Byrne, B., & Fielding-Barnsley, R. F. (1993). Evaluation of a program to
teach phonemic awareness to young children: A one year follow-up.
Journal of Educational Psychology, 85, 104 111.
Byrne, B., & Fielding-Barnsley, R. F. (1995). Evaluation of a program to
teach phonemic awareness to young children: A 2- and 3-year follow-up
and a new preschool trial. Journal of Educational Psychology, 87,
488 503.
Catts, H. W. (1993). The relationship between speech-language impairments and reading disabilities. Journal of Speech and Hearing Research, 36, 948 958.
Catts, H. W., Fey, M. E., Zhang, X., & Tomblin, J. B. (1999). Language
basis of reading and reading disabilities: Evidence from a longitudinal
investigation. Scientific Studies of Reading, 3, 331361.
Chaney, C. (1992). Language development, metalinguistic skills, and print
awareness in 3-year-old children. Applied Psycholinguistics, 13, 485
514.
CTB. (1990). Developing Skills Checklist. Monterey, CA: McGraw-Hill.
Demont, E., & Gombert, J. E. (1996). Phonological awareness as a predictor of recoding skills and syntactic awareness as a predictor of
comprehension skills. British Journal of Psychology, 66, 315332.
Dickinson, D. K., & Snow, C. E. (1987). Interrelationships among prereading and oral language skills in kindergartners from two social
classes. Early Childhood Research Quarterly, 2, 125.
Dubow, E. F., & Ippolito, M. F. (1994). Effects of poverty and quality of
the home environment on changes in the academic and behavioral
adjustment of elementary school-age children. Journal of Clinical Child
Psychology, 23, 401 412.
Dunn, L. M., & Dunn, L. M. (1981). Peabody Picture Vocabulary Test
Revised. Circle Pines, MN: American Guidance Service.
Evans, M. A., Shaw, D., & Bell, M. (2000). Home literacy activities and
their influence on early literacy skills. Canadian Journal of Experimental Psychology, 54, 6575.
Feagans, L., & Applebaum, M. (1986). Validation of language subtypes in
learning disabled children. Journal of Educational Psychology, 78, 308
364.
Feitelson, D., & Goldstein, Z. (1986). Patterns of book ownership and
reading to young children in Israeli school-oriented and nonschooloriented families. Reading Teacher, 39, 924 930.
Gardner, M. F. (1990). Expressive One-Word Picture Vocabulary Test
Revised. Novato, CA: Academic Therapy.
Gillon, G., & Dodd, B. J. (1994). A prospective study of the relationship
between phonological, semantic and syntactic skills and specific reading
disability. Reading and Writing, 6, 321345.
Glasgow, C., & Cowley, J. (1994). The Renfrew Bus StoryAmerican
edition. Centreville, DE: The Centreville School.
Harris, M. M., & Smith, N. J. (1987). Literacy assessment of chapter 1 and
non-chapter 1 homes. Reading Improvement, 24, 137142.
Hoover, W., & Gough, P. B. (1990). The simple view of reading. Reading
and Writing, 2, 127160.
Jastak, S., & Wilkinson, G. S. (1984). Wide Range Achievement Test
Revised. Wilmington, DE: Jastak Associates.
Juel, C., Griffith, P. L., & Gough, P. B. (1986). Acquisition of literacy: A
longitudinal study of children in first and second grade. Journal of
Educational Psychology, 78, 243255.
Klecan-Aker, J. S., & Caraway, T. H. (1997). A study of the relationship
of storytelling ability and reading comprehension in fourth and sixth

This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

946

STORCH AND WHITEHURST

grade African-American children. European Journal of Disorders of


Communication, 32, 109 125.
Lonigan, C. J., Burgess, S. R., & Anthony, J. L. (2000). Development of
emergent literacy and early reading skills in preschool children: Evidence from a latent-variable longitudinal study. Developmental Psychology, 36, 596 613.
Lonigan, C. J., Burgess, S. R., Anthony, J. L., & Barker, T. A. (1998).
Development of phonological sensitivity in two- to five-year-old children. Journal of Educational Psychology, 90, 294 311.
Mason, J. M. (1992). Reading stories to preliterate children: A proposed
connection to reading. In P. B. Gough, L. C. Ehri, & R. Treiman (Eds.),
Reading acquisition (pp. 215243). Hillsdale, NJ: Erlbaum.
McCormick, C. E., & Mason, J. M. (1986). Intervention procedures for
increasing preschool childrens interest in and knowledge about reading.
In W. H. Teale & E. Sulzby (Eds.), Emergent literacy: Writing and
reading (pp. 90 115). Norwood, NJ: Ablex.
Menyuk, P., Chesnick, M., Liebergott, J., Korngold, B., DAgostino, R., &
Belanger, A. (1991). Predicting reading problems in at-risk children.
Journal of Speech and Hearing Research, 34, 893903.
Nation, K., & Snowling, M. J. (1998). Semantic processing and the
development of word-recognition skills: Evidence from children with
reading comprehension difficulties. Journal of Memory and Language, 39, 85101.
National Center for Education Statistics. (2001). The nations report card:
Fourth-grade reading 2000. Retrieved April 6, 2001, from http://nces
.ed.gov/nationsreportcard/reading/results
Perfetti, C. A. (1985). Reading ability. New York: Oxford University
Press.
Pikulski, J. J., & Tobin, A. W. (1989). Factors associated with long-term
reading achievement of early readers. In S. McCormick, J. Zutell, P.
Scharer, & P. OKeefe (Eds.), Cognitive and social perspectives for
literacy research and instruction (pp. 123133). Chicago: National
Reading Conference.
Psychological Corporation. (1989). The Stanford Achievement Test: Eighth
edition. Orlando, FL: Harcourt Brace Jovanovich.
Raz, I. S., & Bryant, P. (1990). Social background, phonological awareness
and childrens reading. British Journal of Developmental Psychology, 8,
209 225.
Richman, N., Stevenson, J., & Graham, P. J. (1982). Pre-school to school:
A behavioural study. In H. R. Schaffer (Ed.), Behavioural development:
A series of monographs. London: Academic Press.
Roth, F. P., Speece, D. L., Cooper, D. H., & de la Paz, S. (1996).
Unresolved mysteries: How do metalinguistic and narrative skills connect with early reading? The Journal of Special Education, 30, 257277.
Roth, F. P., & Spekman, N. J. (1994). Oral story production abilities of
adults with learning disabilities. In R. L. Bloom, L. K. Obler, S. DeSanti,
& J. Ehrlich (Eds.), Discourse in adult clinical populations (pp. 131
148). Hillsdale, NJ: Erlbaum.
Scarborough, H. (1989). Prediction of reading dysfunction from familial
and individual differences. Journal of Educational Psychology, 81,
101108.
Semel, E., Wiig, E. H., & Secord, W. (1987). CELFR: Clinical Evaluation of Language FundamentalsRevised. New York: The Psychological Corporation.
Senechal, M., & LeFevre, J. (2002). Parental involvement in the development of childrens reading skill: A five-year longitudinal study. Child
Development, 73, 445 460.
Senechal, M., LeFevre, J., Thomas, E. M., & Daley, K. E. (1998). Differential effects of home literacy experiences on the development of oral
and written language. Reading Research Quarterly, 33, 96 116.
Shankweiler, D., Crain, S., Katz, L., Fowler, A. E., Liberman, A. M.,
Brady, S. A., et al. (1995). Cognitive profiles of reading-disabled children: Comparison of language skills in phonology, morphology, and
syntax. Psychological Science, 6, 149 156.

Shankweiler, D., Lundquist, E., Katz, L., Stuebing, K. K., Fletcher, J. M.,
Brady, S. A., et al. (1999). Comprehension and decoding patterns of
association in children with reading difficulties. Scientific Studies of
Reading, 3, 69 94.
Share, D. L., Jorm, A. F., MacLean, R., & Mathews, R. (1984). Sources of
individual differences in reading acquisition. Journal of Educational
Psychology, 76, 1309 1324.
Share, D. L., & Silva, P. (1987). Language deficits and specific reading
retardation: Cause or effect? British Journal of Disorders of Communication, 22, 219 226.
Silva, P. A., McGee, R., & Williams, S. M. (1985). Some characteristics of
9-year-old boys with general reading backwardness or specific reading
retardation. Journal of Child Psychology and Psychiatry and Allied
Disciplines, 26, 407 421.
Silva, P. A., Williams, S., & McGee, R. (1987). A longitudinal study of
children with developmental language delay at age three: Later intelligence, reading and behaviour problems. Developmental Medicine &
Child Neurology, 29, 630 640.
Smith, S. S., & Dixon, R. G. (1995). Literacy concepts of low- and
middle-class four-year-olds entering preschool. Journal of Educational
Research, 88, 243253.
Snow, C. E., Barnes, W. S., Chandler, J., Hemphill, L., & Goodman, I. F.
(1991). Unfulfilled expectations: Home and school influences on literacy. Cambridge, MA: Harvard University Press.
Snowling, M., Bishop, D., & Stothard, S. (2000). Is preschool language
impairment a risk factor for dyslexia in adolescence? Journal of Child
Psychology and Psychiatry and Allied Disciplines, 41, 587 600.
Snyder, L. S., & Downey, D. M. (1991). The languagereading relationship in normal and reading disabled children. Journal of Speech and
Hearing Research, 34, 129 140.
Speece, D. L., Roth, F. A., Cooper, D. H., & de la Paz, S. (1999). The
relevance of oral language skills to early literacy: A multivariate analysis. Applied Psycholinguistics, 20, 167190.
Storch, S. A., & Whitehurst, G. J. (2000). The role and structure of
emergent literacy skills in literacy outcomes of Head Start children. In F.
Lamb-Parker, J. Hagan, and R. Robinson (Eds.), Summary of conference
proceedings: Head Starts Fifth National Research Conference, Washington, DC (pp. 116 117). New York: Columbia University Press.
Storch, S. A., & Whitehurst, G. J. (2001). The role of family and home in
the literacy development of children from low-income backgrounds. In
W. Damon (Series Ed.) & P. R. Britto & J. Brooks-Gunn (Vol. Eds.),
New directions for child and adolescent development (Vol. 92, pp.
5371). San Francisco: Jossey-Bass.
Stothard, S. E., & Hulme, C. (1992). Reading comprehension difficulties in
children: The role of language comprehension and working memory
skills. Reading and Writing, 4, 245256.
Stothard, S. E., & Hulme, C. (1995). Comparison of phonological skills in
children with reading comprehension difficulties and children with decoding difficulties. Journal of Child Psychology and Psychiatry, 36,
399 408.
Stothard, S. E., Snowling, M., Bishop, D., Chipchase, M., & Kaplan, C.
(1998). Language-impaired preschoolers: A follow-up into adolescence.
Journal of Speech, Language, and Hearing Research, 41, 407 418.
Sulzby, E., & Teale, W. (1991). Emergent literacy. In R. Barr, M. Kamil,
P. Mosenthal, & P. D. Pearson (Eds.), Handbook of reading research
(Vol. 2, pp. 727758). New York: Longman.
Teale, W. H. (1986). Home background and young childrens literacy
development. In W. H. Teale & E. Sulzby (Eds.), Emergent literacy:
Writing and reading (pp. 110 121). Norwood, NJ: Ablex.
Teale, W. H., & Sulzby, E. (Eds.). (1986). Emergent literacy: Writing and
reading. Norwood, NJ: Ablex.
Torgesen, J. K., & Burgess, S. R. (1998). Consistency of reading-related
phonological processes throughout early childhood: Evidence from
longitudinal-correlational and instructional studies. In J. L. Metsala &

This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

ORAL LANGUAGE AND CODE-RELATED SKILLS


L. C. Ehri (Eds.), Word recognition in beginning literacy (pp. 161188).
Mahwah, NJ: Erlbaum.
Torgesen, J. K., Wagner, R. K., & Rashotte, C. A. (1997). Prevention and
remediation of severe reading disabilities: Keeping the end in mind.
Scientific Studies of Reading, 1, 217234.
Tunmer, W. E., Herriman, M. L., & Nesdale, A. R. (1988). Metalinguistic
abilities and beginning reading. Reading Research Quarterly, 23, 134
158.
Tunmer, W. E., & Nesdale, A. R. (1985). Phonemic segmentation and
beginning reading. Journal of Educational Psychology, 77, 418 427.
U.S. Department of Health and Human Services. (1999). 1999 Head Start
fact sheet. Retrieved April 6, 2001, from http://www2.acf.dhhs.gov/
programs/hsb/research/99_hsfs.htm
Van Kleeck, A. (1998). Preliteracy domains and stages: Laying the foundations for beginning reading. Journal of Childrens Communication
Development, 20, 3351.
Vellutino, F. R., & Scanlon, D. M. (1985). Free recall of concrete and
abstract words in poor and normal readers. Journal of Experimental
Child Psychology, 39, 363380.
Vellutino, F. R., Scanlon, D. M., Sipay, E. R., Small, S. G., Pratt, A., Chen,
R., et al. (1996). Cognitive profiles of difficult-to-remediate and readily
remediated poor readers: Early intervention as a vehicle for distinguishing between cognitive and experiential deficits as basic causes of specific reading disability. Journal of Educational Psychology, 88, 601
638.
Vellutino, F. R., Scanlon, D. M., & Spearing, D. (1995). Semantic and
phonological coding in poor and normal readers. Journal of Experimental Child Psychology, 59, 76 123.
Vellutino, F. R., Scanlon, D. M., & Tanzman, M. S. (1991). Bridging the
gap between cognitive and neuropsychological conceptualizations of
reading disability. Learning and Individual Differences, 3, 181203.
Wagner, R. K., Torgesen, J. K., Laughon, P., Simmons, K., & Rashotte,
C. A. (1993). Development of young readers phonological processing
abilities. Journal of Educational Psychology, 85, 83103.
Wagner, R. K., Torgesen, J. K., & Rashotte, C. A. (1994). The development of reading-related phonological processing abilities: New evidence
of bi-directional causality from a latent variable longitudinal study.
Developmental Psychology, 30, 73 87.
Wagner, R. K., Torgesen, J. K., Rashotte, C. A., Hecht, S. A., Barker,
T. A., Burgess, S. R., et al. (1997). Changing relations between phono-

947

logical processing abilities and word-level reading as children develop


from beginning to skilled readers: A 5-year longitudinal study. Developmental Psychology, 33, 468 479.
White, K. (1982). The relation between socioeconomic status and academic
achievement. Psychological Bulletin, 91, 461 481.
Whitehurst, G. J. (1997). Language processes in context: Language learning in children reared in poverty. In L. B. Adamson & M. A. Romski
(Eds.), Research on communication and language disorders: Contribution to theories of language development (pp. 233266). Baltimore:
Brookes.
Whitehurst, G. J., Arnold, D. H., Epstein, J. N., Angell, A. L., Smith, M.,
& Fischel, J. E. (1994). A picture book reading intervention in daycare
and home for children from low-income families. Developmental Psychology, 30, 679 689.
Whitehurst, G. J., Falco, F., Lonigan, C. J., Fischel, J. E., DeBaryshe,
B. D., Valdez-Menchaca, M. C., et al. (1988). Accelerating language
development through picture-book reading. Developmental Psychology, 24, 552558.
Whitehurst, G. J., & Fischel, J. E. (2000). A developmental model of
reading and language impairments arising in conditions of economic
poverty. In D. Bishop & L. Leonard (Eds.), Speech and language
impairments in children: Causes, characteristics, intervention and outcome (pp. 5371). East Sussex, England: Psychology Press.
Whitehurst, G. J., & Lonigan, C. J. (1998). Child development and emergent literacy. Child Development, 69, 848 872.
Wiig, E. H., Secord, W., & Semel, E. (1992). CELFPreschool. Clinical
Evaluation of Language FundamentalsPreschool. New York: Psychological Corporation.
Wise, B., Ring, J., & Olson, R. K. (2000). Individual differences in gains
from computer-assisted remedial reading. Journal of Experimental Child
Psychology, 77, 197235.
Wolfe, M. (1991). Naming speed and reading: The contribution of the
cognitive neurosciences. Reading Research Quarterly, 26, 123141.
Woodcock, R. W. (1987). Woodcock Reading Mastery TestsRevised.
Circles Pines, MN: American Guidance Service.

Received July 24, 2001


Revision received May 28, 2002
Accepted May 29, 2002 !

You might also like