The relation between reading and writing development in english and esl students

THE RELATION BETWEEN READING AND WRITING DEVELOPMENT IN ENGLISH AND ESL STUDENTS

By

Susan Elizabeth Ball

A thesis submitted in conformity with the requirements For the degree of Doctor of Philosophy

Graduate Department of Education University of Toronto

© Copyright by Susan Elizabeth Ball 2003

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.

1*1 National Library of Canada

Acquisitions and Bibliographic Services395 Wellington Street Ottawa ON K1A0N4 Canada

Bibliotheque nationale du Canada

Acquisitions et sen/ices bibliographiques395, rue Wellington Ottawa ON K1A0N4 Canada

Your file Votre reference

Our tile Notre reference

The author has granted a nonexclusive licence allowing the National Library of Canada to reproduce, loan, distribute or sell copies of this thesis in microform, paper or electronic formats.

The author retains ownership of the copyright in this thesis. Neither the thesis nor substantial extracts from it may be printed or otherwise reproduced without the author’s permission.

L’auteur a accorde une licence non exclusive permettant a la Bibliotheque nationale du Canada de reproduire, preter, distribuer ou vendre des copies de cette these sous la forme de microfiche/film, de reproduction sur papier ou sur format electronique.

L’auteur conserve la propriete du droit d’auteur qui protege cette these. Ni la these ni des extraits substantiels de celle-ci ne doivent etre imprimes ou autrement reproduits sans son autorisation.

0 - 612 - 78365-0

CanadaReproduced with permission of the copyright owner. Further reproduction prohibited without permission.

ABSTRACT

The Relation between Reading and Writing Development in English and ESL Students

Susan Elizabeth Ball

Graduate Department of Education, University of Toronto

Surprisingly little is known about similarities and differences in how reading and writing

skills develop. Most attempts to examine the relation between reading and writing

development have used few measures at a single point in time with small samples of

children. The current study explored the relation between the lower (decoding and

spelling) and higher levels (reading comprehension and story construction) of reading

and writing development for students in Grades 3 and 5/6, and also explored many

underlying cognitive and language processes that are predictive of reading and writing

achievement. In addition, the present study addressed these issues as they relate to

students for whom English is either their first (ENG) or second (ESL) language. A series

of ANCOVAs revealed few significant differences between ENG and ESL students on

reading, writing and cognitive measures; however, significant differences were found on

oral language proficiency (OLP) measures in favor of ENG students. Multiple sequential

regression was undertaken to investigate the underlying cognitive and language processes

that predict reading and writing development. The cognitive ability composite predicted

the most variance in lower-level decoding and spelling skills, and the OLP composite

predicted the most variance in the higher-level reading comprehension and story

construction skills. These results have significant implications for assessment and

programming, particularly for ESL students.

ii


ACKNOWLEDGEMENTS

To my thesis supervisor, Dr. Dale Willows, my utmost respect and admiration. Dale’s

professionalism, her standards of excellence, and her wish to be involved in research that

makes a difference so that all students can learn to read and write, are exemplary

contributions to the field as well as to individual lives. She has been an exemplary role

model and mentor to me. To Dr. Tom Humphries and Dr. Esther Geva for their insight

and expertise, and to Dr. Lesly Wade-Woolley, thank you for the valuable contributions

you made to this thesis.

To my parents, who I am sure are quite proud and also quite relieved that I am finished.

To Christopher and Scott, for understanding when Mommy had to work on her ‘big

paper’ yet again, and finally, to my husband Dave, who has been my biggest supporter all

along this journey. Thank you so much for your encouragement and your support, and

for helping to make this dream a reality. I could not have done this without you.

iii


TABLE OF CONTENTS

CHAPTER ONE: INTRODUCTION.................................................................... 10

Literature Review.........................................................................................10

Overview....................................................................................................... 10

Reading and Writing Development............................................................ 11

ENG and ESL Reading and Writing Development.................................. 11

The Simple Views of Reading and Writing.............................................. 12

The Relation between Reading and Writing............................................. 15

Research on the Relation between Reading and Writing..........................16

The Stages of Reading and Writing Development................................... 18

Reading and Writing Development at Grade 3.........................................20

Reading and Writing Development at Grade 5/6..................................... 21

Interpretation of Stage Theories............................................................... 22

Cognitive and Language Processes Underlying Reading and Writing

Development................................................................................................. 23

Cognitive Processes.................................................................................. 24

Language Processes.................................................................................. 28

Similarities and Differences between ENG and ESL students in Reading,

Writing, Cognitive and Language Development.......................................31

Oral Language Proficiency as it relates to differences between

ENG and ESL students............................................................................. 33

Goals of the Present Study.......................................................................... 34

Research Questions......................................................................................36

CHAPTER TWO: METHOD.................................................................................37

Description of Participants and Measures................................................ 37

Participants.......................................................................................37

Measures........................................................................................... 39

Procedures.................................................................................................... 49

iv


CHAPTER THREE: RESULTS AND DISCUSSION......................................... 50

Overview....................................................................................................... 50

Variable Screening.......................................................................................50

Similarities and differences between ENG and ESL students.................51

Reading Measures..............................................................................53

Writing Measures...............................................................................58

Cognitive Measures........................................................................... 58

Language Measures........................................................................... 59

The relation between lower- and higher-level reading and writing

skills...............................................................................................................60

Variance accounted for in lower- and higher-level reading and

writing.........................................................................................................62

Regression split by language (ENG/ESL)..................................................67

CHAPTER FOUR: GENERAL DISCUSSION....................................................70

Contributions............................................................................................... 70

Limitations....................................................................................................73

Diagnostic and Instructional Implications................................................ 74

References.....................................................................................................76

V


LIST OF TABLES

Table 1. Stage 1 - Beginning Literacy Development - Kindergarten to Grade 2... 19

Table 2. Stage 2 - Confirmation and Fluency - Grade 2 to 4................................. 19

Table 3. Stage 3 - Reading and Writing for Learning the New - Grade 4 to 8......20

Table 4. Languages Represented in the ESL group............................................... 38

Table 5. Participant Characteristics - Fall 1999..................................................... 39

Table 6. Test Measures............................................................................................40

Table 7. Group Comparisons on Reading, Writing, Cognitive and Language

Processing Measures (based on raw scores).............................................54-

Table 8. Partial Correlations Between the Lower- and Fligher-Level Reading

and Writing Skills by Grade (3, 5/6) and by Language (ENG/ESL)........61

Table 9. Factor Loadings of the Reading and Writing Measures - Principal

Component Analysis with Varimax Rotation...........................................64

Table 10. Factor Loadings of the Cognitive and Language Processing Measures

Principal Component Analysis with Varimax Rotation........................... 65

Table 11. Summary of the Variance Accounted for by the Composites on

the Lower- and Higher-Level Reading Skills For ENG and ESL............ 67

Table 12. Summary of the Percentage of Variance Accounted for by the

Composites on the Lower- and Higher-Level Writing Skills

For ENG and ESL Learners......................................................................68

vi


LIST OF FIGURES

Figure B-l. Mean Spelling by Grade and Language

(WRAT Spelling raw scores)................................................................88

Figure B-2. Mean TOWL - III Spelling Errors (# of errors) by Grade

and Language........................................................................................89

Figure B-3. Mean RAN time (measured in seconds) by Grade and Language 90

Figure B-4. Mean Sequences by Grade and Language (CMS raw scores).............. 91

Figure B-5. Mean Receptive Vocabulary by Grade and Language

(PPVT raw scores).................................................................................92

Figure B-6. Mean Syntax by Grade and Language (CELF-III Formulated

Sentences raw scores)............................................................................ 93

Figure B-7. Mean Expressive Vocabulary by Grade and Language

(EVT raw scores)................................................................................... 94

Figure B-8. Mean Listening Comprehension by Grade and Language (Durrell

raw scores)............................................................................................. 94

vii


APPENDICES

Appendix A................................................................................................................84

Appendix B................................................................................................................ 88

Appendix C................................................................................................................98

Appendix D................................................................................................................ 102

Appendix E.................................................................................................................I l l

Appendix F.................................................................................................................113

Appendix G................................................................................................................ 115

Appendix H................................................................................................................ 118

Appendix 1..................................................................................................................114

viii



CHAPTER ONE: INTRODUCTION

For years, much of the research on reading and writing development has had a relatively

narrow focus. Although a great deal of programmatic research has been conducted on the

lower-level skills involved in reading (e.g., decoding/word identification) and, to a lesser

extent, writing (e.g., spelling), until quite recently, much less research has investigated

the acquisition of higher-level reading skills (e.g., reading comprehension), and even less

has studied higher-level writing skills (e.g., story construction). Moreover, with a few

notable exceptions, researchers have largely neglected the study of the interrelations

between the development of reading and writing skills. Furthermore, until quite recently,

the focus of attention has been almost exclusively on the study of reading and writing

processes among children who speak English as a first language and few researchers have

studied the development of English reading and writing skills among students for whom

English is their second language (ESL). This latter shortcoming of the research literature

is particularly concerning given the burgeoning numbers of ESL students learning to read

and write in English language schools.

The purpose of the present study was to investigate the relation between lower-level

(decoding and spelling) and higher-level (reading comprehension and story construction)

reading and writing skills among students in Grades 3 and 5/6 for whom English was

their first (ENG) or second (ESL) language. In addition, the research was designed to

explore similarities and differences in the underlying cognitive and language processes

associated with reading and writing development in ENG and ESL students.

Literature Review

Overview

This review is presented in five sections. The first examines the literature on reading and

writing development for ENG and ESL students, while the second section explores the

relation between reading and writing in these two populations. The third section

10


discusses the cognitive and language variables that researchers have found to correlate

significantly with the acquisition of reading and writing. The fourth section examines the

literature on the similarities and differences between ENG and ESL cognitive, language,

reading and writing development. The final section of the literature review outlines the

goals of the present study.

As mentioned in the Introduction, the literature documenting reading acquisition in

English is substantial whereas the literature on writing acquisition in English is, at this

point, less developed. Very few studies have explored the relation between reading and

writing development in English. Moreover, the research literature relevant to the

development of English language reading and writing skills among ESL learners is even

smaller. This review will integrate the available literature for ENG and ESL learners

throughout, starting with theories of reading and writing development.

Reading and Writing Development

ENG and ESL Reading and Writing Development.

Geva and Verhoeven (2000) document that although a considerable body of research

exists on reading acquisition in children’s first language (LI), much less attention has

been paid to the acquisition of reading in a second language (L2). The literature in L2

reading research at present is still dominated by LI frameworks. These frameworks -

where English is, by far, the most frequent LI represented - conceptualize reading as a

complex information processing operation that draws on underlying component processes

which can be a potential source of individual differences (Daneman, 1996). Efficient

lower-level processing allows the limited capacity system to be devoted to processing

higher-order information during reading comprehension (Stanovich, 1991). L2 reading

has also been conceptualized as a complex cognitive skill, which is characterized by the

development of adequate skill in lower-level graphophonic processing and a gradual

automatization of decoding (McLaughlin, 1990). In Bernhardt’s (1990) multivariate L2

reading model, lower-level word recognition processes are important parameters of L2

11


reading comprehension. Other researchers have also stressed the importance of lower-

level component processes in L2 reading (Haynes & Carr, 1990; Koda, 1994). While

these frameworks exist on reading development in English as a first (ENG) and second

(ESL) language, little theory or research exists on writing acquisition for either ENG or

ESL, with the exception of Haynes and Carr (1990) who applied the component process

framework to writing development in ESL students. The “simple views” of reading and

writing offer a common framework for the discussion of lower- and higher-level reading

and writing development in both ENG and ESL students.

The “Simple Views” of Reading and Writing

The simple views of reading and writing, originally presented by Gough and Tunmer

(1986) and Juel (1988), provide a framework by which to conceptualize the

representative lower- and higher-level reading and writing skills. Although this theory

has been applied specifically to the development of reading and writing in English, the

ESL literature has utilized a framework by which the same lower- and higher-level

reading skills (Perfetti, 1985) were identified. There is no corresponding model of

writing presented within the ESL literature.

Gough and Tunmer (1986), Hoover and Gough (1990), and Gough (1996) outlined a

“simple view” of reading ability that is composed of two factors: decoding (i.e., word

identification) and language comprehension, (a) Word identification is the result of the

application of decoding skills (word analysis based on phonemic awareness - blending -

and grapheme-phoneme correspondences), (b) Language comprehension is the process

by which the meanings of decoded words can be integrated into meaningful sentences

and text structures. As part of the lower-level reading process, phonemic awareness is

required for decoding which then leads to word identification. Once this lower-level skill

(decoding) has been consolidated (leading to fluent - automatized - word identification),

readers are then able to focus on attributing meaning to the text they have decoded

(comprehension). A single underlying process is seen as producing both listening and

reading comprehension, despite differences in speech and written text. Thus, in the

12


simple view of reading, a poor reader (i.e., reading comprehender), is a poor word

identifier, a poor language comprehender, or both. The sequence by which these reading

skills develop is represented therefore, as follows: decoding, word identification, fluency

and reading comprehension. Reading comprehension has come to be viewed as the

“essence of reading” (Durkin, 1993) because it integrates complex cognitive and

language skills, and it depends on accurate and efficient word identification in order to

access meaning from print.

Juel (1988) outlined a “simple view” of writing, also composed of two factors: spelling

and ideation, (a) Spelling involves the application of encoding processes, some of the

same processes as decoding. Encoding (i.e., analysis of a spoken word’s sounds and the

representation of the sounds in print, also based on phonemic awareness - segmentation -

and phoneme-grapheme knowledge) is the process by which a writer represents the

sounds of spoken words in print, (b) Ideation is the ability to generate and organize

ideas, encompassing both the generation of creative thoughts and their organization into

sentence and text structures. Thus, in the simple view of writing, a poor writer (i.e.,

producer of written composition) is a poor speller, a poor ideas generator and organizer,

or both. The sequence by which these writing skills develop is represented, therefore, as

follows: spelling, productivity, the ability to use the conventions of writing, the ability to

use the language (sentence structure) of writing, and finally, the ability to organize ideas

and to construct written compositions. Written composition is a final common pathway

of multiple developmental functions integrating complex cognitive and language skills.

As with reading comprehension, written composition depends on the consolidation of

lower-level skills and fluency, in order to allow the student to focus, in this case, on the

organization of complex ideas into a written composition.

Juel acknowledges that these models of reading and writing may seem somewhat

simplistic. Only two components seem necessary to characterize both the “lower” word-

level skills (i.e., word identification or spelling) and the “higher” text-level skill (i.e.,

reading comprehension or written composition) of reading and writing; however, each

component is complex in its own right, and may be broken down into subcomponents.

13


Poorer lower-level processes may impede the development of higher-level processes. For

example, until the lower-level process of spelling is somewhat automatic - at least for a

number of high frequency words - the attention of the writer may be diverted from

higher-order composing processes (Scardamalia, 1981). In reading, efficient word

identification leads to better comprehension. The “simple view” framework will be

adopted in examining the relation between the lower- (word identification and spelling)

and higher-level (reading comprehension and written composition) reading and writing

skills, and also in evaluating the underlying cognitive and language processes that are

correlated with those lower- and higher-level reading and writing skills.

There are no explicit references to the simple view of reading in the ESL literature;

however, Geva and Clifton (1993) describe reading as involving two interrelated factors:

word identification, or lexical access, and comprehension. They reference Perfetti (1985)

in that “ complex comprehension processes build upon word recognition process”.

According to Perfetti, reading comprehension depends upon the efficient operation of

lower lexical processes. Skilled readers can direct their attention to text comprehension

rather than decoding or word recognition, because of efficient basic processes. There are

no references to the simple view of writing in the ESL literature, but there is also a lack

of literature evaluating writing development in general, beyond the development of

spelling. Within the ESL literature, results from cross-linguistic comparisons of reading,

spelling, cognitive and linguistic processes implicitly fit the simple view of reading in

that different cognitive and language processes have been found to correlate with higher-

leading comprehension) as opposed to lower-level (word identification) skills. This ESL

research has identified that oral language proficiency (OLP) is positively related to

higher-level comprehension (Geva & Ryan, 1993; Verhoeven, 1990) while cognitive

analytical skills are related to lower-level word identification skills.

It is clear from both research and models of reading and writing development in ENG and

ESL that the processes that are of central importance in the beginning stages of

development are not as important (in the sense that they are performed automatically, so

they do not require as much attention) at later stages of development. It is also clear that

14


there are many interdependent processes involved in the development of both reading and

writing skills. This has lead researchers to speculate that the development of reading and

writing skills is related. The automatization of lower-level reading and writing skills,

which are dependent on similar underlying cognitive and language processes, is required

for higher-level skills to develop. Theory and research has evaluated this relation

between reading and writing, primarily at the rudimentary stages of decoding and

spelling. The evaluation of the similarities and differences in the cognitive and language

processes that underlie reading and writing, within the context of stage theories of

literacy development, will enable the interpretation of the relation between reading and

writing.

The Relation between Reading and Writing

There are theoretical models (Ehri, 1986; Frith, 1985) and empirical research (Clarke,

1988; Morris & Pemey, 1984) concerning the relation between reading and spelling

acquisition during the early stages of literacy development. Ehri (1986) theorized that

reading and spelling depend on the same knowledge sources in memory: knowledge

about the alphabetic system, and knowledge about the spelling of specific words.

Similarly, Henderson (1978) theorized that developing word knowledge, composed of

phonemic awareness and orthographic awareness is conceptual in nature and underlies

children’s ability to both read and spell words. Empirically, Henderson found changing

developmental patterns in spelling with an increasing correlation between reading and

spelling in grade-one students as the school year progressed.

While there is considerable research on this relation between selected measures of early

reading and spelling, there is not much research extending beyond that to later stages of

development. There is little research evaluating higher-level reading comprehension as it

relates to higher-level writing (as in ideation, organization of ideas and support for

arguments). Research does exist on the development of story schema (a person’s

conception of what constitutes a story and how a typical story is organized from

beginning to end) but the focus has not been related to the higher-level reading

15


comprehension tasks. Panto (1999) tracked reading and writing development

longitudinally from grade 1 to 4, comparing ENG and ESL learners, but did not focus on

the relation between reading and writing measures across grades. She did, however,

develop a scoring system for story schema based on the work of Nodine, Barenbaum, and

Newcomer (1985). According to this scheme, Panto scored children’s ideas in their

understanding and production of stories and found that their story ideation changed with

development. As children gained more exposure to different types of stories, their ability

to use more complex ideation of their own also changed (Stein & Glenn, 1982; Stein &

Policastro, 1984), but not as a result of their language status. There were no significant

differences between ENG and ESL learners.

As indicated earlier, most examinations of the relation between reading and writing

development have involved small numbers of measures collected at a single point in time,

with small samples of children. Some studies have also attempted to measure underlying

cognitive and language processes, but again, most have focused on only one or two

specific measures. The broadest work investigated differences in the relation among

multiple reading and writing measures in grades 2 and 5, but it did not investigate

underlying cognitive and language processes (Shanahan, 1984). Another large-scale

study, by Juel (1988), tracked reading and writing development longitudinally from grade

1 to 4, but did not focus on the relation between the two skills or the underlying cognitive

and language processes. In addition, neither of the above mentioned studies investigated

these issues from a second-language perspective.

Research on the Relation between Reading and Writing

Shanahan (1984) and Shanahan and Lomax (1986,1988) examined 256 students in grade

two and 251 students in grade five. An exploratory analysis of multiple reading and

writing measures was undertaken using canonical correlational analysis with separate

analyses performed for each grade level and for beginning and proficient readers. The

word recognition factors drawn from the reading set were most related to the spelling

variables of the writing set at both grade levels and at the beginning reader level.

16


Substantial differences were found across the reading level cohorts, however. For

beginning readers, the relationship was based on word recognition and spelling levels.

Thus, the ability to apply basic phonics rules in decoding and the ability to spell

accurately were closely related. As students became more proficient as readers, the

nature of the relation between reading and writing changed with an increasing importance

of sophisticated vocabulary and story structure to writing achievement, and an increasing

importance of reading comprehension to reading achievement. For proficient readers, the

ability to structure prose in complex ways and use a variety of vocabulary in writing was

related to the prose comprehension factor. In none of the analyses was reading or writing

found to explain more than 45 percent of the variance in the opposite test set. Using

these data, Shanahan and Lomax (1986, 1988) then conducted a series of LISREL path

analyses of alternative models of reading-writing relations. Their analyses indicated that

reading and writing influenced each other. The model that best fit their data was an

interactive one. This model postulates that reading can influence writing development

and that writing can influence reading development.

On the basis of the interactive model of reading and writing and other research,

Fitzgerald and Shanahan (2000) present an introductory developmental view of the

relation between reading and writing. They make a specific theoretical contribution by

outlining the critical knowledge required for each of the stages of reading and writing

development using Chall’s (1996) framework. They cite research into reading-writing

connections that analyzes the shared knowledge and cognitive processes between reading

and writing at the different levels of development. This model begins with the

proposition that reading and writing are constellations of cognitive processes that depend

on knowledge representations at various language levels (phonemic, orthographic,

semantic, syntactic, and pragmatic). Reading and writing are related because they

depend on similar knowledge representations and cognitive processes across

developmental stages. Therefore, it would be expected that reading and writing

developments should parallel each other closely. While Fitzgerald and Shanahan (2000)

describe critical cognitive features that are likely to be important to both reading and

writing proficiency at different stages, their model is most useful as an organizing plan.

17


When Chall’s (1983, 1996) model of the stages of reading development and Levine’s

(1998) model of writing development are combined, the understanding of reading and

writing at different developmental stages becomes more fleshed out and applicable.

The Stages of Reading and Writing Development

Chall (1983,1996) proposed broad developmental stages of reading, spanning birth

through adulthood. She contributed significantly to the field of reading research by

showing that reading is not a single skill, and that word recognition, word meaning and

reading comprehension are separate aspects of literacy that need to be developed

differently at the different stages of literacy development. In contrast to reading, few

broad and comprehensive stage theories exist of writing development. Levine’s (1998)

six-stage model of writing development, which parallels Chall’s reading stages, provides

an important contribution to this literature by outlining comprehensive and detailed

expectations regarding the particular components of writing acquired at each stage. In

addition, both models contribute by adding the critical elements of language development

at each stage that influence literacy development.

Chall’s (1983,1996) and Levine’s (1998) models of the stages of reading, writing and

language development will be discussed as they relate specifically to the stages of

development relevant to the current study. Tables 1 to 3 outline the stages of reading,

writing and language development from Initial Literacy, Kindergarten to Grade 2 (Table

1), to the developmental stages of Confirmation and Fluency, Grade 2 to 4 (Table 2) and

Reading and Writing for Learning the New, expected at Grade 4 to 8 (Table 3).

18


Table 1. Stage 1 - Beginning Literacy Development - Kindergarten to Grade 2

Reading_____________________________________________________________- phonemic awareness__________________________________________________- mastery of alphabetic principle (phoneme-grapheme correspondences)___________- word-by-word (dysfluent) reading________________________________________

Writing _____________________________- acquisition of letter and number forms _________________ _- beginning appreciation of spelling accuracy and the use of invented spelling______- introduction to the conventions of capitalization, punctuation and sentence structure

Language____________________________________________________________- vocabulary expansion ________________________________________________- use of unsophisticated language in writing _____________ __________ _* Adapted from Chall (1983,1996) and Levine (1998)

Table 2. Stage 2 - Confirmation and Fluency - Grade 2 to 4

Reading_________________________________________________________________- growing sight vocabulary___________________________________________________- increasing speed and efficiency______________________________________________- reading to confirm existing knowledge________________________________________

Writing_________________________________________________________________- rapidly increasing spelling ability____________________________________________- the integration of conventions (punctuation, capitalization) with language

(morphology, syntax, narrative organization)___________________________________

Language________________________________________________________________- linguistic and cognitive contents of materials remain beneath student’s processing ability- written language is less sophisticated than speech_______________________________* Adapted from Chall (1983,1996) and Levine (1998)

19


Table 3. Stage 3 - Reading and Writing for Learning the New - Grade 4 to 8

Reading__________________________________________________________________- reading to acquire new knowledge____________________________________________- reading rate and efficiency improve considerably ______ _ _ _ _____ ___________- appearance of new and specialized vocabulary - expanding understanding of morphology

Writing__________________________________________________________________- writing occurs with less expenditure of conscious thought_________________________- the ability to produce larger volumes of writing develops__________________________

Language________________________________________________________________- reading content becomes de-contextualized and the linguistic sophistication of what

students are reading catches up to and surpasses the contents of everyday speech_______- written language approximates speech______________ ___________________________* Adapted from Chall (1983,1996) and Levine (1998)

There is research evaluating the underlying cognitive and language processes related with

reading and writing at early stages of literacy development, but not much at the more

advanced stages of literacy development. In the present research, children in grades 3, 5

and 6 were selected because they would be at stages where enough children would have

already developed the lower-level reading and writing skills. This would allow for the

extension of knowledge about the interrelation between the underlying cognitive and

language processes and reading and writing development by comparing both lower- and

higher-level reading and writing skills within the same sample.

Reading and Writing Development in English at Grade 3

According to Chall (1983, 1996), children in grades 2 and 3 are generally still learning to

read. They read primarily to confirm what they know. The language and ideational

content of what they are reading is not as sophisticated as that encountered in daily

conversation, magazines, movies and television. They are acquiring a rapidly expanding

sight vocabulary by which they can read more efficiently and begin to add greater

linguistic sophistication. Concurrently, they are developing word analysis skills, learning

more advanced rules of phonics, and beginning to recognize the basic structure of

morphology (suffixes, prefixes, compound words, roots and syllables). During this time,

20


their language skills are also improving which enables the use of context clues to assist in

decoding and comprehension. The use of context may be particularly more difficult for

ESL learners. The emphasis dining this stage is on fluency and using context. In terms

of writing, according to Levine (1998), children in grades 2 to 4 are learning to gradually

incorporate standards of capitalization, punctuation, syntax, and grammar. Writing is

more often used to state the obvious or relate an experience than to solve a problem or

develop elaborate ideas. During this stage, conversational speech is more sophisticated

and syntactically complex than writing. The language content is likely to be

uncomplicated with the use of simple declarative sentences and few subordinate clauses.

Reading and Writing Development in English at Grades 5/6

Students in grades 4 to 8 are generally at a stage where they read to acquire new

knowledge (Chall, 1983,1996). The linguistic sophistication of what they are reading

catches up and surpasses the content of everyday speech and the media. Textbooks

provide an important source of new knowledge. Comprehension skills play a more

decisive role as reading content becomes de-contextualized, removed from everyday

experience. There is also a demand for the integration of passage reading and memory,

as students must store and recall important facts from what they have read. In addition,

new and specialized vocabulary is introduced during reading. A student now confronts

words in print that are seldom, or never encountered in daily conversation. Students in

grades 4 to 7 write with greater volume (Levine, 1998) and are expected to not only apply

the mechanics of writing with increasing automaticity, but also to review their work and

correct errors in sentence structure, grammar, capitalization, punctuation and spelling.

Language usage becomes increasingly important. Written language starts to approximate

spoken language as language usage becomes increasingly important. Sentences become

longer and contain more sophisticated vocabulary and conceptual content. For the first

time, children use words in writing that they would not likely use in conversation and

writing is no longer used as an end or a process in itself. Instead, it becomes a means of

communicatioa Thus, greater specificity and accuracy in the use of vocabulary and

language gains importance.

21


Interpretation of Stage Theories

While these models provide valuable information regarding expected development, it is

important to recognize that these stages are meant to be useful as a heuristic. They are

not reached automatically by the grades indicated, but are an approximate guideline to

when literacy skills are mastered, and outline a developmental sequence in order to

understand individual differences. Times vary, particularly for students who have

difficulty learning languages, whether it be their first language or second. Some students

encounter more difficulties at one stage than another, and some may reach a certain stage

and never pass beyond it.

There is no research explicitly evaluating the stages of literacy development for ESL

students. In the introduction to the second edition of ChalTs Stages o f Reading

Development (1996), Geva indicated, based on the L2 literature, that there is no reason to

believe that ESL students will not reach the same stages of literacy development as ENG

students. She did suggest, however, that the pace at which they arrive at each stage may

differ. In addition, ESL students may have more difficulty accessing context as required

due to insufficient vocabulary development and lexical access.

Fitzgerald and Shanahan (2000) provide persuasive evidence that reading and writing

rely on correlated mental processes, though the nature of the relation is different at

different age or grade levels. The closest correlations between them are best summarized

by the aspects of learning that are particularly important or most variant at a given stage

of development (such as phonological-orthographic development in the early stages of

literacy learning). It is necessary to establish not just from a proficiency standpoint, but

from a normal development standpoint, which cognitive and language processes underlie

these lower- and higher-level reading and writing skills. It is also important to

understand which cognitive and language processes account for the most variance in

lower- and higher-level reading and writing skills at different stages of development for

children who represent ENG and ESL populations.

22


Because reading and writing are both developmental processes, what is learned at one

stage of development can be qualitatively different from what is learned at another stage

of development. If the underlying cognitive and language processes that are associated

with reading and writing are the same within, but different across the lower and higher

levels of development, more will be understood about the nature of the relation between

reading and writing across development. The present study offers the opportunity to

examine these issues across the lower and higher levels of reading and writing, while also

evaluating the contribution of various cognitive and language processes correlated with

that development. Furthermore, the study extends that understanding to ESL students.

By evaluating similarities and differences in the amount of variance accounted for by the

underlying cognitive and language processes at the lower and higher levels of reading

and writing for ENG and ESL students, valuable information will be gained for the

assessment of written language difficulties, as well as for initial teaching, and remedial

intervention.

Cognitive and Language Processes Underlying Reading and Writing Development

A vast literature exists on the underlying processes that affect the acquisition of reading

and writing skills in English, and how these processes differ in individuals of varying

abilities (e.g., Ceci, 1984; Denckla & Rudel, 1974; Juel, 1988; Scarborough, 1989;

Stanovich, 1986). Researchers have found that the ability to read and write is dependent

on a number of complex, interrelated processes, such as phonemic awareness, rapid

automatic naming, working memory, cognitive capacity, auditory processing skill and

language ability. The majority of this research has focused on the acquisition of reading

skills with the assumption that if an individual has difficulty learning to read, he/she will

also have difficulty learning to write since reading and writing share so many of the same

underlying processes. However, writing is generally acknowledged to be the more

difficult of the two since it is dependent on recall, while reading is dependent on

recognition memory (Juel, Griffith, & Gough, 1986). Juel (1988) found that some

children who were poor writers had difficulties with both form (e.g., spelling) and

process (e.g., content generation), whereas others had difficulties with just one or the

23


other. It can be anticipated that these particular aspects can be a significant problem for

students whose oral language may not be as well developed and understood because both

decoding and comprehension skills are required for writing. The present study offers the

opportunity to evaluate which of the many cognitive and language predictors associated

with reading and writing development, explain the most variance in the lower and higher

levels of reading and writing development for ENG and ESL learners.

Just as there are lower- and higher-level reading and writing skills, it is theorized that

there are lower- and higher-level cognitive and language skills that make a unique

contribution to reading and writing development. The assumption is that lower-level

cognitive and language processing components (e.g., phonemic awareness, rapid

automatic naming, sequencing, working memory, short-term memory) take place prior to

higher-level semantic and syntactic processing (e.g., vocabulary, grammatical judgment)

as the higher-level processes depend on the information supplied by lower-level word

identification processes (Haynes & Carr, 1990). The following subsections describe the

cognitive and language processes that have been found to correlate very highly with

reading and writing achievement.

Cognitive Processes

Phonemic awareness

Research has consistently shown that phonemic awareness is a better predictor of later

reading achievement than IQ or even general language proficiency (Griffith & Olson,

1992). In fact, from a considerable body of research on English literacy, phonological

(i.e., phonemic) awareness has been highlighted as the single most important predictor

(Verhoeven, 2000). It is generally believed that a minimal level of phonemic awareness

is required to learn to read and spell, and that phonological awareness develops further as

literacy abilities develop (Stanovich, 1992). Children who lack phonemic awareness do

not have the tools they need to become successful readers and/or writers. They have not

yet attained adequate awareness of the sound structure of words. They have insufficient

phoneme awareness to allow them to make sense of the alphabetic writing system. Less

24


investigation has been done into the relation between phonemic awareness and writing

ability (Griffith, 1991). However, given that the English writing system is essentially

phonemic, if a child has not yet grasped that words are made up of individual phonemes,

he or she will be unable to master the alphabetic code by which phonemes correspond to

graphemes. Thus, just as learning how to read words requires phonological awareness, so

does learning how to spell (Bradley & Bryant, 1983; Juel et al., 1986).

Research on ESL reading development in school children (e.g., Bruck & Genesee, 1995;

Cisero & Royer, 1995; Dorgunglu et al., 1993; Geva, 2000; Geva, 1998; Geva &

Gholamain, 1998; Geva & Wade-Woolley, 1998; Service 1992; Service & Kohonen,

1995) strongly implicates underlying cognitive and language factors such as phonological

processing skills, phonological memory and naming speed. There is evidence that these

same processes are also important contributors for native speakers of other languages,

ranging from Chinese (So & Siegel, 1997) to Norwegian (Hoien, Lundberg, Stanovich &

Bjaalid, 1995) to Hebrew (Ben-Dror, Bentin, & Frost, 1995). Phonemic awareness and

rapid naming tasks contributed unique variance to word recognition performance for both

ENG and ESL students (Geva, Yaghoub-Zadeh, & Schuster, 2000).

Rapid Automatic Naming (RAN)

Rapid automatic naming is another component of phonological processing that

researchers have found to play an important role in the acquisition and development of

written language skills. Denckla and Rudel (1974,1976) provided evidence for a general

naming retrieval deficit. They showed that poor readers were significantly slower than

good readers in naming continuous lists of single digits, letters, colours and objects.

These tests of naming speed were designed to measure how automatically children recode visual information into a phonologically represented name code. Bowers and Swanson

(1991) described the development of automaticity in the use of lower-level reading skills

as crucial to developing skilled reading. Children who suffer from a retrieval deficit

affecting the rate at which they access and retrieve information are clearly at a

25


disadvantage. Subsequent studies have replicated these findings while controlling for IQ,

reading experiences and socio-economic status (Meyer, Wood, Hart & Felton, 1998).

Longitudinal studies of early reading development have shown that rapid naming tasks in

the pre-literacy period are a major predictor - along with phonemic awareness skills - of

later reading skill (Meyer et al, 1998; Scarborough, 1989). They also show that rapid

naming tasks contribute variance to reading measures independent of the contribution by

phonemic awareness measures. In fact, the ‘‘double deficit” hypothesis, Bowers (1995)

and Bowers and Wolf (1993), stress somewhat separate phonological and rapid naming

mechanisms. In this two-factor theory, speed of processing is the critical component

tapped in rapid naming tasks, and such tasks contribute specific variance to orthographic

processing skill.

Less research is available on the role of RAN in the acquisition of writing skills in

English. However, just as fluent reading is dependent on the automatization of lower-

level reading processes such as decoding, it can be surmised that written composition is

largely dependent on the automatization of spelling patterns. Until the lower-level

spelling and writing conventions (capitalization, punctuation) are automatic, the attention

of the writer may be diverted from higher-level composing/ideation (Bereiter, 1980;

Scardamalia, 1981).

Research on ESL reading development in school children, as previously indicated,

strongly implicates naming speed in reading development (e.g., Brack & Genesee, 1995;

Cisero & Royer, 1995; Dorgunglu et al., 1993; Geva, 1998; Geva & Gholamain, 1998;

Geva & Wade-Woolley, 1998; Service 1992; Service & Kohonen, 1995). In the ESL

research literature, the potential relation between RAN performance and writing

development is virtually unexplored.

26


Memory: Sequencing, Short Term Memory, Working Memory

As a child learns to decode and spell, she or he must segment words into their written

symbols and associate those symbols with specific sounds. To identify words, the child

must hold the sounds in active working memory and then synthesize them, blending them

to form a recognizable word. Poor readers have been found to perform less well than

good readers on a variety of short-term and active working memory tasks, such as the

digit span test (digits forward and backward), and to have difficulty recalling in order,

strings of letters, digits, nonsense syllables or words (Brady & Shankweiler, 1991; Mann

& Brady, 1988). In fact, comprehension problems have also been found in some cases to

be predominantly due to working memory difficulties (Gottardo, Stanovich, & Siegel,

1996; Mann, Cowin, & Schoenheimer, 1989).

Working memory, in particular, has been found to play an important role in second

language word identification and reading comprehension (Geva & Ryan, 1993; Geva &

Wade-Wooley, 1998; Gholamian & Geva, 1999; Harrington & Sawyer, 1992). Geva and

Ryan (1993) found that working memory plays an even more important role in second

language reading than in first language reading. These researchers interpreted that this

finding was due to the heavier demands posed by working memory because of the lack of

automaticity in executing lower-level component processes in second language than first

language reading.

Verbal memory: Memory for Words, Sentences and Stories

Prediction studies have found that children’s memory for sentences and stories are more

strongly related to word identification and reading comprehension skills than are their

scores on digit span, word span and word repetition measures (Scarborough, 1998).

Among these language measures, phonemic awareness, expressive vocabulary, sentence

imitation and story recall were the strongest predictors of skilled word identification and

reading comprehension, whereas speech perception and articulation were weakest. In

contrast, nonverbal visual and motor tasks provided little or no prognostic information.

27


Language Processes

Vocabulary

The literature on reading acquisition and vocabulary shows that once decoding skills have

become automatized, insufficient knowledge of word meanings may be a primary cause

of academic failure (Becker, 1977, Biemiller, 2001). Robbins & Ehri (1994) showed that

children with larger vocabularies learned more words from storybook readings than

children with smaller vocabularies. Research has indicated that L2 and economically

disadvantaged students who were good decoders in grades one and two start to show a

decrease in academic achievement soon after (Becker, 1977; Chall, Jacobs & Baldwin,

1990) which may be due to vocabulary deficits. Therefore, while vocabulary in the early

stages of reading may not play a significant role in the acquisition of lower-level reading

skills such as word recognition or word attack skills, it becomes increasingly important in

higher-level reading comprehension. Since vocabulary deficiencies are a primary cause

of academic failure in grades 3 through 12 (Baumann & Kameenui, 1991; Becker, 1977;

Stanovich, 1986), children with less-developed vocabularies will become increasingly

disadvantaged as the focus on learning to read changes to one on reading to learn (Chall,

1983,1996). Vocabulary, as a measure of oral language proficiency, is of particular

significance to the ESL literature. Geva and Petrulis-Wright (1998) found that regardless

of ENG-ESL status, children in grades 1 and 2 who have a more expanded vocabulary

and a more sophisticated command of grammar, are also better decoders and

comprehenders.

Syntax/Grammatical Judgment

Syntax involves word order and grammatical construction. The inability to understand

and use syntax has the potential to interfere with sentence comprehension, even if the

child is adept at decoding. Children who have mastered the basic syntax needed to

understand simple constructions may have more difficulty with the complex

constructions found in more advanced textbooks, especially when the content is

28


unfamiliar. Phonological processing and syntactic awareness have been shown to be

significant contributors to the development of reading skills for native English speakers

(Siegel, 1993) and for English-as-a-second-language speakers (Geva & Petrulis-Wright,

1998). In addition, Da Fontoura and Siegel (1995) found that the syntactic skills of

bilingual children bom in Canada still lagged behind the skills of monolingual children in

grades 4, 5 and 6.

Oral Language Proficiency (OLP)

Rice (1989) identified four major dimensions of oral language proficiency (OLP):

phonology (the sound system); semantics (the system of meanings); morphology (the

rales of word formation); and syntax (the rales of sentence formation). In addition, OLP

includes the ability to understand and use speech appropriately in social contexts

(pragmatics). In the L2 research, considerable attention has focused on the contribution

of OLP to reading. This is based on the assumption that because L2 OLP and reading are

both language based processes, achievement in one should contribute to achievement in

the other (e.g., Cummins, 1984). A number of studies have documented the role of oral

language proficiency in L2 reading and supported the existence of a positive relationship

between OLP and reading comprehension (Geva & Ryan, 1993; Verhoeven, 1990).

However, the relationship between OLP and more basic reading skills, such as word

recognition and decoding, has been shown to be perhaps less critical (Durgunoglu, Nagy

& Hancin-Bhatt, 1993; Geva & Siegel, 1999; Geva & Wade-Woolley, 1998; Geva,

Wade-Woolley & Shany, 1997; Gholamain & Geva, 1998). These studies showed that

elementary school children can decode and spell words that may or may not be part of

their vocabulary, even when their OLP is not fully developed. These findings were

surprising as it was generally expected on the basis of theories of language and literacy

development that oral language proficiency was related to reading development.

Cummins’ language interdependence hypothesis (1979) suggests that there is a

significant relationship between children’s skills in acquiring native and foreign

languages. That is, only children who have deficient language and cognitive skills in

29


their first language would exhibit similar difficulties in learning a foreign language.

Conversely, children who have strong cognitive and language skills in their first language

are likely to develop similar strengths when learning a foreign language. Cummins

(1980, 1984) also indicated that it takes an ESL student 5 to 7 years to acquire the

necessary cognitive academic language proficiency skills to achieve comparably with

native-speaking first language students. Cummins indicated that most ESL students are

able to communicate in social situations and day-to-day classroom interactions within

one or two years (basic interpersonal communication skills). However, students may

require from five to seven years to develop the ability to understand the academic

language used in textbooks and to express in English the increasingly complex and

abstract vocabulary, syntax and concepts encountered in the higher grades.

Support for Cummins’s language interdependence hypothesis comes from the studies

comparing LI and L2 basic literacy skills among young bilingual or L2 learners (Bruck

& Genesee, 1995; Durgunoglu, Nagy & Hancin, 1993; Geva & Clifton, 1993; Geva,

Wade-Wooley & Shany, 1993; Geva & Siegel, 1991; Gholamain & Geva, 1997). Results

from these studies suggest that parallel LI and L2 reading skills are positively correlated.

The positive and moderate-to-high correlations between parallel LI and L2 word

recognition skills across different languages, with differing orthographies, suggest that

common underlying cognitive and language processes play an important role in

explaining the correlations between parallel LI and L2 reading tasks. They also suggest

that LI and L2 individual differences in the development of these skills can be predicted

to some extent on the basis of the specific underlying cognitive and language abilities.

The cognitive and language processing skills identified in these studies (i.e., phonological

processing, working memory, orthographic knowledge and RAN, or speed of lexical

access) appear to play a similar role in LI and L2 reading skills acquisition.

Chiappe and Siegel (1999) also found evidence to support Cummins’s language

interdependence hypothesis in cross-linguistic phonological transfer. For both ENG and

ESL children in their sample, reading difficulties appeared to be strongly linked with

impaired phonological processing. Thus, phonological processing has been shown to

30


play a critical role in reading acquisition for both first and second language speakers.

Other than the finding that the ESL learners had poorer grammatical sensitivity and lower

scores on an oral cloze task, the profiles of the ESL children were very similar to those of

native English speakers. Reading difficulties in English occurred with approximately the

same frequency for native and non-native speakers of English. These findings are

consistent with the findings of DaFontoura and Siegel (1995). Although the Portuguese-

English bilingual children in their study had no difficulties on word reading tasks, their

oral cloze scores were significantly lower than those of English monolingual children.

Although it would be hoped that this difference would diminish in time, DaFontoura and

Siegel (1995) found that the syntactic skills of these children, bom in Canada, still lagged

behind the skills of English monolingual children in grades 4, 5, and 6. The present study

will compare similarities and differences in the development of ENG and ESL children’s

syntactic skills, and also assess the contribution that judgement of syntax makes, as part

of an oral language proficiency composite, to higher-level reading and writing skills.

Similarities and Differences between ENG and ESL students in Reading, Writing

Cognitive and Language Development

Evidence is accumulating that while there are few differences in reading, writing, or even

cognitive measures between ENG and ESL students (Panto, 1999), there are significant

differences in language processes (e.g., receptive vocabulary and grammatical judgment),

favoring ENG students, which take longer to disappear (DaFontoura & Siegel, 1995;

Geva, 1998; Geva & Pefrulis-Wright, 1998). These findings seem to support Cummins’s

5-to-7-year timeline for ESL students to achieve comparable language proficiency (e.g.,

vocabulary and syntax knowledge) with native speakers. An apparent inconsistency is

found in that the ESL research does document that ESL students have achieved

comparable performances on some of the higher-level academic tasks (reading

comprehension and written language), although at earlier stages of development (Geva &

Petrulis-Wright, 1998; Panto, 1999).

31


Panto (1999) compared the spelling and higher-level writing ability of ENG and ESL

students. She found that ENG and ESL students performed similarly on a variety of

spelling and writing measures, and that there were no differences in the patterns of

growth in spelling and writing development in these children across grades 1 to 4, as well

as no significant differences between ENG and ESL students in their higher-level story

schemas.

Geva (1998) compared ENG and ESL students’ performance on reading, writing,

cognitive and language measures in grades 1 and 2. Measures of word recognition,

spelling, phonemic awareness, rapid naming, working memory, receptive and expressive

vocabulary, grammatical judgment and listening comprehension were collected. None of

the children categorized as ESL in this study had been in an English-speaking

environment for more than 2 years. Literacy instruction in the primary grades in the

school setting in which these data were collected involved a balanced combination of

activities and instructional approaches (Willows, 1996) designed to enhance children’s

general language development, literacy skills and phonemic awareness and alphabetic

coding (Lloyd, 1983). The results on the reading and writing tasks did not support an

expected ENG advantage. A significant difference was not found on word recognition

between ENG and ESL students. Although a significant difference was found on spelling,

to the advantage of ENG students, it was noted that the group differences were not that

striking. Results comparing the cognitive and language skills of ENG and ESL students

indicated that on measures of phonological processing and letter naming speed (RAN),

the two groups did not differ significantly. In contrast, ENG children showed an

advantage on working memory measures for verbal information and for numbers, and on

all of the oral language proficiency measures (receptive vocabulary, expressive

vocabulary, grammatical judgment and listening comprehension).

A series of multiple regressions was then undertaken to examine the role of the

underlying cognitive and language processes in accounting for variance in the reading

and listening skills in the ENG and ESL groups. The lower-level phonemic awareness

and RAN (letter naming speed) tasks were found to be the only significant predictors of

32


the lower-level word attack, word recognition and spelling skills in each of the language

groups. Measures of oral language proficiency played a significant role in accounting for

higher-level listening comprehension. Measures of higher-level reading comprehension

and written composition were not collected in this study. In addition, the children in this

study were in grades 1 and 2, during the earlier stage of beginning literacy development.

The fact that ESL students had less developed English vocabulary knowledge and less

command of English syntax at this stage was not surprising. What was less obvious was

the fact that while proficiency on these oral language measures played a significant role

in explaining higher-level listening comprehension, it was less crucial in the development

of lower-level L2 word recognition skills.

Oral Language Proficiency as it relates to differences between ENG and ESL

Petrulis-Wright (1998) found that LI and L2 children differed in English OLP, but not on

reading tasks in grades 1 and 2. She also found that OLP played a greater role in

predicting reading comprehension than in predicting basic reading skills, and that it

contributed significantly to the explained variance of L2 basic reading skills, but not to

LI. She concluded that LI children are utilizing analytical skills, while L2 children are

utilizing OLP skills to achieve the same performance. Three to four percent of the

variance of basic reading skills was accounted for by the interaction between language

status, and vocabulary and grammatical judgment. That is, the relationship between these

two language measures and basic reading skills differed depending on whether the

children spoke English as a first or second language. The authors hypothesized that the

significant interaction of LI -L2 status with OLP may be due to the fact that cognitive-

language resources, such as phonological processing skills and rapid automatic naming

speed, of the L2 speakers may be more fully taxed. They are forced to use these resources to develop various aspects o f spoken language concurrently, in addition to

developing reading skills. In summary, although some studies indicate that differences

may be found between LI and L2 children in terms of their phonological processing

profiles, these differences do not appear to be reflected in their ability to learn basic

reading and spelling skills.

33


In a review of the L2 literature, Geva (1998) concluded that a complex interaction of

processes is involved in children’s L2 reading development at both the lower and higher

levels of development. Furthermore, Nassaji and Geva (1999) found that L2 reading

comprehension is an interactive multivariate process in which the underlying cognitive

and language processes are highly intercorrelated. Each of the component processes

contributed to each other and to reading comprehension overall.

The question that arises is what generalizations and conclusions can be made from

comparisons between students assessed in English as their first language (ENG) and

students assessed in English as their second language (ESL)? Results from the ENG

research literature indicate that if ESL students have adequate exposure and instruction in

their second language, strong positive correlations should be expected between parallel

ENG and ESL measures of various component reading skills (e.g., accurate and fast word

recognition; the ability to decode unfamiliar words; efficient text reading efficiency; and

reading comprehension). This finding informs expectations regarding the pace of ESL

students’ reading development. However, less research is available to inform decision

making adequately regarding the expected pace of writing development as compared with

first language speakers (ENG). Furthermore, the research on oral language development

indicates that there may be delays beyond what would be expected given the positive

correlations found between LI and L2 lower- and higher-level reading skills. The present

study offers the opportunity to compare reading, writing and oral language development

for both ENG and ESL students at more advanced stages of reading and writing

development when it would be expected that lower-level skills would be consolidated

(e.g., automatic).

Goals of the Present Study

Previous studies have compared reading, writing, cognitive and language skills for ENG

and ESL students in a comparable multilingual population within Metropolitan Toronto

(Geva, 1998; Panto, 1999; Petrulis-Wright, 1998). The results of those studies suggest

34


comparable literacy skills at grades one and two for ENG and ESL students, with

superior OLP skills for ENG students (Geva, 1998; Petrulis-Wright, 1998) and

comparable spelling and writing skills for grades one to four ENG and ESL students

(Panto, 1999).

Evidence from the reading acquisition literature on second languages has indicated that

OLP is not related to beginning word identification skills (Geva & Petrulis-Wright, 1999)

but it is documented to be related to later comprehension tasks (Clifton & Geva, 1994;

Geva & Ryan, 1993; Verhoeven, 1990). The present study offers the opportunity to

examine the cognitive and language processes that are associated with lower- and higher-

level reading (decoding and reading comprehension) and writing (spelling and story

construction) skills at Grades 3, 5 and 6 within the same sample. It also offers the

opportunity to compare the skills and abilities of ENG and ESL students.

The present study extends current knowledge in several areas. It contributes to the

literature on the relation between lower- and higher-level reading (decoding; reading

comprehension) and writing (spelling; story construction). It contributes to the ESL

literature by investigating lower and higher levels of reading and writing development at

the more advanced stages of literacy development in Grades 3, 5 and 6. It also

investigates which of the many underlying cognitive and language processes associated

with reading and writing achievement explain the most variance for lower and higher

levels of reading and writing skills for both ENG and ESL students. Thus, the following

research questions will be asked.

35


Research Questions:

1. Do ENG and ESL students in grades 3 and 5/6 differ significantly in their

performance on a) reading, b) writing, c) cognitive, and d) language measures?

2. What is the relation between lower-level and higher-level reading and writing

skills across Grade (3, 5/6) and Language (ENG/ESL)?

3. Which cognitive and language composites (Cognitive Ability and OLP) explain

the most variance in the lower (word attack and spelling) and higher levels

(reading comprehension and story construction) of reading and writing

development?

36


CHAPTER TWO: METHOD

In order to address these research questions, two age/grade groups of ENG and ESL

children (grades 3 and 5/6), representing an “earlier” and “later” stage of literacy

development were assessed on a number of reading and writing measures as well as a

number of cognitive and language processing measures.1

Description of Participants and Measures

The present study was part of a large-scale longitudinal project examining the

development of early literacy skills. As a sub-set of this larger investigation, it made use

of data collected in the fall of 1999, and added measures specific to the current research

questions.

Participants

The participants in the present study were children drawn from eighteen public schools in

a lower-middle-class suburb of a large multi-ethnic metropolitan center in Canada. Five

years earlier, these children had been in schools that had been part of a “balanced

literacy” research project in which teachers and school administrators had participated in

professional development (Lloyd, 1993; Willows, 1996).2

At the time of the completion of the present study (1999), the children were in the

beginning of the fall term of their respective grades (3, 5 and 6). Of the 289 children

participating in the study, 110 (34.6 %) were classified as English as a Second Language

(ESL). Children were classified as ESL if they spoke a language other than English at

home and if a language other than English was their first language. As outlined in Table

1 On the average, children in grade 3 and 5/6 would be expected, to be functioning in reading/writing stages 2 and 3, respectively.2 Especially given the context between ENG and ESL, it would be important to ensure that groups were equated on nonverbal intelligence. In the research proposal, two nonverbal measures had been included, but the school board would not approve their use. As an alternative, cognitive measures, such as short-term and working memory, with extremely little language demands, were used.

37


4, the main languages spoken by the ESL learners were Punjabi, Chinese, Arabic, Urdu

and Vietnamese. Other languages groups that were represented included Tamil,

Portuguese, Gujaerati, Patois, Spanish, Italian, German, Korean, Latvian, and French.

Table 4 - Languages Represented in the ESL group

Home Language # of Students Home Language # of Students

Punjabi 51 Gujaerati 2

Chinese 16 Patois 2

Arabic 9 Spanish 1

Urdu 9 Italian 1

Vietnamese 7 German 1

Tamil 5 Latvian 1

Portuguese 4 French 1

In order to compare two samples of children - representative of earlier and later stages of

literacy functioning - two age/grade groups were formed, grade 3 and grade 5/6. On the

average, grade 3 children (assessed near the beginning of the school year) would have

been expected to be functioning in reading and writing Stage 2, in which the lower-level

skills (decoding and spelling) are not yet automatized, whereas grade 5/6 children (also

assessed near the beginning of the school year) would have been expected to be

functioning in reading and writing Stage 3 in which the higher-level skills (reading•3

comprehension and written composition) play a more prominent role. Table 5 presents a

summary of the participant characteristics.

3 Combining data from these later grades made sense because, according to both Chall’s (1996) and Levine’s (1998) stages o f reading and writing development, students in grades 5 and 6 would function similarly. Combining data from grades 5 and 6 also increased the number of ESL students in the “later” literacy stage group. Appendix A provides evidence of the comparability of the 5 and 6 groups on reading, writing, cognitive and language processing measures.

38


Table 5. Participant Characteristics - Fall 1999

Grade 3 Grade 5/6

(n= 118) (n=171)

ENG ESL ENG ESL

Language n = 56 n = 62 n = 123 n = 48

Mean Age 8.45 years 8.33 years 10.76 years 10.95 years

SD Age 3.30 months 3.38 months 6.70 months 6.62 months

Measures

Table 6 provides a list of the measures collected for children in grades 3 and 5/6.

Overall, there were 24 measures (4 reading, 7 writing, 8 cognitive, and 4 language). Four

standardized reading tests provided measures of word attack, word recognition, fluency

and reading comprehension. The seven writing measures were obtained from two

sources, a standardized test of spelling provided one measure and a writing sample taken

in response to a picture prompt provided six measures of components reflecting different

aspects of writing. These included contextual conventions, contextual language, story

construction and understanding of story schema, as well as number of words written

(productivity) and the number of spelling errors (spelling errors). The eight cognitive

processing measures included measures of phonemic awareness, rapid automatic naming,

short-term auditory memory, working memory, sequencing, as well as memory for

words, sentences and stories. The four language processing measures included measures

of receptive vocabulary, expressive vocabulary, grammatical judgment and listening

comprehension. On the basis of a principal components analysis, an oral language

proficiency (OLP) composite score was created, combining sentence memory from the

cognitive battery with the four language measures.

39


Table 6. Test Measures

Type Name Skill

Reading WRMT-R Word Attack * Pseudoword decoding

WRAT-III Reading * Word recognition

Biemiller Test of Reading Processes Reading fluency (time)

Gates-MacGinitie * Reading comprehension

Writing WRAT-III Spelling *

TOWL-III Writing Sample

Spelling

Contextual Conventions (TOWL-III) * Capitalization/punctuation

Contextual Language (TOWL-III) * Sentence structure/grammar

Story Construction (TOWL-III) * Plot/setting/character/theme

Productivity Number of words written

Spelling errors Number of spelling errors

Story Schema (Nodine, 1985) Elements of story structure

Cognitive Test of Auditory Analysis Skills (TAAS) Phonemic awareness

RAN (CELF-III) * Rapid Automatic Naming

Sequences (CMS) * Sequencing

Numbers Forward (CMS) * Short-term memory

Numbers Backward (CMS) * Working memory

Verbal Learning (WRAML) * Memory for words

Sentence Memory (WRAML) * Memory for sentences

Story Memory (WRAML) * Memory for stories

Language PPVT-III * Receptive vocabulary

EVT* Expressive vocabulary

Formulated Sentences (CELF-III) * Grammatical judgment/syntax

Durrell Analy sis of Reading Difficulty Listening comprehension

* Asterisks denote standardized tests


Reading Measures

Of the 4 reading measures, three are representative of lower-level reading skills: word

attack, word recognition and reading time. The fourth measure is representative of

higher-level reading skill: reading comprehension. Each of these measures is discussed

in turn.

Lower-level reading skills

Pseudoword Decoding. The Word Attack subtest of the Woodcock Johnson Reading

Mastery Test (WJRMT) was used to measure decoding skills. The child is asked to read

pronounceable pseudowords, or non-words (e.g., “ift”, “mancingful”) that can be

decoded by using English grapheme-phoneme conversion rules. The test consists of 45

items, and is terminated when the child makes six consecutive errors.

Word Recognition. The word recognition subtest of the Wide Range Achievement Test -

Third Edition (WRAT-3) (Jastek & Wilkinson, 1995) was used to measure context free

word recognition. The child is asked to read a list of 42 isolated words. The list starts

with simple words (e.g., “in”, “cat”, “book”) and progresses to more difficult words (e.g.,

“stretch”, “horizon”, and “itinerary”). The test is terminated when a child reads

incorrectly ten consecutive words. The maximum score attainable on this test is 57.

Reading Speed. The Biemiller Test of Reading Processes is a timed reading task that

assesses the student’s ability to use context to facilitate word identification. The student

is asked to read the passage as quickly as they can. The maximum score is the time (in

seconds) that it takes for the student to read the passage.

41


Higher-level reading skill

Reading Comprehension. The Gates MacGinitie Reading Comprehension Test was

administered to assess reading comprehension. The child is instructed to pay close

attention to the stories as he/she is reading, because he/she will be asked questions about

the story. The child is not provided with any assistance while reading. When the child

has completed reading the passage, they must read and answer the questions that follow

the story. It is important to note that two different test booklets were used. The grade 5

and 6 level booklet had more difficult processing and language demands than at the grade

3 level booklet. The maximum score for both the grade 3 and the grade 5/6 booklets is

48.

Writing Measures

Two writing measures were collected. The first measure, WRAT spelling, was chosen as

a representative lower-level writing skill. The second measure, a writing sample taken

lfom the TOWL-III, was chosen specifically to look at the different components of

writing in a standardized fashion. However, given the biases of standardized norms with

an ESL population, raw scores were used. There are scoring measures that represent

lower level skills, such as the conventions of writing (punctuation and capitalization), the

language of writing (sentence structure and grammar), productivity (number of words),

and number of spelling errors. Scoring measures representing higher-level skills include

story construction (the cognitive component of what a story is made of) which involves

understanding the importance of a beginning, middle, and end to a story structure as well

as the presence of conflict, and the resolution of that conflict for plot development. In

addition, a measure called story schema, looks specifically at how the student is able to

represent their understanding of a story. Each of these scoring measures is discussed in

turn.

Writing Sample. A standardized writing sample was collected using the Test of Written

Language - Third Edition (TOWL-III). The children were given 45 minutes to write a

42


story that went with the visual prompt. According to instructions, stories were only

scored if they contained more than 40 words. The writing sample was then scored

according to the last three subtests (the first two represent lower level skills, and the last

represents higher level skill) of the TOWL-III. Descriptions follow:

Lower-level writing skills

Spelling. The spelling subtest of the Wide Range Achievement Test - Third Edition

(WRAT-3) (Jastek & Wilkinson, 1995) was used to measure context free spelling. The

child is asked to spell a list of 42 isolated words. The list starts with simple words (e.g.,

“in”, “cat”, “book”) and progresses to more difficult words (e.g., “stretch”, “horizon”,

and “itinerary”). The test is terminated when a child spells incorrectly ten consecutive

words. The maximum score on this test is 55.

Contextual conventions. This subtest of the TOWL-III measures mastery of the arbitrary

conventions of written language: punctuation, spelling and capitalization. A score is

assigned out of 18.

Contextual language. This subtest measures the use of language in writing, paying

particular attention to sentence structure, grammar and vocabulary. A score is assigned

out of 29.

Higher-level writing skills

Story construction. This subtest evaluates the student’s use of prose, action, sequencing

and theme. A score is assigned out of 21. It measures the cognitive components of story

writing in that the student’s understanding of what constitutes a story (beginning, middle,

end, conflict and resolution of conflict) is assessed.

Three unstandardized measures supplemented the previously reported standardized

measures from the TOWL-III. A simple word count was taken as a measure of

43


Productivity - the number of words written in a 45 minute time period. In addition, a

simple count of the number of spelling errors that occurred within the writing sample was

taken as a measure of the students’ spelling errors made within the context of their

writing. The third measure was taken from research on children’s development of a story

schema.

Story schema. This measure assessed children’s understanding of the requirements of a

story. To study the development of children’s writing, researchers have developed

category systems to analyze and classify children’s written compositions. The four

categories used as a measure of story schema in this study are based on the work of

Nodine et al. (1985) whose four categories used to classify story development are

expressive, descriptive, story-like and story. Inter-rater reliability was .87 for

composition classification. Children’s writing samples were scored on the basis of which

category best described their stories as follows:

Expressive: A writing sample was judged to be expressive if it did not contain a story line or narrative. Both this category and the next one, descriptive writing, lack a story line, but unlike descriptive writing, the expressive writing does not relate to the story in any manner.

Descriptive: A writing sample was judged to be descriptive if it described a picture, person, place or thing, but failed to integrate them into a story. There was no evidence of story schema.

Story-Like: A writing sample was judged to be “story-like” if it identified a setting but failed to describe a complication or resolution.

Story: This category was used to describe writing samples that contained a setting, a conflict, and a resolution that related to one another. A story schema was evident.

Cognitive Processing Measures

Eight processing measures were collected on the basis of the research findings. First they

were evaluated separately as to whether there were significant differences between the

44


first language and second language speakers on each of these measures. Then, they were

put into a factor analysis to see how they loaded into factors, in an attempt to simplify the

regression analyses. These measures loaded onto two factors: a short-term auditory

memory factor (consisting solely of the numbers forward subtest) and a cognitive ability

factor which included the TAAS (Rosner), RAN, working memory and sequencing

measures. The factor analysis follows in the Results section.

Test of auditory analysis skills (TAAS; Rosner & Simon. 197IT The TAAS taps into a

child’s syllable and phoneme awareness. The first few items require the segmentation of

words into syllables, but the task becomes increasingly difficult as participants are asked

to perform more complex phoneme manipulations including the deletion of initial,

medial, and final sounds and the splitting of consonant blends. There were two trial

demonstration words, followed by test items. The child was asked to repeat a word

spoken by the examiner, and then repeat it again but without saying a particular syllable

or phoneme (e.g., “ Say ‘dog’. Now say it again without the /d/”). Testing was

discontinued after two consecutive errors and the child’s score was recorded as the item

number prior to the two errors (consistent with the TAAS instructions). The child’s score

was the number of items he/she got correct ranging from 0 to 22. The extended version

of the TAAS, items 14 to 22, was used to provide items that would challenge the students

in grades 5 and 6. This task has been shown to have a moderate to high reliability

(a=.78) using Cronbach’s alpha (Yopp, 1988).

Rapid automatized naming (RAN). Tests of rapid automatic naming speed are designed

to measure how automatically children recode visual information into a phonologically

represented name code. The Rapid Automatic Naming task (RAN) task used in this

study recorded how rapidly and accurately children identified a series of geometric

shapes presented in a series of 10 in 5 rows. This task was taken from the Clinical

Evaluation of Language Fundamentals - Third Edition (CELF-III). The maximum score

on this test is how many seconds it takes the child to name the objects.

45


Sequencing. The Sequences subtest of the Children’s Memory Scale (CMS) was

administered as a measure of sequencing and working memory because, like the memory

for numbers backwards subtest, the child was required to hold information in his or her

memory and manipulate it before recalling it. However, this subtest differs in that it also

requires the child to recall sequences of information, from well-learned sequences such as

the alphabet, to sequences that required the child to add numbers by 3’s. The maximum

score on this test was 84.

Short-term auditory memory. The Memory for Numbers Forwards subtest of the

Children’s Memory Scale (CMS) was administered as a measure of short-term auditory

memory. Each child was asked to remember a series of numbers of increasing length.

The maximum score on this test was 16.

Working memory. The Memory for Numbers Backwards subtest of the Children’s

Memory Scale (CMS) was administered as a measure of working memory. Each child

was required to remember a series of numbers of increasing length, manipulate them in

memory and recall them in a backward order.

Memory for words. The Verbal Learning subtest of the Wide Range Assessment of

Memory and Learning (WRAML) was administered as a measure of memory for words.

Each child was required to recall a long list of words over four presentations with the

opportunity to learn the words with repetition. For children 8 years and younger (grade

3), the maximum score is 52, and for children 9 years and up (grade 5/6), the maximum

score is 64.

Memory for sentences. The Sentence Memory subtest of the WRAML was administered

as a measure o f memory for sentences. Each child was required to recall a number o f

sentences of increasing length and complexity. Memory for sentences has been found to

distinguish between LI and L2 learners because of the heavy language/syntax demands

inherent in the task. This measure was collected as part of the Cognitive battery, but as

an individual subtest, it loaded heavily on the Oral Language Proficiency factor and was

46


therefore included in the composite. The maximum score on this subtest is 40 regardless

of age.

Memory for stories. The Story Memory subtest of the WRAML was administered as a

measure of memory for stories. Each child was required to recall as many details in two

stories that they were able to. This subtest assessed children’s ability to recall verbal

detail with the benefit of context. The maximum score for children 8 years and younger

(grade 3) is 45, and the maximum score for children 9 years and up (grade 5/6) is 51.

Language Measures

Four language measures were collected. Their contribution was assessed separately as

well as in an Oral Language Proficiency (OLP) composite. This composite was

constructed on the basis of a principal components analysis on which these five items

loaded heavily.

Receptive vocabulary. The Peabody Picture Vocabulary Test - Third Edition (PPVT-III)

is a measure of receptive vocabulary. Each child was shown four pictures on a page (e.g.,

“dog”, “brush”, “chair”, “car”) and was then asked to point to one item (e.g., “can you

point to the picture of a chair?”). The test consists of 204 words of increasing difficulty.

The test is discontinued when the child responds incorrectly to eight items in a block of

twelve questions. The maximum score on this test is 204.

Expressive vocabulary. The Expressive Vocabulary Test (EVT) is a measure of

expressive vocabulary. Each child was shown a picture (e.g., “bus”) and was then asked,

“Can you tell me what this is a picture of ”? The test consists of 190 items of increasing

difficulty. The test is discontinued when the child responds incorrectly to 5 consecutive

items.

Syntax/grammatical judgment. The Formulated Sentences subtest of the CELF-III was

administered as a measure of syntax and grammatical judgment. Many studies have

47


shown that this skill discriminates between LI and L2 learners. The maximum score on

this test is 22.

Listening comprehension. The Durrell Analysis of Reading Difficulty (Durrell, 1970)

assesses listening comprehension. The test is comprised of three short stories (about a

paragraph in length) that are read aloud to the child. Each child was instructed to pay

attention while the researcher read the story and then was asked questions about the story.

Following each story the child was asked to retell the story. This was followed by 5

questions - 4 factual questions and one requiring inference. The maximum score on this

test is 6 (corresponding to attributed grade level).

The Sentence Memory subtest of the WRAML, as described in the Cognitive Measures

section, was administered as a measure of memory for sentences. As previously

mentioned, memory for sentences has been found to distinguish between LI and L2

learners because of the heavy language/syntax demands inherent in the task. Although

this measure was collected as part of the Cognitive battery, it loaded heavily on the Oral

Language Proficiency composite.

48


Procedures

As indicated earlier, this research comprised a sub-study within the context of a larger

project, assessing the development of literacy skills. As part of the larger project,

children were tested on a variety of standardized reading and writing measures. As part

of the current study, children were also assessed on a variety of cognitive and language

processing measures. With the exception of the reading comprehension task and the

writing sample that were administered within the class setting, the children were seen

individually, in a quiet setting, by members of the research team (graduate students in

psychology and education). The walls of all testing rooms were purposely bare, without

letters, words or writing samples on display that might interfere with the accuracy of the

testing results.

To avoid bias associated with using norms standardized on LI populations, standardized

test (e.g., WRAT-III, word recognition) raw scores were not converted to percentiles or

standard scores. Instead, all analyses for both standardized and informal tasks were

based on raw scores.

49


CHAPTER THREE: RESULTS AND DISCUSSION

Overview

The results are presented in three major parts corresponding to the three main research

questions. The first part examines whether any significant differences exist between

ENG and ESL students on the reading and writing academic measures, and the cognitive

and language processing measures. The second part examines the relation between

lower- and higher-level reading and writing measures by Grade (3, 5/6) and by Language

(ENG/ESL). The third part examines how much variance in the lower- and higher-level

reading and writing skills is accounted for by the underlying cognitive and language

processes.

Variable Screening

As previously mentioned, analyses were performed on raw scores (i.e., standardized test

raw scores were not converted to percentiles or standard scores) to avoid problems

associated with using norms standardized on English populations for ESL children. Prior

to analysis, all variables were examined for accuracy of data entry, missing values,

normalcy of distributions and other assumptions of univariate and multivariate analyses

(using SPSS 9.0). When variables were found to possibly violate the normality

assumption, two sets of analyses were performed. Analyses were run on both

transformed (via square root and logarithmic transformations which brought the

distributions within acceptable norms) and non-transformed variables. Differences in the

results of these analyses were negligible, with no change in the level of significance.

Therefore, analyses using non-transformed variables will be discussed, since

interpretation o f these results is more meaningful.

Regression analyses, following the recommendation of Cohen and Cohen (1983) (see

also Aiken & West, 1991; Holbeck, 1997), involved the use of z scores. The independent

variables used in the interaction terms within the multiple sequential regression (i.e.,

50


Language X OLP) needed to be centered from the mean prior to their inclusion in the

regression analyses. The use of z scores addresses the potential problem of

multicollinearity that could result from this inclusion of the independent variables and

their interaction terms in the regression analyses.

Similarities and differences in ENG and ESL students on reading, writing, cognitive

and language measures

The first of the four major questions concerned similarities and differences in the skills

and abilities of ENG and ESL students. This section presents the results addressing this

first question. The question is stated as follows:

1. Do ENG and ESL students in grades 3 and 5/6 differ significantly in their

performance on a) reading, b) writing, c) cognitive, and d) language measures?

Preliminary analyses were first performed to determine whether it would be legitimate to

combine the results from grades 5 and 6 , since children in both grades would,

theoretically, represent the same stage of literacy development in the reading and writing

stages frameworks used to guide the research. 2 x 2 factorial ANCOVAs were conducted

with 2 levels of Grade (5, 6 ) and 2 levels of Language (ENG, ESL) to determine the

similarities in the patterns of results for grades 5 and 6 . Based on these analyses the

decision was made to combine grades 5 and 6 . The results for the combined grade 5/6

group were in the same direction as those for the grade 5 and 6 groups separately, and

combining the two grades increased the sample size for the ESL group. Table A -1 in

Appendix A summarizes the means, standard deviations, maximum scores, F, p and f

values for all of the measures in grades 5 and 6 by Language (ENG and ESL).

ANCOVAs were undertaken to ensure that age was not contributing to the pattern of

results, given the wide age range in the combined grade 5/6 group. The primary analyses

did not control for chronological age because the purpose of the research was to evaluate

development according to stage theory, and the contribution of chronological age was

established in the preliminary analyses.

51


The primary analyses addressing the first main research question involved 2 X 2 factorial

ANOVAs. These were performed for each measure by Grade (3, 5/6) and by Language

(ENG/ESL) in order to determine whether significant differences existed on the reading,

writing, cognitive and language measures of ENG and ESL students.4 The ANOVAs

were followed by planned comparisons between Grade (3, 5/6) to see where the

significant differences existed. Table 7 summarizes the means, standard deviations,

maximum scores, F,p and rj2 (i.e., effect size) values for all of the measures (by Grade

and Language) and indicates comparisons on which the performance of the ENG and

ESL groups differed significantly.

Eta square {ij2) effect sizes were calculated as an alternative to d. Eta square ranges in

value from 0 to 1, as d values do. An ij2 value of 0 indicates that there are no differences

in the mean scores among groups. A value of 1 indicates that there are differences

between at least two of the means on the dependent variable and that there are no

differences on the dependent variable scores within each of the groups (i.e., perfect

replication). In general, tj2 is interpreted as the proportion of variance of the dependent

variable that is related to the factor. Traditionally, rj2 values of .01, .06, and .14 represent

small, medium and large effect sizes, respectively.

The findings on the reading, writing, cognitive and language measures will be discussed

in order.

4 This procedure was chosen instead of MANOVAs because there were multiple Reading, Writing, Cognitive and Language measures that were clearly correlated and not independent. When the Dependent Variables are correlated, they measure the same facets of behavior in slightly different ways. Since a MANOVA creates a composite DV that maximizes group differences, not much would be gained by the inclusion of several measures of the same thing. Even moderately correlated Dependent Variables diminish the power of the MANOVA (Tabachnik & Fidell, 2001).

52


Reading Measures

Significant main effects of Grade were found for Word Attack, F (4, 288) = 15.89,/? <

.001, WRAT Reading, F (4,288) = 86.67,/? < .001, and for Reading Time, F (4,284) =

7.53,/? < .0 1 , indicating that the scores on these different reading measures increased

with age. Thus, the grade 5/6’s read faster than the grade 3’s and were more fluent and

automatized. Significant main effects of Grade were not found for Reading

Comprehension.5

Surprisingly, there were no significant main effects of Language (ENG, ESL), nor were

there any significant interactions with Language. This means that on both lower-level

(word decoding, word recognition, reading time) and higher-level (reading

comprehension) reading tasks, there were no significant differences between the ENG

and ESL students. One might have expected differences between the ENG and ESL

students in the time it took them to read graded passages and on the higher-level reading

comprehension task with its more complex language demands. This was not the case.

5 The comparison of Grade on the reading comprehension task is meaningless because different tests were involved at grade 3 and 5/6. The actual raw scores are very similar between the grade 3 and 5/6 tests because the tests are designed to be appropriately difficult for the age group. The test was the same for grade 5 and 6.

53


Table 7.

Group Comparisons on Reading, Writing, Cognitive and Language Processing Measures (based on raw scores)

GRADE3

GRADES5/6

ENG ESL F P< ENG ESL F P K

n = 56 n = 57 n= 123 n = 48Reading Measures

Word Attack Maximum score = 45

X = 26.71 S.D.= 8.21

X = 27.25 S.D. = 9.62

.386 n/s .003 X = 30.50 S.D. = 7.67

X = 31.65 S.D. = 7.40

.632 n/s .004

WRAT Word Recognition Maximum score = 57

X = 31.51S.D. = 3.98

X = 31.46 S.D. = 4.61

.037 n/s .0 0 0 X = 36.60 S.D. = 4.98

X = 37.21 S.D. = 4.30

.161 n/s .0 0 1

Reading Comprehension Maximum score = 48

X = 28.21 S.D. = 9.32

X = 26.46 S.D. = 10.17

.422 n/s .004 X = 27.46 S.D. = 9.39

X = 25.56 S.D. = 8.06

2.24 n/s .013

Reading Time (seconds) X = 51.60 S.D. = 19.93

X = 56.65 S.D. = 33.08

.414 n/s .004 X = 48.63 S.D. = 26.43

X = 44.00 S.D. = 14.03

.419 n/s .003

Rep

rodu

ced

with

pe

rmis

sion

of

the

copy

right

ow

ner.

Furth

er

repr

oduc

tion

proh

ibite

d w

ithou

t pe

rmis

sion

.

GRADE3

GRADES5/6

ENG ESL F P< r ENG ESL F P<

Writing Measures

WRATSpellingMaximum score = 55

X = 25.57 S.D. = 3.03

X = 26.72 S.D. = 4.18

3.69 p=. 06 .03 X = 31.45 S.D. = 5.64

X = 33.60 S.D. = 4.56

3.88 .05 .0 2

TOWL-III Spelling Errors (#) X = 10.45 S.D. = 5.74

X = 9.26 S.D. = 7.51

.806 n/s .0 0 0 X = 9.12 S.D. = 7.26

X = 6.87 S.D. = 4.29

3.19 p=. 07 .0 2

TOWL-IIIContextual Conventions Maximum score = 18

X = 2.64 S.D. = 2.48

X = 3.46 S.D. = 2.62

3.31 p=. 07 .029 X = 4.88 S.D. = 3.20

X = 5.12 S.D. = 2.98

.0 1 0 n/s .0 0 0

TOWL-IIIContextual Language Maximum score = 29

X = 9.73 S.D. = 3.16

X = 9.95 S.D. = 3.41

.084 n/s .0 0 1 X = 13.11 S.D. = 4.04

X = 13.14 S.D. = 3.47

.065 n/s .0 0 0

TOWL-III Story Construction Maximum score = 21

X = 10.71 S.D. = 3.20

X = 10.45 S.D. = 3.55

.0 0 2 n/s .0 0 0 X = 11.99 S.D. = 3.92

X = 12.37 S. D. = 3.18

.223 n/s .0 0 1

Story Schema Maximum score = 4

X = 3.22 S.D.= .95

X - 2.96 S.D. = 1.04

.647 n/s . 0 0 0 X = 3.47 S.D. = .90

X = 3.44 S.D. = .81

.006 n/s .0 0 0

Productivity (# of words) X = 103.49 S.D. = 43.50

X = 99.67 S.D. = 41.89

.025 n/s .0 0 0 X = 134.96S.D. = 45.89

X = 148.52 S.D. = 42.92

.2 2 1 n/s .014

Rep

rodu

ced

with

pe

rmis

sion

of

the

copy

right

ow

ner.

Furth

er

repr

oduc

tion

proh

ibite

d w

ithou

t pe

rmis

sion

.

GRADE3

GRA51

DES6

ENG ESL F P K V2 ENG ESL F P<Cognitive Measures

TAASMaximum score = 22

X = 14.41 S.D. = 5.75

X = 15.29 S.D. = 5.74

.529 n/s .005 X = 15.65 S.D. = 5.64

X = 16.51 S.D. = 5.57

.487 n/s .003

RAN (seconds) X = 36.71 S.D. = 10.38

X = 35.59 S.D. = 9.75

.910 n/s .009 X = 30.46 S.D. = 9.31

X = 25.66 S.D. = 3.95

9.62 .0 0 1 .05

CMS - Sequences Maximum score = 84

X = 45.11 S.D. = 9.75

X = 44.31 S.D. = 11.88

.028 n/s .0 0 0 X = 53.47 S.D. = 10.28

X = 57.08 S.D. = 9.02

3.31 p=. 07 .019

CMS - Numbers Forward Maximum score = 16

X = 7.61 S.D. = 1.58

X = 7.63 S.D. = 1.91

.033 n/s .0 0 0 X = 8.46 S.D. = 1.79

X = 9.00 S.D. = 1.55

3.02 n/s .018

CMS - Numbers Backward Maximum score =14

X = 3.88 S.D. = 1.27

X = 4.14 S.D. = 1.55

1.29 n/s . 0 1 2 X = 4.55 S.D. = 1.58

X = 4.83 S.D. = 1.49

.681 n/s .004

WRAML - Verbal Learning Gr. 3 Maximum score = 52 Gr. 5/6 Maximum score = 64

X = 25.42 S.D. = 6.08

X = 25.94 S.D. = 6.41

.248 n/s .0 0 2 X - 31.78 S.D. = 7.45

X = 32.25 S.D. = 9.14

.032 n/s .0 0 0

WRAML - Sentence Memory Maximum score = 40

X = 13.69 S.D. = 3.58

X = 12.49 S.D. = 3.48

1 .8 8 n/s .017 X = 17.35 S.D. = 4.20

X = 16.44 S.D. = 4.34

2.29 n/s .014

WRAML - Story Memory

Gr. 3 Maximum score = 45 Gr. 5/6 Maximum score =51

X = 23.92 S.D. = 7.05

X = 23.07 S.D. = 7.36

.135 n/s .0 0 1 X = 26.04 S.D. = 7.84

X = 24.60 S.D. = 7.43

1.89 n/s .011

Rep

rodu

ced

with

perm

issi

on

of the

co

pyrig

ht

owne

r. Fu

rther

re

prod

uctio

n pr

ohib

ited

with

out

perm

issi

on.

GR/U)E GRA15/1

DES5

ENG ESL F P < v 2 ENG ESL F PKLanguage measures

Receptive Vocabulary Maximum score = 204

X = 116.23 S.D. = 22.18

X = 106.94 S.D. = 19.75

3.02 p=.08 .0 2 X = 142.92 S.D. = 17.14

X = 132.81S.D. = 18.47

14.45 .0 0 1 .08

Expressive Vocabulary Maximum score =190

X = 77.71 S.D. = 14.16

X = 76.40 S.D. = 11.19

.0 0 0 n/s .0 0 0 X = 98.98 S.D. = 15.72

X = 94.47 S.D. = 10.79

4.00 .05 .0 2

SyntaxMaximum score = 22

X = 13.11 S.D. = 3.07

X = 12.45 S.D. = 3.56

.713 n/s .007 X = 15.69 S.D. = 3.46

X = 14.39 S.D. = 3.60

5.11 .05 .03

Listening Comp. Maximum score = 6

X - 3.46 S.D. = .80

X = 3.43 S.D. = .88

.003 n/s .0 0 0 X = 4.39 S.D. = 1.06

X = 4.04 S.D. = 1.00

3.91 .05 .0 2

Rep

rodu

ced

with

pe

rmis

sion

of

the

copy

right

ow

ner.

Furth

er

repr

oduc

tion

proh

ibite

d w

ithou

t pe

rmis

sion

.

Writing Measures

Significant main effects of Grade were found on all of the writing measures, as follows:

WRAT Spelling, F(A, 288) = 114.74,/? < .001; Spelling Errors, F (4, 288) = 5.46,/? <

.05; Productivity, F (4,288) = 51.76,/? < .001; Contextual Conventions, F (4,288) =

27.42,/? < .001; Contextual Language, F (4,288) = 49.58,/? < .001; Story Construction,

F (4,288) = 12.13,/? < .001; and Story Schema, F (4,288) = 8.92,/? < .01. These results

indicate that grade 5/6’s spell more accurately than grade 3’s, and that grade 5/6’s also

write longer stories with more developed punctuation, sentence structure and story

structure than grade 3’s do.

The only significant main effect of Language found on the seven writing measures

favored ESL on WRAT Spelling, F( 4, 288) = 9.58,/? < .01. No significant interactions

were found. Planned comparisons between Grade 3 and 5/6 revealed that it was the

Grade 5/6 ENG and ESL students who differed significantly in their spelling, favoring

ESL, F (2, 170) = 3.88, p < .05, with a small effect size of rj2 = .03. For graphic

representation and discussion of these findings, see Appendix B.

Cognitive Measures

Significant main effects of Grade were found on six of the eight cognitive measures, with

trends on the two remaining measures, as follows: RAN, F (4,279) = 54.90,/? < .001,

Sequences, F (4,278) = 64.37,/? < .001, Numbers Forward, F (4,279) = 24.53,/? < .001,

Numbers Backward, F (4, 278) - 12.36,/? < .001, Memory for Words, F (4, 279) = 45.16,

p < .001, Sentence Memory, F (4,278) = 4.87,/? < .05, and a trend on Story Memory, F

(4,278) = 3.57,/? = .06, and the TAAS, F (4,278) - 2.95,p = .08. These results indicate

that the grade 5/6’s were able to rapidly name geometric shapes, sequence information,

remember numbers in both a forward and backward sequence, and remember more words

and sentences than grade 3’s.

58


The only significant main effect of Language, favoring ESL students, was found for

RAN, F (4, 279) = 7.66, p < .01, and no significant interactions were found. This result

indicates that ESL students were able to name geometric shapes faster than ENG

students. Planned comparisons between Grade (3, 5/6) revealed that the significant

difference between ENG and ESL students was again at the grade 5/6 level, F (2, 170) =

9.63,p < .001, ij2 — .05. This result may in fact help to interpret the spelling differences

favoring ESL children as research supports a positive relation between spelling and rapid

automatic naming (Denckla & Rudel, 1974,1976 and Wolf & Bowers, 1999). RAN

measures how automatically children recode visual information into a phonologically

represented name code (Denckla & Rudel, 1974, 1976), and such tasks have been shown

to contribute specific variance to orthographic processing (Bowers, 1995). For graphic

representation and discussion of these findings, see Appendix B.

Language Measures

Significant grade effects were found for Receptive Vocabulary, F (4, 280) = 123.86,

p < .001, Judgment of Syntax, F (4,278) = 25.98,/? < .001, Expressive Vocabulary,

F (4, 280) = 127.59,/? < .001, and Listening Comprehension, F (4, 279) = 37.94,

p < .001. Thus, grade 5/6’s have a larger receptive and expressive vocabulary than grade

3’s. Grade 5/6’s are also better able to formulate sentences with proper grammatical

syntax, and possess a more advanced listening comprehension than grade 3’s.

In addition to significant Grade effects, significant main effects of Language, favoring

ENG students, were found for Receptive Vocabulary, F (4, 280) = 19.34,/? < .001, and

for Judgment of Syntax, F (4, 278) = 4.87, p < .05. A trend was found in the same

direction for Expressive Vocabulary, F (4,280) = 3.29, p = .07. As would be expected,

ENG students have both a larger store of receptive vocabulary, and a more advanced

ability to formulate sentences and judge grammatical syntax. No significant interactions

were found. Planned comparisons revealed that the significant differences between ENG

and ESL students were found at grade 5/6, although the pattern was similar in grade 3.

For discussion and graphic representation of these results, see Appendix B.

59


Given the significant differences found in these oral language proficiency measures

between ENG and ESL students, to the advantage of ENG students, it is important to

highlight the lack of differences that were found on the academic measures, at both the

lower- and higher-levels of reading and writing development, as will be addressed in the

next section.

The Relation Between Lower- and Higher-Level Reading and Writing Skills

The second of the three major questions concerned the relation between reading and

writing in ENG and ESL students. The results addressing this question are presented in

this section. The question is stated as follows:

2. What is the relation between lower-level and higher-level reading and writing

skills across Grade (3, 5/6) and Language (ENG/ESL)?

Partial correlations were undertaken controlling for chronological age, because the grade

5 and 6 samples were combined, and it was important to ensure that the results were not

unduly influenced by chronological age.6 Preliminary correlational matrices were

undertaken for all of the reading and writing measures. Results indicated a strong

correlation between reading and spelling across Grade and Language. For discussion and

presentation of the results, see Appendix C.

Tables 8 presents the partial correlations between lower- and higher-level reading and

writing measures by Grade (3, 5/6) and by Language (ENG/ESL). Using the Bonferroni

approach to control for Type 1 error across the many correlations, a p- value of less than

.005 (.05/10) was required for significance. In the table, the correlations that reached this

level of significance are bolded.

6 A comparison of the same correlations between reading and writing measures, by Grade and Language, without controlling for age showed very similar results with almost exactly the same significance levels. Therefore, partial correlations were used to most accurately reflect the relation between the measures, without the effect o f chronological age.

60


Table 8.

Partial Correlations Between the Lower- and Higher-Level Reading and Writing Skills by Grade (3,5/6) and by Language (ENG/ESL)

Grade 3 Grac e 5/6ENG ESL ENG ESL

Word Attack and Spelling .74 .77 .72 .52

Reading Comprehension and Story Construction

.31 .54 .51 .35

Note: Based on Bonferrotvi adjustment,/? <.001 in bold, For ENG, r = .31,p < .05; For ESL, r = .35,p < .05

These correlations convincingly demonstrate the strong relation between lower-level

word attack (pseudoword decoding) and spelling across Grade and Language. The

relation between word attack and spelling remained consistently strong for both ENG and

ESL across Grade; however, the relation between higher-level reading comprehension

and story construction was more variable.

For ENG, the relation between higher-level reading comprehension and story

construction was not significant at the p < .001 level in grade 3, but, was significant in

grade 5/6 (r = .51 ,P< .001). This pattern would be expected given that the consolidation

of lower-level reading and writing skills would be in the lower grade (stage) in order to

further the development of higher-level skills in the higher grade (stage). It is in grades 5

and 6 that the ENG students would have consolidated both the lower-level reading and

spelling skills, as well as the complex linguistic sophistication required for the higher-

level reading and writing skills. Interestingly, for ESL, the correlation between higher-

level reading comprehension and story construction was moderately significant in grade 3

(r = .54, p < .001) and not significant atp< .001 in grade 5/6. Possible interpretations of

this finding include that in grade 3, the test items may be fairly common whereas in grade

5/6, the items become more difficult and go beyond what would ordinarily be expected.

The language demands in grade 5/6 thus become more complex and abstract, whereas in

grade 3, the language demands are not as complex, and are more literal. It is possible that

the correlation between higher-level reading comprehension and story construction is

61


significant for ESL students in grade 3 and not in grade 5/6 precisely because the

language demands of the tasks are less complex. It may be that it takes the ESL students

longer to develop the degree of linguistic sophistication that ENG students do in grades 5

and 6 (as evidenced in the significant differences found in OLP measures), and this is

reflected in the relation between their higher-level reading and writing skills. The likely

reason as to why there are significant differences on oral language measures and not on

written ones, is that the language demands are greater on the oral language measures

(e.g., receptive vocabulary, expressive vocabulary, and judgment of syntax) than they are

on the written tests (e.g., reading comprehension and story construction). It may be that

the written language tasks used in the current study did not adequately tap the linguistic

differences that exist between the ENG and ESL students.

Given the differences between ENG and ESL in the relation between lower- and higher-

level reading and writing tasks, it is valuable to explore how much of the variance in

lower- and higher-level reading and writing is accounted for by the underlying cognitive

and language processes.

Variance accounted for in lower- and higher-level reading and writing

The last of the three major research questions concerns the prediction of reading and

writing skills from the underlying cognitive and language processes. The question

follows:

3. Which cognitive and language composites (Cognitive Ability and OLP) explain

the most variance in the lower (word attack and spelling) and higher levels

(reading comprehension and story construction) o f reading and writing

development?

Preliminary correlational matrices for reading and writing measures, with cognitive and

language measures were undertaken across Grade (3, 5/6) and Language (ENG/ESL). In

62


summary, the major findings indicated that although both cognitive and language

processes were represented across the measures, a pattern appeared in which mainly

cognitive processes were correlated with lower-level word attack for both ENG and ESL,

and predominantly language measures were correlated with higher-level reading

comprehension across Grade and Language. For more details of these results, see

Appendix D.

A principal component factor analysis was run in order to reduce the large number of

reading, writing, cognitive and language variables into a smaller number of components.

This procedure also allowed for the discovery of those variables that formed coherent

subsets that were relatively independent of each other. The variables that were correlated

with one another, but largely independent of other subsets of variables, were then

combined into factors6. Those factors, as will be discussed, are thought to reflect

underlying processes that have created the correlations among variables. The most

commonly used method of rotation, Varimax, was used after extraction of these

components, in order to maximize the variance. The factors that emerged for the reading

and writing measures using this procedure, were a strong Decoding/Encoding factor

(word attack, word recognition, spelling, reading time and reading comprehension), an

Age factor (chronological age), and a Story Schema factor (Story Construction and Story

Schema). For the cognitive and language processing measures, three factors emerged: A

Cognitive Ability factor (TAAS, RAN, Sequencing, and Working Memory), a Short-term

Memory factor (numbers forward) and an Oral Language Proficiency factor (Receptive

and Expressive Vocabulary, Syntax, Listening Comprehension and Sentence Memory).

Table 9 shows the respective loadings for the reading and writing measures.

The first factor was named the Decoding/Encoding Factor because it contained measures

pertaining to those written language processes. Interestingly, it included all four reading

measures (both lower and higher level skills) as well as the lower-level Spelling measure.

6 Technically, Principal Component Analysis produces components while Factor Analysis produces factors. Since these two statistical techniques are both applied to a set of variables for the purpose of discovering which variables form subsets, for ease of discussion, the results will be called factors.

63


The second factor was named Age because it involves the effect that chronological age

had on reading and writing performance. This was to be expected given the nature of the

sample in which data was collected for three grades. The third factor was named the

Story Schema factor because it was comprised of the two writing measures that tap

higher-level cognitive understanding of stories, Story Construction and Story Schema.

Notably, this factor separated out from the other writing measures, such as Spelling,

Contextual Conventions and Contextual Language.

Table 9 - Factor Loadings of the Reading and Writing Measures - Principal Component

Analysis with Varimax Rotation

Components

Measures 1 2 3

Word Attack .782

Word Recognition .767 .417

Spelling .756 .513

Reading Fluency -.732

Reading Comprehension .695 .405

Contextual Conventions .638

Contextual Language .534 .477 .385

Chronological Age .8 8 8

Story Schema .842

Story Construction .799

Productivity .612 .457

Note: Values in boldface represent the highest loadings for each factor.

The cognitive and language measures loaded onto factors that made sense conceptually as

the reading and writing factors did. Table 10 shows the factor loadings for the cognitive

and language measures.

64


Table 10 - Factor Loadings of the Cognitive and Language Processing Measures

Principal Component Analysis with Varimax Rotation

Components

Measures 1 2 3

Receptive Vocabulary .856

Expressive Vocabulary .805

Syntax .719

Listening Comprehension .691

Sentence Memory . 6 6 8 .426

Working Memory .676

Sequences .662

TAAS .652

RAN -.610

Short-term Memory .785


The first factor was named the Oral Language Proficiency Factor because it contained all

of the measures pertaining to oral language, with the addition of Sentence Memory. The

second factor was named Cognitive Ability because it contained those lower-level

Cognitive measures that researchers have used to control for cognitive ability apart from

language. The third factor was named the Short-Term Memory Factor because it

contained the one measure that taps short-term auditory memory. This factor was not

used in the multiple sequential regressions because when it was used in an exploratory

manner, it did not account for much of the variance. In addition, it was not reviewed in

the research to be as important a contributor to reading and writing as cognitive ability

and oral language proficiency. The two factors, Cognitive Ability and OLP were then

used to create composites (by first converting the raw scores to z scores, and then by

calculating the overall sum of the individual measures within the factors and averaging

them). The composites rather than the individual measures were entered into the multiple

65


sequential regressions in order to increase power by reducing the total number of

predictors. Further exploration was undertaken within Grade (3, 5/6) to determine if the

cognitive and language factors that were used to create the composites were stable across

development. For more details of these results, see Appendix E.

To investigate the contribution of these composites to reading and writing measures,

sequential multiple regression analyses were conducted. The dependent variables were

the four lower- and higher-level reading and writing measures (Word Attack, Spelling,

Reading Comprehension and Story Construction) . 7 The predictors were the two

composites that were identified in the principal components analysis (Cognitive Ability

and Oral Language Proficiency). Following the recommendations of Cohen and Cohen

(1983) (See also Aiken & West, 1991) the variables used in the interaction term were

centered, that is, each variable was subtracted from the mean score for that variable, prior

to their inclusion in the regression analyses (by transforming them into z scores). This

procedure is used to address the potential problem of multicollinearity that can result

from the inclusion of the independent variables and their interaction terms in the

regression analyses.

The sample was split by Language (ENG/ESL) to investigate similarities and differences

in the contributions of the two composites to the lower- and higher-level reading and

writing measures. The predictors were the composites entered in the following order:

Grade was entered first in order to control for any differences dues to age. The Cognitive

Ability composite was entered before the Oral Language Proficiency composite based on

the assumption that the execution of the lower-level Cognitive processing measures (e.g.,

phonemic awareness, speed of naming, working memory, sequencing) takes place prior

to the higher-level semantic and syntactic Language processing. In other words, higher-

level processes depend on the information supplied by lower-level visual and word

identification processes (Haynes & Carr, 1990). This also allowed for the control of

Cognitive Ability, to examine how much of the variance OLP accounted for above and

7 Story Construction, the standardized measure, was used to represent the higher-level writing task as Story Construction and Story Schema were moderately correlated (r =•7 , p < .01). However, Story Schema also was included in regressions undertaken for the purposes o f further exploration in Appendix I.

66


beyond the Cognitive Ability composite (Geva, 1999). Since one could argue that the

cognitive measures are confounded to some extent with OLP, particularly for the ESL

group, further exploration was undertaken with the OLP composite entered before the

Cognitive Ability composite. For details of these regressions, see Appendix F.

Sequential Regression Split by Language (ENG/ESL)

The Cognitive Ability composite accounted for a comparable percentage of the variance

for both ENG and ESL students. There were, however, differences in the contribution of

the OLP composite, with more of the variance accounted for by OLP in ENG than ESL

students, as seen in Tables 11 and 12. The contribution of the OLP composite was

greater on the higher- level Reading and Writing measures (Reading Comprehension and

Story Construction). With the exception of Spelling, the percentage of total variance

accounted for, was higher for ENG than ESL students.

Table 11.

Summary of the Percentage of Variance Accounted for by the Composites on the Lower- and Higher-Level Reading Skills For ENG and ESL Learners

Word Attack Reading ComprehensionR2 change________________ R2 change______

ENG ESL ENG ESLGrade 4 y ** 6.3 ** 0 .1 0. 18Cognitive Ability 23 7 *** 25.8 *** \ 2 9 *** 12 9 ***OLP 9 3 *** 1.7 39.3 ** 27 7 ***Total Variance 37.7 33.8 52.2 40.7* p < .05, **p < .01, *** p < .001

8 The fact that Grade did not account for any of the variance on the Reading Comprehension task is not surprising given that two different tests were used for grade 3 and 5/6. Thus, the difference between the two cannot be meaningfully compared.

67


Table 12.

Summary of the Percentage of Variance Accounted for by the Composites on the Lower- and Higher-Level Writing Skills For ENG and ESL Learners

Spelling Story ConstructionR2 change________________ R2 change

ENG ESL ENG ESLGrade 22.7 *** 40.2 *** 2.6 * g g ***Cognitive Ability 2 *** g *** g 4 *** 6.7 **OLP g g *** 3.7 17 2 *** 2.6Total Variance 46.7 55.8 26.2 19.1*p< .05, **p< .01, ***p < .001

A combined analysis, presented in Appendix G, confirmed that Language, when entered

as a factor, was not significant, nor were there any significant interactions between

Language and the reading and writing measures. The evident patterns, as seen in Tables

11 and 12, are similar with the combined analysis, as seen in Appendix G. Cognitive

Ability accounted for the most variance in lower-level word attack and spelling for both

ENG and ESL. These results indicated that it was oral language ability that accounted for

the variance in higher-level reading comprehension for both ENG and ESL, and in

higher-level story construction for ENG only. It is important to note that for story

construction, there was still a considerable amount of the variance that was not accounted

for. While OLP accounted for 16.7 percent of the variance in story construction for ENG

students, it did not contribute significantly for ESL students. Only 25 percent of the total

variance was accounted for in ENG story construction, and even less (18 percent) for

ESL story construction. This finding means that some other unmeasured variable was

accounting for that difference.

These results indicate very clearly that Cognitive Ability accounts for more of the

variance in lower-level word attack and spelling. Oral Language Proficiency accounts

for more of the variance, beyond the contribution of Cognitive Ability, in higher-level

reading comprehension, notably for both ENG and ESL students, and in stoiy

construction for ENG students.

68


Further exploration was undertaken to explore how ESL students were able to achieve

comparably with ENG students on higher-level story construction despite significantly

lower OLP. Stepwise regressions were undertaken for both of the higher-level writing

measures (story construction and story schema). Results indicated that while OLP and

Grade accounted for the variance in ENG story construction, a smaller contribution of

OLP and Cognitive Ability accounted for the variance for ESL story construction. For

details of these findings, see Appendix H. Further exploration was initially attempted to

determine the contribution of the various components within the Cognitive Ability and

OLP composites, however, given the number of regressions undertaken, the small sample

size, and the necessary caution required with which to interpret the findings, the analyses

were not included. In addition, further exploration was undertaken to look at the

regressions by Grade in order to determine if the structure of the predictors is different in

Grade 3 than it is in Grade 5/6. Given the small sample size when presented by Grade,

the results of these regressions also should be interpreted with caution. No firm

conclusions can be drawn. For details of these results, see Appendix I.

69


CHAPTER FIVE: GENERAL DISCUSSION

The results will be presented as a general discussion of the contributions, limitations, and

the diagnostic and instructional implications that will outline the practical relevance of

the findings of the current study.

Contributions

An important contribution of the current study is the evaluation of many reading and

writing skills, and cognitive and language abilities for both ENG and ESL students at

more advanced stages of development than have previously been explored. Significant

Grade effects on almost all of the measures, indicated that, in fact, as expected, grade 5/6

students’ reading, writing, cognitive and language abilities were more developed than

were the grade 3’s (Chall, 1990; Fitzgerald & Shanahan, 2000; Levine, 1998).

The contribution the current study makes in terms of comparing ENG and ESL

development is that while surprisingly few significant differences were found on reading,

writing, and cognitive measures even at the higher-levels of reading and writing

development, significant differences remained on the majority of oral language measures,

favoring ENG. In particular, on receptive vocabulary, a moderate effect size indicates

that in fact, a highly significant difference (p < .001) existed between ENG and ESL

students on that particular language measure. While these results confirmed the

expectation that ENG-speaking students’ receptive and expressive language skills,

judgment of syntax, and listening comprehension were more developed than those of

ESL students, the findings seemed inconsistent with the fact that no significant

differences were found on the higher-level reading and writing measures (reading

comprehension and story construction). The fact that widely used standardized measures

were utilized, indicates that this may also be how ESL students appear to be functioning

in the classroom. It would be overstating the findings to say that ENG and ESL students

do not differ in reading comprehension and written language composition. One can only

say that they do not differ on the measures used in this study. These measures may be

70


lacking the level of complexity needed to tap into the linguistic differences that exist

between the groups. It also may be that more linguistically complex scoring methods are

needed within these measures, such as sentence complexity (Mean Length of Utterance),

or breadth of vocabulary. The question that emerges is how the ESL students are able to

achieve comparable higher-level writing skills despite significantly lower OLP. The

results of exploratory stepwise regressions found that for ENG students, while the OLP

composite and Grade accounted for the variance in Story Construction, for ESL students,

it was the OLP and the Cognitive Ability composites that accounted for that variance. It

may be the use of these cognitive ability processes that distinguishes ESL from ENG

students, thus allowing for comparable results even though significant differences in OLP

favored ENG students. It is also possible that nonverbal intelligence plays some role in

accounting for the organization of ideas in the ideation required for higher-level written

composition.

The fact that a significant difference remained in oral language skills at the more

advanced stage, in grade 5/6 ESL students, despite their having received their education

in English, is a significant contribution to the ESL literature, and corroborates

DaFontoura and Siegel (1995). The ESL students likely had less exposure to the more

complex English vocabulary and syntax because English was not the language spoken at

home. Although they received exposure at school to English-speaking teachers, it

appears that the exposure was not sufficient to bring them to the level of their English-

speaking peers. These results indicate that ESL children who come from homes in which

indirect or basic interpersonal communication is almost exclusively in their first

language, may have difficulty with the de-contextualized nature of communication and

language demands at the more advanced stages of literacy development (Juel, 1988).

The strong and consistent relation between lower-level decoding and spelling for both

ENG and ESL students in grade 3 and 5/6 corroborated Shanahan (1984). The ability to

spell accurately and apply basic phonics rules in decoding was closely related, with the

correlation between the two accounting for approximately 49 percent of the variance. In

fact, regression analyses confirmed that these lower-level skills shared similar underlying

71


cognitive processes, as the Cognitive Ability composite accounted for most of the

variance in the lower-level reading and spelling skills for both ENG and ESL students. In

contrast, the correlation between higher-level reading and writing (reading

comprehension and story construction) varied. This relation between higher-level

reading comprehension and story construction was significant in grade 5/6 for ENG

students, as would be expected according to stage theories of literacy development.

These higher-level reading and writing skills depend on the automatization of lower-level

reading and spelling skills. However, for ESL students, the relation between higher-level

reading comprehension and story construction was significant (p< .001) at grade 3 and

not at grade 5/6. This finding is likely, to a certain extent, also to be a result of the

measures used. The reading comprehension measure taps more ordinary words at the

grade 3 level than at the grade 5/6. Thus, the ESL students are better able to consolidate

the relation between reading comprehension and written composition when the language

demands of the test are not as complex. The written language tests did not have the same

level of language demands that the oral language tests did.

The current study confirms that lower-level cognitive processes accounted for more of

the variance in lower-level word attack and spelling skills, while higher-level oral

language processes accounted for more of the variance in the higher-level reading

comprehension and story construction skills. These results would be expected according

to stage theories of reading and writing development (Chall, 1990; and Levine, 1998).

This finding also corroborated the ESL literature that has identified the cognitive

predictors of word identification and spelling, and the language predictors of reading

comprehension (Geva, 1998; Geva & Petrulis-Wright, 1998; Panto, 1999; DaFontoura &

Siegel, 1995). However, the current study did so at a more advanced stage of

development (Stage 3), assessed the contribution of many cognitive and language

predictors that previously have been studied individually, and extended the ESL literature

by documenting that relation with higher-level writing.

72


Limitations

One limitation of the present study is the constraint placed by collecting data as part of a

longitudinal study. Ideally, children in a lower grade would also have been studied in

order to assess beginning literacy development (stage 1) as well as confirmation and

fluency (stage 2) and reading and writing for learning the new (stage 3). This would

allow for a wider range of development for comparison. The use of measures was also a

constraint of the longitudinal study. Ideally, reading comprehension and written

composition measures would be chosen that would tap the more complex linguistic

demands that differentiate ENG from ESL students. Furthermore, the collection of a

measure of nonverbal intelligence would have allowed for a comparison that might have

accounted for the lack of significant differences between ENG and ESL on these higher-

level reading and writing measures. In addition, comparisons across grade have an

unavoidable confound such that some of the schools have grades 5 and 6 in them, and do

not have grade 3. It is therefore not possible to ensure that the catchment areas have

identical demographics. A caution must therefore be taken in interpreting grade and/or

possible social class differences. In terms of the generalizations that can be drawn from

the current study, the fact that the ESL children were tested as part of a heterogeneous

population that spoke many different languages, rather than one (or two) specific LI

backgrounds, is a limitation. There was limited information available regarding the

amount of English the ESL students speak at home, the length of time they had been in

Canada, or other English-speaking country, and whether they were also literate in another

language or orthography. The current study compared similarities and differences in

reading, writing, cognitive and language measures between English-speaking students,

and those for whom English was their second language, defined in the broadest sense.

This allowed for an understanding of the pace of literacy and language acquisition for

ESL students within the context of an English education system; however, general L2

conclusions cannot be drawn for the reasons previously discussed. Further studies are

needed to replicate these results before strong conclusions can be drawn.

73


Diagnostic and Instructional Implications

A significant conclusion from this study is that the development of various aspects of oral

and written language processing skills is not synchronous. These findings have important

implications for assessment and instruction in that assessment can proceed separately for

oral and written language, and that the assessment of potential at-risk status may take

place even when ESL oral skills are not comparable with ENG. When assessing lower-

level reading and writing skills in the earlier grades, it could in fact, be detrimental to

refrain from assessment until the child has developed adequate oral language skills as it is

clear that cognitive processes explain the most variance in lower-level reading and

spelling, while oral language accounts for more of the variance in higher-level reading

comprehension and story construction. The child at risk may require early identification

and intensive instruction focusing on the development of word identification and spelling

skills in addition to the development of oral language skills. The results of the current

study also indicate that continued oral language programming for ESL students is

important even at more advanced stages of literacy development.

Cummins (1976,1980,1984) indicated that most ESL students are able to communicate

in social situations and day-to-day classroom interactions within one or two years;

however, they may require from five to seven years to develop the ability to understand

the academic language used in textbooks and to use English to express the increasingly

complex and abstract concepts encountered in the higher grades. The results of the

current study have shown that although there are few significant academic differences

between ENG and ESL students in grades 3,5/6, there were significant differences in oral

language proficiency favoring ENG students in grades 5/6. ESL students have more

difficulty with complex sentences and vocabulary even at grade 5/6. Efforts must be

made to understand this gap in oral language proficiency and program for it. Closing the

gap in language proficiency is a formidable challenge for ESL students because they

must catch up with a moving target as ENG students continue to grow every year in

language abilities (range of vocabulary, command of complex syntax, etc.). Particularly

at the higher stages of literacy development, academic success depends on students’

74


gaining access to and comprehending the language of text that is found only in books.

Academic language is much more complex than conversational language in terms of

vocabulary and syntax. Thus, students’ knowledge of academic language and their

ability to use academic language coherently in their own writing is crucially dependent

on the amount and variety of what they read, in addition to the language models they

hear. The fact that ESL students will be catching up in academic language over a number

of years means that if they are to acquire the necessary academic English skills, their

language development must be promoted across the curriculum, and taught concurrently

with reading and writing.

75


REFERENCES

Aiken, L.S., & West, S.G. (1991). Multiple regression: Testing and interpreting interaction. Newbury Park, CA: Sage.

Baumann & Kameenui (1991). Research on vocabulary instruction: An ode to Voltaire. In Flood, James, Jensen, Julie, Lapp, Diane, Squire, James, (Eds.), Handbook o f research on teaching the English language arts, (pp. 604-632). Toronto, Ontario: Macmillan Publishing Company.

Becker (1977). Teaching reading and language to the disadvantaged: what we have learned from field research. Harvard Educational Review, 47, 518-543.

Bernhardt, E.B. (1990). A model of L2 text reconstruction: The recall of literary text by learners of Jerman. In A. Labarca & L.M. Bailey (Eds.), Issues in L2: Theory as practice/practice as theory (pp. 21-43). Norwood, NJ: Ablex.

Ben-Dror, I., Bentin, S., & Frost, R. (1995). Semantic, phonological, and morphologic skills in reading disabled and normal children: Evidence from perception and production of spoken Hebrew. Reading Research Quarterly, 30, 876-893.

Bereiter, C. (1980). Development in writing. In L.W. Gregg & E.T. Steinberg (Eds.), Cognitive processes in writing (pp.763-96). Hillsdale, NJ: Erlbaum.

Biemiller, A.J. (1981). Biemiller Test o f Reading Processes. Toronto,Canada: Guidance Centre, Faculty of Education, University of Toronto Press.

Biemiller, A.J. (2001). Teaching vocabulary. American Educator, 25,24-28.

Bowers, P. & Swanson, L. (1991). Naming speed deficits in reading disability: Multiple measures of a singular process. Journal o f Experimental Child Psychology, 51,195- 219.

Bowers, P. & Wolf (1993). Theoretical links among naming speed, precise timing mechanisms and orthographic skill in dyslexia. Reading and Writing: An Interdisciplinary Journal, 5, 69-85.

Bradley & Bryant (1983). Categorizing sounds and learning to read - a causal connection. Nature, 301,419-421.

Brady, S., & Shankweiler, D. (1991). Phonological processes in literacy. Hillsdale, NJ: Erlbaum.

Brack, M. & Genesee, F. (1995). Phonological awareness in young second language learners. Journal o f Children’s Language, 22, 307-324.

Ceci, S. (1984). Developmental study of learning disabilities and memory. Journal o fExperimental Child Psychology, 38, 352-366.

76


Chall, J. (1967). Learning to read: The great debate. New York, NY: McGraw-Hill, Inc.

Chall, J. (1983). Learning to read: The great debate. Updated edition. New York, NY:McGraw-Hill, Inc.

Chall, J. (1996). Learning to read: The great debate, (revised with a new forward) New York, NY: McGraw-Hill, Inc.

Chall, J., Jacobs, & Baldwin (1990). The reading crisis: Why poor children fall behind. Cambridge, MA: Harvard University Press.

Chiappe, P., & Siegel, L.S. (1999). Phonological awareness and reading acquisition in English- and Punjabi-speaking Canadian children. Journal o f Educational Psychology, 91,20-28.

Cisero, C. A., & Royer, J.M. (1995). The development and cross-language transfer of Phonological awareness. Contemporary Educational Psychology, 20, 275 -303.

Clarke, L. (1988). Invented versus traditional spelling in first graders’ writings: Effectson learning to spell and read. Research in the Teaching o f English, 22, 281-309.

Cohen, J. & Cohen, P. (1983). Applied Multiple Regression/Correlation Analysis for the Behavioral Sciences. Hillsdale, NJ :Lawrence Erlbaum Associates.

Coyne, M.D., Kame-enui, E.J., & Simmons, D. (2001). Prevention and intervention in beginning reading: Two complex systems. Learning Disabilities Research and Practice, 16(2), 62-73.

Cummins, J. (1991). Conversational and academic language proficiency in bilingual contexts. In j. H. Hulstijn and J.F. Matter (eds) Reading in Two Languages (pp. 75- 89). Amsterdam: AILA.

Cummins, J. (1984). Bilingualism and Special Education: Issues in Assessment and Pedagogy. Clevendon: Multilingual Matters.

Cummins, J. (1980). Psychological assessnebt of immigrant children: Logic or intuition? Journal o f Multilingual and Multicultural Development 1, 97-111.

Cummins, J, (1979). Linguistic interdependence and the educational development of bilingual children. Review o f Educational Research, 49, 222-251.

DaFontoura, H.A., & Siegel, L.S. (1995). Reading, syntactic, and working memory skills of bilingual Portugese-English Canadian children. Reading and Writing: An Interdisciplinary Journal, 7, 139-153.

77


Daneman, M. (1996). Individual differences in reading skills. In R. Barr, M. L. Kamil, P. Mosenthal, & P. D. Pearson (Eds.), Handbook o f reading research (Vol. 2, pp. 512- 538). Mahwah, NJ: Erlbaum.

Denckla, & Rudel. (1974). Rapid “automatized” naming of pictured objects, colors, letters, and numbers by normal children. Cortex, 19, 186-202.

Denckla & Rudel. (1976). Rapid “automatized” naming (R.A.N.): Dyslexia differentiated from other learning disabilities. Neuropsychologia, 14,471-479.

Dunn, L.M., & Dunn, L.M. (1981). Peabody Picture Vocabulary Test- Revised. Circles Pines, Minnesota: American Guidance Service.

Dorgunoglu, A.Y., Nagy, W.E., & Hancin-Bhatt, B.J. (1993). Cross-language transfer of phonological awareness. Journal o f Educational Psychology, 85,453-465.

Durkin, D. (1993). Teaching them to read. (6 th Edition). Boston, MA: Allyn and Bacon.

Durrell, D.D. (1970). Durrell Analysis o f Reading Difficulty. New York: Harcourt, Brace & Janovich.

Foorman, B.R. (1995). Research on the “great debate”: Code-oriented versus whole language approaches to reading instruction. School Psychology Review, 24, 376-392.

Ehri, L. C. (1986). Sources of difficulty in learning to spell and read. In M. Wolraich &D. Routh (Eds.), Advances in developmental and behavioural pediatrics (pp. 121-195). Greenwich, CT: JAI.

Fitzgerald, J. & Shanahan, T. (2000). Reading and writing relations and their development. Educational Psychologist, 35(1), 39-50.

Frith (1985). Unexpected spelling problems. In U. Frith (Ed.), Cognitive processes in spelling (pp.495-516). London: Academic.

Geva, E. (1998). Issues in the development of second language reading: Implications for instruction and assessment. In T. Nunes & P. Bryant (Eds.) Investigating Literacy Research and Practice.

Geva, E., & Clifton, S. (1993). The development of first and second language reading skills in early French Immersion. Canadian Modern Language Review, 50, 646-667.

Geva, E., & Petrulis-Wright, J. (1998). The role of English language proficiency in the reading development of LI and L2 primary level children.

Geva, E. & Ryan, E.B. (1993). Linguistic and cognitive correlates of academic skills in first and second languages. Language Learning, 43, 5-42.

78


Geva, E., & Siegel, L. (2000). Orthographic and cognitive factors in the concurrent development of basic reading skills in two languages. Reading and Writing: An Interdisciplinary Journal. 12:1-30.

Geva, E. & Verhoeven, L. (2000). Introduction: The development of second language reading in primary children - research issues and trends. Scientific Studies o f Reading, 4(4), 261-266.

Geva, E. & Wade-Woolley, L. (1998). Component processes in becoming English- Hebrew biliterate. In A.Y. Durgonoglu & L. Verhoeven (Eds),, Acquisition o f literacy in a multilingual context: A cross-cultural perspective (pp.85-110). Mahwah, NJ: Lawrence Erlbaum Associates, Inc.

Geva, E., Wade-Woolley, L., & Shany, M. (1997). The concurrent development of spelling and decoding in two different orthographies. Journal o f Reading Behaviour, 25, 383-406.

Geva, E., Yaghoub-Zadeh, & Schuster, B. (2000). Understanding individual differences in word recognition skills of ESL children. Annals o f Dyslexia, 50,123-154.

Gholamain, M., & Geva, E. (1998). The concurrent development of word recognition skills in English and Farsi. Language Learning.

Gottardo, A., Stanovich, K.E., & Siegel, L.S. (1996). The relationships between phonological sensitivity, syntactic processing and verbal working memory in the reading performance of third-grade children. Journal o f Experimental Child Psychology, 63, 563-582.

Gough (1996). The beginning of decoding.

Gough, P. B., & Tunmer, W. E. (1986). Decoding, reading and reading disability. Remedial and Special Education, 1 ,6-10.

Griffith (1991). Phonemic awareness helps first graders invent spelling and third graders remember correct spellings. Journal o f Reading Behavior, 23, 215-232.

Harrington, M., & Sawyer, M. (1992). L2 working memory capacity and L2 reading skills. Studies in Second Language Acquisition, 14, 25-38.

Haynes, M., & Carr, T. (1990). Writing system background and second language reading: A component skills analysis of English reading by native speakers of Chinese. In T. Carr & B. Levy (Eds.), Reading and its development: Component skills approaches (pp. 375-421). San Diego: Academic.

79


Hoien, T., Lundberg, I., Stanovich, K.E., & Bjaalid, I.K. (1995). Components of phonological awareness. Reading and Writing: An Interdisciplinary Journal, 7,171- 188.

Hoover, W.A., & Gough, P.B. (1990). The simple view of reading. Reading and Writing: An Interdisciplinary Journal, 2, 127-160.

Jastak, S. & Wilkinson, G.S. (1984). Wide Range Achievement Test (WRAT-R). Wilmington, Delaware: Jastak Associates Inc.

Juel, C., Griffith, L., & Gough, P. (1986). Acquisition of literacy: A longitudinal study of children in first and second grade. Journal o f Educational Psychology, 78, 243-255.

Juel, C. (1988). Learning to read and write: A longitudinal study of 54 children from first through fourth grades. Journal o f Educational Psychology, 80,437-447.

Koda, K. (1994). Second language reading research: Problems and possibilities. Applied Psycholinguistics, 15, 1-28.

Levine, M. D. (1998). Developmental variation and learning disorders. Cambridge,MA: Educators Publishing Service, Inc.

Lloyd, S. (1983). The Phonics Handbook. London: Jolly Learning Publishers.

McLaughlin, B. (1990). Theories of second language learning. Applied Linguistics, 2, 113-158.

Mann, V. A., & Brady, S. (1988). Reading disability: The role of language deficiencies. Journal o f Consulting and Clinical Psychology, 56, 811-816.

Mann, V. A., Cowin, E., & Shoenheimer, J. (1989). Phonological processing, language comprehension and reading ability. Journal o f Learning Disabilities, 22, 76-89.

Mann, V.A., Shankweiler, D., & Smith, S.T. (1985). The association betweencomprehension of spoken sentences and early reading ability: The role of phonetic representation. Journal o f Child Language, 11, 627-643.

Meyer, M., Wood, F., Hart, L., & Felton, R. (1998). Selective predictive value of rapid automatized naming in poor readers. Journal o f Learning Disabilities, 31 (2), 106-117.

Moats, L.C. (1999). Teaching reading is rocket science: What expert teachers o f reading should know and be able to do. Item No. 372, 6/99. American Federation of Teachers: Washington, DC.

Morris & Pemey (1984). Developmental spelling as a predictor of first-grade reading achievement. The Elementary School Journal, Vol. 84, No. 4,441-457.

80


Nassaji & Geva, E. (1999). The contribution of phonological and orthographic processing skills to adult ESL reading: Evidence from native speakers of Farsi. Applied Psycholinguistics 20, 241-267.

National Reading Panel (2000). Teaching children to read: An evidence-based assessment of the scientific research literature on reading and its implications for reading instruction. Available athttp://www.nichd.nih.gov/publications/nrp/smallbook.htm.

National Research Council (1998). Preventing reading difficulties in young children. Washington, DC: National Academy Press.

Nodine, B., Barenbaum, E., & Newcomer, P. (1985). Story composition by learning disabled, reading disabled, and normal children. Learning Disability Quarterly, 8 , 167- 179.

Panto, J. (1999). A Five-year longitudinal study of spelling and writing development: Differences and similarities in LI and L2 elementary students. Unpublished masters thesis, OISE, University of Toronto.

Perfetti, C. A. (1984). Reading acquisition and beyond: Decoding includes cognition. America Journal o f Education, 92,40-60.

Perfetti, C. A. (1985). Reading ability. New York: Oxford University Press.

Petrulis-Wright, J. (1998). The role of oral language proficiency in the development of LI and L2 basic reading skills in young children. Unpublished masters thesis, OISE, University of Toronto.

Ricard, R.J., & Snow, C.E. (1990). Language use in and out of context. Journal o f Applied Developmental Psychology, 11, 251-266.

Rice, M.L. (1989). Children’s language acquisition. American Psychologist, 44, 149- 156.

Robbins, C., & Ehri, L.C. (1994). Reading storybooks to kindergartners helps them learn new vocabulary words. Journal o f Educational Psychology, 8 6 , 54-64.

Scarborough, H. (1989). Prediction of reading disability from familial and individual differences. Journal o f Educational Psychology, 81, 101-108.

Scardamalia, M. (1981). How children cope with the cognitive demands of writing. In C.H. Fredericksen & J.F. Dominic (Eds.), Writing: The nature, development and teaching o f written communication. Vol. 2: Writing: Process, development, and communication (pp.81-104). Hillsdale, NJ: Erlbaum.

81


http://www.nichd.nih.gov/publications/nrp/smallbook.htm

Service, E. (1992). Phonology, working memory and foreign-language learning. Quarterly Journal o f Experimental Psychology, 45 A, 21-50.

Service, E., & Kohonen, V. (1995). Is the relation between phonological memory and foreign language learning accounted for by vocabulary acquisition? Applied Psycholinguistics, 16, 155-172.

Shanahan & Lomax (1998). A developmental comparison of three theoretical models of the reading-writing relationship. Research in the Teaching o f English, Vol. 22, No.2, 196-212.

Shanahan & Lomax, (1986). An analysis and comparison of theoretical models of the reading-writing relationship. Journal o f Educational Psychology. Vol. 78, No.2, 116- 123.

Shanahan, T. (1984). Nature of the reading-writing relation: An exploratory multivariate analysis. Journal o f Educational Psychology. Vol. 76, No. 3,466-477.

Siegel, L.S. (1993). The development of reading. In H. W. Reese (Eds.), Advances in child development and behavior (Vol. 24, pp. 63-97). San Diego, CA: Academic Press.

Siegel, L.S., & Ryan, E.B. (1988). Development of grammatical sensitivity, phonological and short-term memory skills in normally achieving and learning disabled children. Developmental Psychology, 24, 28-37.

Siegel, L.S., & Ryan, E.B. (1989). The development of working memory in normally achieving and subtypes of learning disabled children. Child Development, 60,973- 980.

Snow, C.E. (1991). The theoretical basis for relationships between language and literacy in development. Journal o f Research in Childhood Education, 6 , 5-10.

Stanovich, K.E. (1992). Speculations on the causes and consequences of individual differences in early reading acquisition. In P. B. Gough, L.C. Ehri, & R. Treiman (Eds.), Reading Acquisition, Hillsdale, New Jersey: Erlbaum.

Stanovich, K.E. (1991). Changing models in reading and reading acquisition. In L.Rieben & C. Perfetti (Eds.), Learning to read: Basic research and its implications (pp. 19-31). Hillsdale, NJ: Erlbaum.

Stanovich, K.E. (1986). Matthew effects in reading: Some consequences of individual differences in the acquisition of literacy. Reading Research Quarterly, 21, 360-407.

Stein, N. & Glenn, C. (1982). Children’s concept of time: The development of a story schema. In W. Freidman (Ed.), The developmental psychology o f time. New York: Academic Press.

82


Stein, N. & Policastro, M. (1984). The concept of a story: A comparison between children’s and teacher’s viewpoints. In H. Mandl, N. Stein & T. Trabasso (Eds.). Learning and comprehension o f text (pp. 113-155), Hillsdale, NJ: Ablex.

So, D., & Siegel, L.S. (1997). Learning to read Chinese: Semantic, syntactic,phonological and working memory skills in normally achieving and poor Chinese readers: Reading and Writing: An Interdisciplinary Journal, 9, 1-21.

Swanson, H.L., Ashbaker, M., & Lee, C. (1996). The effects of processing demands on the working memory of learning disabled readers. Journal o f Experimental Child Psychology.

Swanson, H.L. (1992). Generality and modifiability of working memory among skilled and less skilled readers. Journal o f Educational Psychology, 64,473-488).

Torgesen, J. K. (2000). Individual differences in response to early interventions in reading: The lingering problem of treatment resisters. Learning Disabilities Research and Practice, 15,55-64.

Torgesen, J.K., Rashotte, C.A., and Alexander, A.W. (2001). Principles of fluencyinstruction in reading: Relationships with established empirical outcomes. In M. Wolf (Ed.), Dyslexia, fluency, and the brain. Timonium, MD: York Press.

Tunmer, W.E., & Hoover, W.A. (1992). Cognitive and linguistic factors in learning to read. In P.B. Gough, L.C. Ehri, & R. Treiman, (Eds.), Reading Acquisition (pp. 175- 214). Hillsdale, NJ: Erlbaum.

Verhoeven (1990). Components in early second language reading and spelling, Scientific Studies o f Reading, 4 (4), 313-330.

Wade-Woolley, L., & Siegel, L.S. (1997). The spelling performance of ESL and native speakers of English as a function of reading skill. Reading and Writing: An Interdisciplinary journal, 9, 387-406.

Willows, D.M. (1996). Balanced literacy for the 90’s: An evaluation o f a new approach. Presented as part of a symposium entitled “Phonics within a balanced literacy program, at the annual meeting of the National Reading Conference, Charleston,South Carolina.

Wolf, M., & Bowers, P.G. (1999) The double-deficit hypothesis for the developmental dyslexias. Journal o f Educational Psychology, Vol. 91, No.3, 415-438.

Woodcock, R.W. (1987). Woodcock Reading Mastery Tests-Third Edition. Minnesota: American Guidance Service.

83


Reproduced

with perm

ission of the

copyright ow

ner. Further

reproduction prohibited

without

permission.

Appendix A

Table A-l

Group Comparisons for Grade 5 and 6 on Reading, Writing, Cognitive and Language Processing Measures (based on raw scores)

GRADE5

GRADE6

ENG ESL F P < r ENG ESL F P < 9 2

n = 83 n = 23 n = 40 n = 25Reading MeasuresWord Attack Maximum score = 45

X = 30.06 S.D.= 8.06

X = 31.81 S.D. = 7.55

.846 n/s .008 X = 31.42 S.D. = 6.80

X = 31.52 S.D. = 7.43

.003 n/s .0 0 0

WRAT Word Recognition Maximum score = 57

X = 36.02 S.D. = 4.97

X = 36.59 S.D. = 4.77

.2 2 0 n/s . 0 0 2 X = 37.80 S.D. = 4.85

X = 37.76 S.D. = 3.85

.0 0 1 n/s .0 0 0

Reading Comprehension Maximum score = 48

X = 26.92 S.D. = 9.26

X = 24.26 S.D. - 8.53

1.51 n/s .015 X = 28.57 S.D. = 9.68

X = 26.76 S.D. = 7.58

.611 n/s .0 1 0

Time (seconds) X = 50.42 S.D. = 29.47

X = 48.72 S.D. — 17.15

.055 n/s .0 0 1 X = 44.97 S.D. = 18.50

X = 39.66 S.D. = 8.70

1.83 n/s .029

84

Reproduced

with perm

ission of the

copyright ow

ner. Further


without

permission.

GRADE5

GRADE6

ENG ESL F P< ENG ESL F P< v 2Writing MeasuresWRATSpellingMaximum score = 55

X = 30.90 S.D. = 5.79

X = 31.91 S.D. = 4.16

.603 n/s .03 X = 31.45 S.D. = 5.64

X = 33.60 S.D. = 4.56

4.12 .05 .06

TOWL-III Spelling Errors (#) X = 9.08 S.D. = 7.50

X = 7.50 S.D. = 4.72

.6 6 8 n/s .007 X = 9.20 S.D. = 6.89

X = 6.04 S.D. = 3.75

4.56 .05 .070

TOWL-IIIContextual Conventions Maximum score =18

X = 4.63 S.D. = 3.20

X = 4.04 S.D. = 3.25

.552 n/s .029 X - 5.35 S.D. = 3.18

X = 6.12 S.D. = 2.36

1.04 n/s .017

TOWL-IIIContextual Language Maximum score = 29

X = 12.76 S.D. = 4.15

X = 12.65 S.D. = 3.47

.0 1 0 n/s .0 0 0 X = 13.76 S.D. = 3.80

X = 13.60 S.D. = 3.47

.028 n/s .0 0 0

TOWL-III Story Construction Maximum score = 21

X = 11.43 S.D. = 4.01

X = 13.08 S.D. = 2.82

3.39 p=.06 .035 X = 13.02S.D. = 3.56

X = 11.72 S. D. = 3.40

2.16 n/s .034

Story Schema Maximum score = 4

X = 3.38 S.D.= .93

X = 3.75 S.D. = .55

2.65 n/s .031 X = 3.61 S.D. = .84

X = 3.20 S.D. = .91

3.66 p=.06 .057

Productivity (# of words) X = 127.20 S.D. = 46.18

X = 149.43 S.D. = 44.34

4.14 .05 .042 X = 149.69 S.D. = 42.08

X = 147.68 S.D. = 42.48

.034 n/s .0 0 1

85

Reproduced

with perm

ission of the

copyright ow

ner. Further


without

permission.

GRADE5

GRADE6

ENG ESL F P< V2 ENG ESL F P< V2Cognitive Measures

RosnerMaximum score = 22

X = 15.40 S.D. = 5.58

X = 15.68 S.D. = 5.48

.042 n/s .0 0 0 X = 16.17 S.D. = 5.79

X = 17.24 S.D. = 5.65

.527 n/s .008

RAN(seconds)

X = 31.04 S.D. = 9.80

X = 26.60 S.D. = 4.79

4.53 .05 .042 X = 29.25 S.D. = 8.20

X = 24.80 S.D. = 2.82

6.71 .0 1 .098

CMS -Sequences Maximum score = 84

X = 52.68 S.D. = 11.13

X = 54.39 S.D. = 9.68

.439 n/s .004 X = 55.12 S.D. = 8.12

X = 59.56 S.D. = 7.76

4.65 .05 .070

CMS -Numbers Forward Maximum score =16

X = 8.39 S.D. = 1.73

X = 8.69 S.D. = 1.49

.564 n/s .005 X = 8.60 S.D. = 1.94

X = 9.28 S.D. = 1.59

2 .1 0 n/s .033

CMS - Numbers Backward Maximum score =14

X = 4.48 S.D. = 1.59

X = 4.39 S.D. = 1.40

.061 n/s .0 0 1 X = 4.70 S.D. = 1.57

X = 5.24 S.D. = 1.47

1 .8 6 n/s .029

WRAML - Verbal Learning Gr. 3 Maximum score = 52 Gr. 5/6 Maximum score = 64

X = 31.46 S.D. = 7.67

X = 30.86 S.D. = 8.68

.105 n/s .0 0 1 X = 32.42 S.D. = 7.01

X = 33.52 S.D. = 9.53

.271 n/s .004

WRAML - Sentence Memory Maximum score = 40

X = 16.95 S.D. = 4.18

X = 15.45 S.D. = 4.74

2.08 n/s .0 2 0 X = 18.20 S.D. = 4.18

X = 17.32 S.D. = 3.83

2.29 n/s .014

WRAML - Story MemoryGr. 3 Maximum score = 45 Gr. 5/6 Maximum score =51

X = 25.54 S.D. = 6.99

X = 23.43 S.D. = 7.54

1.58 n/s .015 X = 27.07 S.D. = 9.37

X = 25.68 S.D. = 7.31

.756 n/s .0 1 2

86

Reproduced

with perm

ission of the

copyright ow

ner. Further


without

permission.

GRADE5

GRADE6

ENG ESL F P< r ENG ESL F P< f}2Language MeasuresReceptive Vocabulary Maximum score = 204

X = 140.74 S.D. = 17.56

X = 128.82 S.D. = 19.52

7.83 .0 1 .071 X - 147.45 S.D. = 15.50

X =136.48 S.D. = 17.02

7.08 .0 1 .103

Expressive Vocabulary Maximum score = 190

X = 98.18 S.D. = 15.61

X = 94.69 S.D. = 12.81

.956 n/s .009 X = 100.65 S.D. = 16.02

X = 94.28 S.D. = 8.80

3.40 p=.07 .052

SyntaxMaximum score = 22

X = 15.59 S.D. = 3.84

X = 14.21 S.D. = 3.47

2.38 n/s .023 X = 15.90 S.D. = 2.53

X = 14.56 S.D. = 3.77

2.87 n/s .044

Listening Comp. Maximum score = 6

X = 4.38 S.D. = .97

X = 4.17 S.D. = 1.07

.807 n/s .008 X = 4.42 S.D. = 1.23

X = 3.92 S.D. = .95

2.97 n/s .046

87

Appendix B

Planned comparisons between Grade 3 and 5/6 ENG and ESL on Writing, Cognitive and Language Measures

Writing Measures

Planned comparisons were undertaken to see where the significant differences lay in the

writing and cognitive measures between ENG and ESL students, particularly given the

significant Grade effects. The results will be discussed and presented graphically.

Figure B-l

1. Mean Spelling by Grade and Language (WRAT Spelling Raw Scores)

■ BIG

■ ESL

Students received a point for spelling the target word correctly, thus, a higher score

indicates more words spelled correctly. While this was the only significant difference,

three notable trends were present. The grade 3 students showed a trend on Spelling,

favoring ESL students, F (2, 117) = 3.69, p = .06, with a small effect size of i]2 = .03.

The grade 3 students also showed a trend on Contextual Conventions, F (2,117) = 3.31,

p = .07, with a small effect size of rj2 = .03. This finding is not surprising given that

accuracy of spelling is a main scoring feature of the Contextual Conventions measure.

A third trend was found in grade 5/6 on the TOWL (# of) Spelling Errors measure, again

with the ESL students outperforming the ENG students, F (2,159)= 3.19,/? = .07, with a

small effect size, i]2 = .02 (See Figure IB-2.) There were no significant interactions with

Language on any of the writing measures.

88


Figure B -2. Mean TOWL - III Spelling Errors (# of errors) by Grade and Language

12

10

8

6

4

2

oGrade 3 G rades 5/6

To explore factors that might account for the apparent ESL advantage on spelling

measures, an ANCOVA controlling for chronological age was undertaken on the

measures of Cognitive Ability. As will be discussed in Question 4, a Cognitive-Ability

composite (comprised of the TAAS, RAN, working memory and sequencing ability) was

created. No significant differences were found between ENG and ESL groups on this

Cognitive Ability composite for either the grade 3 group, F (2, 117) = .001, p = .98,

rj2 ~ .00, or for the grade 5/6 group, F (2, 171) — .08,/? = .78, ij2 = .00. However, as the

results in the Cognitive measures section will show, a significant difference and a trend

were found on two of the individual measures that comprise the Cognitive Ability

composite (RAN and Sequencing), with the ESL students outperforming the ENG

students.

Overall, the pattern of results on the writing measures indicates that, with the exception

of more accurate spelling in grade 5/6 for ESL students, and a trend favoring grade 3 ESL

students’ use of the conventions of writing (with a spelling component), there were no

other significant differences found in the remainder of the writing measures. The ESL

students’ use of language in writing (sentence structure), their ability to construct stories,

and their schema for stories, were developed to the same extent as the ENG students.

89


Surprisingly, there were no significant differences on the higher-level writing task, Story

Construction, with its inherently more complex cognitive and language demands.

Cognitive Measures

Figure B -3. Mean RAN time (measured in seconds) by Grade and Language

Grade 3 Grade 5/6

HENG ■ ESL

Performance on RAN is measured by the time in seconds it takes for the student to name

geometric objects as rapidly as they can (5 objects repeated 20 times each in random

sequence). Therefore, the shorter the amount of time, the better the performance. On this

graph, one can clearly see that the grade 5/6 ESL students took less time to name the

objects in question than the ENG students. This finding may, in fact, help to explain the

spelling differences favoring the ESL children. There is research to support a positive

relation between spelling and rapid automatic naming (Denckla & Rudel, 1974,1976 and

Bowers, 1995). RAN measures how automatically children recode visual information

into a phonologically represented name code (Denckla & Rudel, 1974,1976), and such

tasks have been shown to contribute specific variance to orthographic processing

(Bowers, 1995).

Within these planned comparisons between Grades (3, 5/6), a trend was evident in Grade

5/6 favoring ESL students on the Sequencing measure, F (2, 170) = 3.36,p = .07, with a

small effect size, rj2 ~ .02, as can be seen in Figure 1C -2. This task measured the

90


students’ ability to recall rote information, such as the alphabet, the days of the week and

the months of the year forward and backward in sequences as quickly as possible.

Figure B-4 Mean Sequences by Grade and Language (CMS Raw Scores)

60

50

40

30

20

10

0Grade 3 Grade 5/6

There is also research to support the positive relation between spelling and sequencing

ability. As students spell words, they must segment each word into its constituent speech

sounds and represent the sounds with the appropriate symbols by holding the sounds in

active working memory and sequentially representing each sound in the word. Thus,

sequencing ability and memory are important in the segmentation and synthesis that is

required for spelling (Lieberman, Shankweiler, Liberman, Fowler & Fischer, 1977).

While it is impossible to know for sure why the obtained differences existed in this

sample, it is likely that the ESL students’ faster and more automatic recoding of visual

information and their better-developed sequencing ability positively impacted their

spelling.

Language Measures

The results for all four language measures, receptive vocabulary, expressive vocabulary,

syntax and listening comprehension, are summarized in Figures 2d 1-4. On receptive

vocabulary, planned comparisons between Grade (3, 5/6) revealed a trend in grade 3,

favoring ENG students, F (2, 108) = 3.03,p = .08, with a small effect size, tj2 = .03. A

significant difference was found in grade 5/6, favoring ENG students, on the receptive

91


vocabulary measure, F (2,170) = 14.45,/? < .001, with a moderate effect size, i]2 = .08.

There was no significant interaction with Language.

Figure B-5. Mean Receptive Vocabulary by Grade and Language (PPVT Raw Scores)

Grade 3 Grade 5/6

Figure B-5 clearly shows the superiority of the ENG students in their store of receptive

vocabulary, with both a trend at grade 3, and a significant difference at grade 5/6. This

Language effect is not surprising. The language demands at the beginning of grade 3 are

not as complex. In grade 5 and 6 , the language demands are more complex because they

refer now to abstract de-contextualized vocabulary that is removed from everyday

experience. Thus, it is not surprising that ESL students would have more difficulty with

these more challenging vocabulary items in grades 5 and 6 than ENG students.

An analysis of the mean number of items achieved on the receptive vocabulary task in

grade 3 indicates that ENG and ESL students were able to handle comparably

challenging items such as “ co-operating, microscope, and archery”. However, in grade

5/6, ESL students have greater difficulty with items that the ENG students are able to

answer, such as “pedestrian, and constrained”. Thus, when the material is more difficult

and a larger vocabulary and more developed syntax are required, ESL students have more

difficulty on items that they would have little exposure to beyond what they would

experience in written text. This same pattern is also evident on the syntax and the

expressive vocabulary tasks. A significant difference was found, favoring ENG students,

in their judgment of syntax, F(2, 170) = 5.1 \ , p < .05, with a small effect size, ij2 = .03.

92


Figure B - 6. Mean Syntax by Grade and Language (CELF-III - Formulated Sentencesraw scores)

Grade 3 Grade 5/6

This graph shows the superiority for grade 5/6 ENG students over ESL students in their

understanding and judgment of syntax. This finding is consistent with Da Fontoura and

Siegel (1995), who found that the syntactic skills of bilingual children bom in Canada

still lagged behind the skills of monolingual children in grades 4, 5 and 6 .

93


Figure B - 7. Mean Expressive Vocabulary by Grade and Language (EVT raw scores)

Grade 3

■ BIG

■ ESL

No significant difference was found on expressive vocabulary in grade 3, however, a

significant difference was found, favoring ENG students, in grade 5/6 as can be seen in

this graph, F (2, 170) = 4.00,/? < .05, with a small effect size, q2 = .02.

The same pattern was found on listening comprehension. No significant differences were

found in grade 3; however, the difference approached significance in grade 5/6, favoring

ENG students. This trend can be seen in Figure B - 8

Figure B - 8 . Mean Listening Comprehension by Grade and Language (Durrell raw

scores).

■ ENG■ ESL

94


The results for judgment of syntax, expressive vocabulary and listening comprehension

are consistent with what was found on the receptive vocabulary measure and can be

explained according to the stages of literacy development in the same manner. The

vocabulary and syntax demands are not as complex in early grade 3 as they are in grades

5 and 6 . According to the stages of literacy development, comprehension becomes more

decisive in grades 5 and 6 because of increasing linguistic sophistication that surpasses

the content of everyday speech. Listening comprehension integrates complex cognitive

and language skills, and depends on knowledge of vocabulary, concepts, text structures,

syntax and pragmatics.

With the exception of the trend found in receptive vocabulary, no other significant

differences were found on the language measures for this grade 3 sample. This lack of

significant differences between the grade 3 ENG and ESL groups was very surprising,

given that the children had a language other than English as their home language and that

many of the children would have spoken little English prior to school entry (into half-day

senior kindergarten). Many of these children, tested in the beginning months of grade 3,

would have had fewer than 3 years of regular English language use in school. Based on

Cummins’s (1984) 5-to-7-years estimate for achieving cognitive academic language

proficiency, the grade 3 ESL children would have been expected to differ substantially

from their ENG counterparts on measures of oral language processes. But they did not.

To ensure that the grade 3 ESL sample was truly “second language”, the children were

traced back in the longitudinal data set to senior kindergarten (1996/1997) when language

measures were obtained for them. Analyses of their language skills at that time showed

that the ENG children were significantly better than the ESL children on both vocabulary

and syntactic measures (p < .001). Thus, these children were truly “second language”

learners. Details of group comparisons are provided in Table B-9 and B-10.

95


Table B-9.

Group Comparisons on Language and Literacy Measures (based on raw scores) for the Grade 3 sample (1999/2000) tracked back to Senior Kindergarten (1996/1997)

S.K.

ENG ESL F P<

n= 114 n = 94

Language Measures

Receptive Vocabulary X = 64.50 S.D. = 14.53

X = 47.96 S.D. = 14.89

64.45 .0 0 1

Linguistic Concepts (CELF) X = 17.38 S.D. = 2.68

X = 15.07 S.D. = 4.15

22.96 .0 0 1

Sentence repetition X = 15.84 S.D. = 2.55

X = 13.37 S.D. = 3.45

34.32 .0 0 1

Wepman Auditory Discrimination (# correct)

X = 16.54 S.D. = 2.42

X = 16.16 S.D. = 3.49

.843 n/s

Rosner X = 2.29 S.D. = 2.40

X = 2.34 S.D. = 2.49

.018 n/s

96


Table B-10.

Group Comparisons on Language and Literacy Measures (based on raw scores) for the 16 Grade 3 students (1999/2000) tracked back to Senior Kindergarten (1996/1997)

S.K.

ENG ESL F P<

n = 7 n = 9

Language Measures

Receptive Vocabulary X = 69.29 S.D. = 15.51

X = 51.67 S.D. = 17.26

4.47 .05

Linguistic Concepts (CELF) X = 18.14 S.D. = 1.07

X = 15.44 S.D. = 4.07

2 .8 8 n/s

Sentence repetition X = 16.86 S.D. = 1.35

X = 12.89 S.D. = 3.82

6.79 .05

Wepman Auditory Discrimination (# correct)

X = 17.00 S.D. = 2.08

X = 17.33 S.D. = 2.18

.096 n/s

Rosner X = 7.16 S.D. = 3.97

X = 5.71S.D. = 3.03

.558 n/s

Given the significant differences found in these oral language proficiency measures, to

the advantage of ENG students, it is important to highlight the lack of differences that

was found on the academic measures, at both the lower- and higher-levels of reading and

writing development. The second research question will address the relation between

lower- and higher-level reading and writing by Grade (3, 5/6) and Language (ENG/ESL).

97


Appendix C

The results of these correlational analyses will be discussed in the following order. The

first table (Tables C-l) compares the partial correlations (controlling for chronological

age) between reading and writing measures for grade 3 ENG and ESL students. The

following table (Table C-2) compares grade 5/6 ENG and ESL students on those same

measures. Using the Bonferroni approach to control for Type 1 error across the many

correlations, ap- value of less than .005 (.05/10) was required for significance. In the

table the correlations that reached this level of significance are bolded.

98


Reproduced

with perm

ission of the

copyright ow

ner. Further


without

permission.

Appendix CTable C-l. Partial Correlations (controlling for chronological age) of Reading Measures with Writing Measures for Grade 3 ENG and ESL Students

Grade 3

Writing Measures

Spelling ContextualConventions

Productivity ContextualLanguage

StoryConstruction

Story Schema

Reading Measures

Word Attack ENG .74 .30 .25 .34 .38 .09

ESL .77 .45 .15 .53 .36 .20

WRAT Reading ENG .71 .40 .18 .52 .50 .20

ESL .74 .46 .17 .53 .33 .15

Fluency (Reading Time) ENG -.51 -.29 -.19 -.37 -.25 -.09

ESL -.48 -.32 -.20 -.46 -.21 -.01

Reading Comprehension ENG .59 .30 .18 .38 .30 .06

ESL .68 .51 .28 .62 .53 .20

Note: For ENG students df= 52, r = .30,/? < .05; r = .34, p < .01; r = .50,/? < .001; For ESL students df= 55, r - .28,/? < .05; r = .33,/? < .01; r - .45,/? < .001; Based on Bonferroni adjustment, p < .001 in bold.

98

The Relation between Reading and Writing for Grade 3 ENG and ESL Students

The results of the correlational analysis presented in Table C-l show that 18 of the partial

correlations (controlling for chronological age) between reading and writing measures in

grade 3 were statistically significant, with a larger number of significant correlations for

ESL (12) than ENG (6 ) students. The strongest correlations for both ENG and ESL grade

3 students were between spelling and all four reading measures, ranging from r = .48 to

•77,/? < .001. The correlations between these reading measures and spelling were

comparable for both ENG and ESL.

For ENG, the only significant correlations beyond reading and spelling were found

between WRAT Reading and Contextual Language (r = .52, p < .001), and WRAT

Reading and Story Construction (r = .50,/? < .001). In contrast, ESL students had several

significant correlations between different reading and writing measures, such as Word

Attack and Contextual Language (r - .53,p < .001), and Fluency (Reading Time) and

Contextual Language (r = -.46, p < .001). In addition, for ESL higher-level Reading

Comprehension was significantly correlated with all three writing measures: lower-level

Contextual Conventions (r = .51 ,P < .001); Contextual Language (r = .62, p < .001); and

higher-level Story Construction (r = .53, p < .001). For ENG students, Reading

Comprehension was, as previously mentioned, correlated with Spelling (r= .59, p<

.001).

99


Reproduced

with perm

ission of the

copyright ow

ner. Further


without

permission.

Table C-2.

Partial Correlations (controlling for chronological age) of Reading Measures with Writing Measures for Grades 5/6 ENG and ESL Students

Grades 5/6

Writing Measures

Reading Measures

Spelling ContextualConventions

Productivity ContextualLanguage

StoryConstruction

Story Schema

Word Attack ENG .72 .46 .21 .50 .37 .34

ESL .52 .40 .29 .43 .12 .05

WRAT Reading ENG .78 .47 .19 .54 .36 .26

ESL .58 .31 .09 .41 .13 .09

Fluency (Reading Time) ENG -.65 -.34 -.24 -.46 -.38 -.18

ESL -.69 -.17 -.11 -.31 -.20 -.05

Reading Comprehension ENG .63 .24 .29 .53 .51 .18

ESL .34 .36 .21 .14 .35 .20

Note: For ENG students df= 123, r = .21, p < .05; r = .24, p < .01; r = .34, p < .001; For ESL students d f = 4 8 , r = .31, p < .05 ; r = .40, p < .01; r = .52, p < .001;Based on Bonferroni adjustment, p < .001 in bold.

100

The Relation between Reading and Writing for Grade 5/6 ENG and ESL Students

The results of the correlational analysis presented in Table C-2 show that 19 of the partial

correlations (controlling for chronological age) between reading and writing measures in

grade 5/6 were statistically significant, with more for ENG (16) than for ESL (3). As was

found in grade 3, the strongest correlations were between the four reading measures and

spelling, for both grade 5/6 ENG and ESL, with the exception of higher-level Reading

Comprehension for ESL.

For ENG, every reading measure was significantly correlated (ranging from r = .3 - .5,

p < .001) with the higher-level writing measure, Story Construction. In addition, almost

every other writing measure was significantly correlated (ranging from r = .5 to .6,P<

.001) with the higher-level reading measure, Reading Comprehension, (with the

exception of Productivity and Story Schema). Furthermore, the higher-level reading and

writing skills themselves, Reading Comprehension and Story Construction, were

moderately correlated, r - .51, p < .001.

In contrast, there were no significant correlations beyond reading and spelling for ESL

students in grade 5/6. This finding is somewhat surprising given that, for ESL students in

grade 3, this was where the largest number of significant correlations was found.

In view of the pattern of correlations between reading and writing skills across Grade (3,

5/6) and Language (ENG, ESL), it is important to highlight how the lower- and higher-

level reading and writing measures correlate to address the second part of question 2

directly.

101


Appendix D

Correlating cognitive and language processes with lower- and higher-level reading

and writing skills

Tables D-l-4 present the partial correlations (controlling for chronological age) between

the cognitive and language processing measures and the lower- and higher-level reading

and writing skills. Tables D-l and D-2 show this relation with the cognitive measures

and Tables D-3 and D-4 show this relation with the language measures by Grade (3, 5/6)

and by Language (ENG/ESL). The results of the correlational analyses will be discussed

in the following order. First, the partial correlations (controlling for chronological age)

between the cognitive measures and the lower- and higher-level reading and writing skills

will be presented by Language (ENG/ESL) in grade 3 (Table D-l) and then grade 5/6

(Table D-2). Second, the partial correlations (controlling for chronological age) between

the language measures and the lower- and higher-level reading and writing skills will be

presented by Language (ENG/ESL) in grade 3 (Table D-3) and grade 5/6 (Table D-4).

Using the Bonferroni approach to control for Type 1 error across the many correlations, a

/?-value of less than .001 was required for significance. The correlations that reached this

level of significance are bolded in the tables.


Table D-l

Partial Correlations (controlling for chronological age) between lower- and higher-level Reading and Writing Measures and the underlying Cognitive Processing Measures for Grade 3 ENG and ESL students

Grade 3

Cognitive Measures

Lower- and Higher-TAAS RAN Sequences

#Forward

#Backward

VerbalLearning

SentenceMemory

StoryMemory

Level Measures

Word Attack ENG .58 -.48 .47 .28 .40 .06 .44 .24

ESL .54 -.35 .44 .19 .27 .18 .38 .06

Spelling ENG .41 -.43 .39 .11 .39 .09 .29 .22

ESL .55 -.35 .51 .36 .36 .24 .42 .08

Reading ENG .39 -.36 .47 .38 .46 .31 .54 .49Comprehension

ESL .32 -.42 .42 .35 .35 .18 .45 .40

Story ENG .19 -.17 .15 .13 .00 .33 .39 .29Construction

ESL .18 -.25 .41 .17 .32 .22 .25 .22

Note: For ENG students, r =.28-.33, p<.05; r =.36-.42, p < .01; r = ,46-.58, p < .001; For ESL students, r = .21-31, p < .05; r = .34 -.40, p < .01; r = .42-.55, p < .001; Based on Bonferroni adjustment, p <.001 in bold

Rep

rodu

ced

with

perm

issi

on

of the

co

pyrig

ht

owne

r. Fu

rther

re

prod

uctio

n pr

ohib

ited

with

out

perm

issi

on.

The Relation Between the Lower- and Higher-Level Reading and Writing Skills and the

Cognitive Processes for ENG and ESL students in Grade 3

The results of the correlational analyses presented in Table D-l show that 13 of the

partial correlations (controlling for chronological age) between Word Attack, Spelling

and Reading Comprehension and cognitive processes in grade 3 were statistically

significant, ranging from r = .46 - .59,/? < .001. There were a comparable number of

significant correlations for both ENG (7) and ESL (6 ).

Lower-level Word Attack was significantly correlated with cognitive measures (with the

TAAS and Sequences) for both ENG (r = .59,/? < .001; r = .47,/? < .05) and ESL (r =

•54,/? < .001; r = .44,/? < .05). In contrast, lower-level Spelling was significantly

correlated with cognitive measures for ESL only with the TAAS (r = .55, p < .001);

Sequences (r = .51 , P < .05); and Sentence Memoiy (r = .42, p < .001). Higher-level

Reading Comprehension was significantly correlated with four cognitive measures for

ENG and with one for ESL. There were no significant correlations with Story

Construction for either ENG or ESL.

Thus, at grade 3, lower-level Word Attack is the only measure that shows consistent

significant correlations with cognitive measures for both ENG and ESL students. Lower-

level Spelling has significant correlations with cognitive measures for ESL students only,

and Reading Comprehension has significant correlations with cognitive measures

predominantly for ENG students.


Table D-2

Partial Correlations (controlling for chronological age) between lower- and higher-level Reading and Writing Measures and the underlying Cognitive Processing Measures for Grade 5/6 ENG and ESL students

Grade 5/6

Cognitive Measures

Lower and Higher-

TAAS RAN Sequences #Forward

#Backward

VerbalLearning

SentenceMemory

StoryMemory

Level Measures

Word Attack ENG .56 -.52 .59 .38 .42 .29 .39 .18

ESL .56 -.35 .39 .27 .33 .07 .44 .12

Spelling ENG .43 -.46 .59 .32 .33 .30 .39 .24

ESL .25 -.43 .43 .12 .05 .11 .25 .20

Reading ENG .42 -.43 .42 .26 .28 .28 .47 .47Comprehension

ESL .37 -.17 .09 .01 .14 .31 .42 .49

Story ENG .29 -.29 .36 .22 .23 .21 .38 .21Construction

ESL .12 -.23 .14 .12 .17 .17 .07 .28

Note: For ENG students, r =. 21-.29, p < .01; r - .30-.58, p < .001; For ESL students, r = .30-.33, p < .05; r = .35-.44, p < .01; r = ,48-.56, p < .001Based on Bonferroni adjustment, p < .001 in bold

Rep

rodu

ced

with

pe

rmis

sion

of

the

copy

right

ow

ner.

Furth

er

repr

oduc

tion

proh

ibite

d w

ithou

t pe

rmis

sion

.


Cognitive Processes for ENG and ESL students in Grade 5/6

The results of the correlational analyses presented in Table D-2 show that 22 of the

partial correlations (controlling for chronological age) between lower- and higher-level

reading and writing skills and cognitive measures in grade 5/6 were statistically

significant, ranging from r — .30 - .59, p < .001. There were many more significant

correlations for ENG (20) than ESL (2).

For ENG, there were significant correlations between all four reading and writing

measures and the majority of cognitive measures. However, the pattern did show a larger

number of significant correlations with the lower- than the higher-level reading and

writing skills. For ESL, there were only two significant correlations, lower-level Word

Attack was significantly correlated with the TAAS (r = .56, p < .001) and higher-level

Reading Comprehension was significantly correlated with Story Memory (r =. 49, p <

.001) in contrast to the many significant correlations for ENG students.


Table D-3

Partial Correlations (controlling for chronological age) between lower- and higher-level Reading and Writing Measures and the underlying Language Processing Measures for Grade 3 ENG and ESL students

Grade 3

Language Measures

Reading and WritingLower- and Higher- Level Measures

PPVT EVT Syntax ListeningComprehension

OLPComposite

Word Attack ENG .24 .28 .34 .30 .41

ESL .27 .13 .21 .04 .20

Spelling ENG .23 .29 .34 .17 .37

ESL .28 .23 .41 .18 .36

Reading Comprehension ENG .39 .45 .54 .46 .66

ESL .56 .50 .42 .33 .56

Story Construction ENG .18 .14 .42 .25 .39

ESL .39 .28 .29 .22 .22

Note: For ENG students, r =.28-.30,/><05; r =.34-.42,/? < .01; r = A5-.66,p < .001;For ESL students, r =.26-.29 p < .05; r =.33-.41,/? < .01; r =.50-.56 , p < ,001;Based on Bonferroni adjustment,/? <.001 in bold

Rep

rodu

ced

with

pe

rmis

sion

of

the

copy

right

ow

ner.

Furth

er

repr

oduc

tion

proh

ibite

d w

ithou

t pe

rmis

sion

.


Language Measures for ENG and ESL students in Grade 3

The results of the correlational analyses presented in Table D-3 show that the 7 partial

correlations (controlling for chronological age) in grade 3 that were statistically

significant, ranging from r = .45 - .66, p < .001, were between the language measures and

higher-level reading comprehension. There were a comparable number of significant

correlations for both ENG (4) and ESL (3).

All of the significant correlations were between higher-level Reading Comprehension and

the language measures for both ENG and ESL. For ENG, there were significant

correlations between Reading Comprehension and Expressive Vocabulary (r = .45, p <

.001), Grammatical Judgment of Syntax (r = .54, p < .001), Listening Comprehension (r

= .46, p < .001), and the OLP composite (r = .66, p < .001). For ESL, there were

significant correlations between Reading Comprehension and Receptive Vocabulary (r -

.56,p < .001), Expressive Vocabulary (r = .50, p < .001), and the OLP composite (r =

.56, p < .001).

There were no significant correlations between lower-level Word Attack or Spelling, or

higher-level Story Construction and the language measures in Grade 3 for either ENG or

ESL students.


Table D-4

Partial Correlations (controlling for chronological age) between lower- and higher-level Reading and Writing Measures and the underlying Language Processing Measures for Grade 5/6 ENG and ESL students

Grade 5/6

Language Measures

Reading and WritingLower- and Higher- Level Measures

PPVT EVT Syntax ListeningComprehension

OLPComposite

Word Attack ENG .45 .36 .26 .31 .44

ESL .19 .16 .25 .13 .25

Spelling ENG .43 .45 .23 .28 .44

ESL .35 .26 .22 .11 .29

Reading Comprehension ENG .64 .63 .41 .54 .71

ESL .43 .51 .55 .59 .69

Story Construction ENG .36 .38 .28 .34 .44

ESL .17 .11 .04 .12 .13

Note: for ENG students, r = 2 3 -2 1 , p < .01; r =.31-.71, p < .001; For ESL students, r = .35, p < .01; r = .42-.69, p < .001Based on Bonferroni adjustment,/? <.001 in bold

Rep

rodu

ced

with

pe

rmis

sion

of

the

copy

right

ow

ner.

Furth

er

repr

oduc

tion

proh

ibite

d w

ithou

t pe

rmis

sion

.


Language Measures for ENG and ESL students in Grade 5/6

The results of the correlational analyses presented in Table D-4 show that 19 of the

partial correlations (controlling for chronological age) between lower- and higher-level

reading and writing skills and language measures in grade 5/6 were statistically

significant and ranged from r = .34 to .71 ,P< .001. As with the cognitive measures in

grade 5/6, there were a larger number of significant correlations for ENG (16) than ESL

(3).

For ENG, there were significant correlations between all four lower- and higher- level

reading and writing measures and the language measures. For ESL, there were only

significant correlations between higher-level Reading Comprehension and the language

measures.

Summary of the correlations between Lower- and Higher-Level Reading and Writing,

and Cognitive and Language Measures for ENG and ESL students

One of the most predominant patterns seen in these partial correlations between reading,

writing, cognitive and language measures was that Word Attack was the one measure for

which the cognitive measures were correlated for both ENG and ESL students in grade 3.

In addition, Reading Comprehension was the one measure for which many of the

language measures were significantly correlated for both ENG and ESL across grade (3

and 5/6). For ENG, the largest number of significant correlations was present in grade

5/6 with both cognitive and language measures. For ESL, the largest number of

correlations was present in grade 3, predominantly with the cognitive measures.

Given these differences in the correlations between the lower- and higher-level reading

and writing measures and the cognitive and language measures for ENG and ESL

students, it is valuable to examine how much of the variance is accounted for by these

underlying cognitive and language measures.


Appendix EFactor Analysis by Grade

Further exploration was undertaken within Grade (3, 5/6) to determine if the cognitive

and language factors that were used to create the composites were stable across

development. Factor loadings are presented by grade in Table 1 (Grade 3) and Table 2

(Grade 5/6).

Table 1 - Factor Loadings of the Cognitive and Language Processing Measures in Grade 3


Components

Measures 1 2 3



Story Memory .749


Syntax .621 .309

Sequences .747

Sentence Memory .515 .676

Working Memory .610

TAAS .670

RAN -.677




Table 2 - Factor Loadings of the Cognitive and Language Processing Measures in Grade 5/6


Components

Measures 1 2 3


Syntax .737



Sentence Memory .619 .408

Working Memory .699

TAAS .658

RAN -.610


Sequences .524

Story Memory .368 .6 6 6

Verbal Learning .637


These analyses indicate that the factors identified for the combined group are relatively

stable across development when split by Grade (3, 5/6).

112


Appendix F

Sequential Regression Split by Language (ENG/ESL) with OLP entered before Cognitive Ability

Since one could argue that the cognitive measures are confounded to some extent with

OLP, particularly for the ESL group, further exploration was undertaken with the OLP

composite entered before the Cognitive Ability composite. It is important to note

however, that the ability to develop language skills also partly depends on cognitive

ability. A comparison of the two methods indicates that OLP, as would be expected,

accounted for more of the variance when entered first, particularly for reading

comprehension, although word attack remained similar. In terms of writing measures,

spelling remained similar for ENG, and the two composites evened out in their

contribution for ESL. For story construction, the contribution of OLP was larger for

ENG, and both OLP and Cognitive Ability contributed for ESL. The pattern changed for

ENG story schema, but remained similar for ESL.

Table 1. Summary of the Percentage of Variance Accounted for by the Composites on the Lower- and Higher-Level Reading Skills for ENG (n = 179) and ESL (n = 110) Learners

Word Attack Reading ComprehensionR2 change________________ R2 change______

ENG ESL ENG ESLGrade 4 7 ** 6.3 ** 0 .1 0.11OLP 2i \ *** 7 2 * * 50.8 *** 36.2 ***Cognitive Ability j 2 q *** 20.3 *** 1.3 * 4 4 **Total Variance 37.7 33.8 52.2 40.7* p < .05, ** p < .01, *** p < .001

1 The fact that Grade did not account for any of the variance is not surprising given that two different tests were used for Grade 3 and 5/6. Thus, the difference between the two cannot be meaningfully compared.

113


Table 2. Summary of the Percentage of Variance Accounted for by the Composites on the Lower- and Higher-Level Writing Skills for ENG (n = 179) and ESL (n = 110) Learners


ENG ESL ENG ESLGrade 22.7 *** 40.2 *** 2 .6 * 9.8 **OLP 7̂ 4 *** 7 g *** 2 2 .8 *** 5.3 **Cognitive Ability 6 .6 *** 7 9 *** 0 .8 4.0*Total Variance 46.7 55.8 26.2 19.1* p < .05, ** p < .01, *** p < .001

Table 3. Summary of the Percentage of Variance Accounted for by the Composites on Story Schema for ENG (n = 179) and ESL (n = 110) Learners

Story Schema R2change

ENG ESLGrade 1 .8 7.9 **OLP 5.1 ** 6 .2 **Cognitive Ability 2 .0 0 .0

Total Variance 8.9 14.1* p < .05, * * p < .01, ***p < .001Sequential Regression by Language for Story Schema with the Cognitive Ability composite entered before OLP

Table 4. Summary of the Percentage of Variance accounted for by the Composites on Story Schema for ENG (n = 179) and ESL (n = 110) Learners


ENG ESLGrade 1 .8 7 9 **Cognitive Ability 4 4 ** .5OLP 2.7* 5.7*Total Variance 8.9 14.1*p < .05, * * p < .01, ***p < .001


Appendix G

Accounting for variance on the lower- and higher-level reading and writing

measures in the combined sample

Table 1 presents the results of the sequential regressions for the four lower- and higher-

level reading and writing measures. Cognitive Ability accounted for the most variance in

lower-level Word Attack, Grade (followed by Cognitive Ability) accounted for the most

variance in lower-level Spelling, while OLP accounted for the most variance in both

higher-level Reading Comprehension and Story Construction.

Table 1.

Summary of the Percentage of Variance Accounted for by the Composites on the Four Measures of Lower- and Higher-Level Reading and Writing Skills

Word Attack

R2change

Reading Comprehension

R2change

SpellingR2

change

Story Construction

R2changeGrade ̂2 *** 0 .0 27.0 *** 5 i ***

Language 0 .2 0.9 1.9 ** 0 .0

Cognitive Ability 2 4 .4 *** 1 2 .8 *** 13.8 *** ^ 4 ***

OLP 6 .0 *** 32.0 *** y 4 *** q 4 ***

Interactions 0.4 0 .1 0 .0 0 .6

(Lang X Cognitive Ability)(Lang X OLP)

Total Variance 36.0 45.8 49.1 21.5

* p < .05, **p < .01, *** p < .001

Lower-level reading measure - Word Attack

Both Cognitive Ability and OLP contributed significantly to the regression for Word

Attack. Grade (3, 5/6) contributed 5.2 percent of the explained variance (p < .001) while

Language did not. Cognitive Ability accounted for 24.4 percent of the variance in Word

115


Attack, and Oral Language Proficiency (OLP) accounted for a further 5.3 percent.

Cognitive Ability made the most overall contribution to Word Attack, accounting for

24.4 percent of the total 36.2 percent of the variance ip < .001). There were no

significant interaction effects for Language. These results were not surprising given that

Word Attack is a lower-level reading skill with demands on many cognitive skills such as

phonemic awareness, working memory, and sequencing needed to manipulate the sounds

of language in order to read phonemic approximations of pseudowords.

Higher level reading measure - Reading Comprehension

For higher-level Reading Comprehension, both Cognitive Ability and Oral Language

Proficiency composites contributed significantly to the regression. Grade and Language

did not account significantly for any of the explained variance. However, Cognitive

Ability accounted for 12.8 percent of the variance (p < .001) while the OLP composite

made the largest contribution, accounting for a further 34.2 percent (p < .001). There

were no significant interaction effects with Language. These results were not surprising

given that Reading Comprehension is a higher-level reading skill with complex language

and cognitive demands.

Lower-level writing measure - WRAT Spelling

Grade (3, 5/6) significantly accounted for 27 percent of the total variance (p < .001) in

lower-level Spelling. Language accounted for only 1.9 percent of the variance. Both

Cognitive Ability and OLP composites contributed significantly to the regression, with

the Cognitive Ability composite accounting for most of the variance, at 13.8 percent

(p < .001), and OLP a further 6 .6 percent (p <.001). No interaction effects with Language were found. It was not surprising that the Cognitive Ability composite

explained the most variance for Spelling, as it did for Word Attack, both of the lower-

level reading and writing skills. As with Word Attack, the speller was required to

manipulate, remember and sequence the sounds of language in order to spell both new

and already acquired words.

116


Higher level writing measure - Story Construction

For higher-level Story Construction, Grade accounted for 5.1 percent of the variance and

Language did not contribute significantly. The Cognitive Ability and OLP composites

contributed a further 6.4 percent, and then 9.9 percent. No interaction effects with

Language were found. The OLP composite explained the most variance in the higher-

level writing skill, as it did for the higher-level reading skill - Reading Comprehension.

117


Appendix H

Stepwise Regression Split by Language (ENG/ESL) on Story Construction

Further exploration was undertaken to explore how ESL students were able to achieve

comparably with ENG students on higher-level story construction despite significantly

lower OLP.

Table 1. Summary of the Percentage of Variance Accounted for by the Composites on Higher-Level Story Construction for ENG Learners

Story Construction R2change

ENGOLP 23.2 *Grade 2 .2 *Total Variance 25.4* p < .05, **p < .01, *** p < .001

Table 2. Summary of the Percentage of Variance Accounted for by the Composites on Higher-Level Story Construction for ESL Learners

Story ConstructionR2change

ESLOLP 13.3 *Cognitive Ability 4.0*Total Variance 17.3* p < .05, ** p < .01, *** p < .001

These results indicate that OLP accounted for less of the variance for ESL story

construction than it did for ENG. Oral Language Proficiency still accounted for more of

the variance overall in story construction for both ENG and ESL students. However, for

ENG students, grade accounted for the remainder of the explained variance, while for

ESL students, Cognitive Ability accounted for the remainder of the explained variance.

It may be the contribution of Cognitive Ability, and the smaller contribution of OLP that

distinguishes ESL from ENG students, thus allowing for comparable results on story

construction despite significant differences in OLP favoring ENG students.

118


Stepwise Regression Split by Language (ENG/ESL) on Story Schema

Table 3. Summary of the Percentage of Variance accounted for by the Composites on Story Schema for ENG Learners


ENGOLP 6 .8 *Total Variance 6 .8 ** p < .05, * * p < . 01, ***/?<.001

Stepwise Regression Split by Language (ENG/ESL)

Table 4. Summary of the Percentage of Variance Accounted for by the Composites on Story Schema for ESL Learners


ESLOLP 13.0*Total Variance 13.0* p < .05, ** p < .01, *** p < .001

These results indicate that OLP actually accounted for more of the variance for ESL story

schema than it did for ENG. This higher-level writing task tapped a different skill than

story construction.

Further exploration was initially attempted to determine the contribution of the various

components within the Cognitive Ability and OLP composites, however, given the

number of regressions undertaken and the small sample size, the analyses were not

included.


Appendix I

Sequential Regression Split by Grade (3,5/6) and Language (ENG/ESL) with

Cognitive Ability entered before OLP

Exploratory regressions were undertaken in order to determine if the structure of

predictors was stable across Grade. However, given the small sample size when presented

by Grade, the results of these regressions should be interpreted with caution. No firm

conclusions can be drawn.

Table 1. Summary of the Percentage of Variance Accounted for by the Composites on the Lower- and Higher-Level Reading Skills for ENG (n = 52) and ESL (n = 55) Learners in Grade 3

Word Attack R2change

Reading Comprehension R2change

ENG ESL ENG ESLCognitive Ability 24 3 *** 2 7 2 *** 29 4 *** 2 5 7 **OLP 9.5 ** 1 .2 32 5 *** 219 ***Total Variance 33.8 28.4 50.9 37.7* p < .05, * * p < .01, *** p < .001

Table 2. Summary of the Percentage of Variance Accounted for by the Composites on the Lower- and Higher-Level Writing Skills for ENG (n = 52) and ESL (n = 55) Learners in Grade 3

Spelling R2change


ENG ESL ENG ESLCognitive Ability 14.2 ** 34.8 *** 1.1 14.6**OLP 5.5 5.0* 20.6 *** 6.4 *Total Variance 19.7 39.8 2 1 .6 2 1 .0* p < .05, ** p < .01, *** p < .001

Table 3. Summary of the Percentage of Variance accounted for by the Composites on Story Schema for ENG (n = 52) and ESL (n = 55) Learners in Grade 3


ENG ESLCognitive Ability 0.3 2 .8

OLP 4.2 7.5 *Total Variance 4.5 10.3* p < .05, ** p < .01, *** p < .001

120


Sequential Regression Split by Grade (3,5/6) and Language (ENG/ESL) with

Cognitive Ability entered before OLP

Table 1. Summary of the Percentage of Variance Accounted for by the Composites on the Lower- and Higher-Level Reading Skills for ENG (n = 123) and ESL (n = 48) Learners in Grade 5/6



ENG ESL ENG ESLCognitive Ability 25.3 *** 28.7*** 1 0 7 *** 9.1 *OLP 1 0 l *** 3.5 42.5 *** 38.6 ***Total Variance 35.4 32.2 53.1 47.8* p < .05, * * p < .01, *** p < .001

Table 2. Summary of the Percentage of Variance Accounted for by the Composites on the Lower- and Higher-Level Writing Skills for ENG (n = 123) and ESL (n = 48) Learners in Grade 5/6

Spelling R2 change


ENG ESL ENG ESLCognitive Ability 22.2 *** 7.7 9.8*** 1.6OLP 13.0*** 7.5 16.3 *** .3Total Variance 35.2 15.2 26.2 1.9* p < .05, ** p < .01, *** p < .001

Table 3. Summary of the Percentage of Variance accounted for by the Composites on Story Schema for ENG (n = 123) and ESL ( n = 48) Learners in Grade 5/6


ENG ESLCognitive Ability 11.5*** .6

OLP 2.0 3.8Total Variance 13.5 4.5* p < .05, ** p < .01, *** p < .001


Sequential Regression Split by Grade (3,5/6) and Language (ENG/ESL) with OLP entered before Cognitive Ability

Table 1. Summary of the Percentage of Variance Accounted for by the Composites on the Lower- and Higher-Level Reading Skills for ENG (n = 52) and ESL (n = 55) Learners in Grade 3



ENG ESL ENG ESLOLP 2 i 9 *** 6.5 * 46 8 *** 31 7 ***Cognitive Ability 2 i 9 *** 4.1 * 6.0 *Total Variance 33.8 28.4 50.9 37.7* p < .05, ** p < .01, *** p < .001

Table 2. Summary of the Percentage of Variance Accounted for by the Composites on the Lower- and Higher-Level Writing Skills for ENG (n = 52) and ESL (n = 55) Learners in Grade 3


ENG ESL ENG ESLOLP 12.7 ** 14.8 ** 2 i 3 *** 13.4 **Cognitive Ability 7.0* 25.0 *** 0.3 7.6*Total Variance 19.7 39.8 2 1 .6 2 1 .0

* p < .05, ** p < .01, *** p < .001

Table 3. Summary of the Percentage of Variance accounted for by the Composites on Story Schema for ENG (n = 52) and ESL (n = 55) Learners in Grade 3

Story Schema R2 change

ENG ESLOLP 3.1 9.9*Cognitive Ability 1.4 0.4Total Variance 4.5 10.3* p < .05, ** p < .01, *** p < .001

122


Sequential Regression Split by Grade (3,5/6) and Language (ENG/ESL) with OLP entered before Cognitive Ability

Table 1. Summary of the Percentage of Variance Accounted for by the Composites on the Lower- and Higher-Level Reading Skills for ENG (n = 123) and ESL (n = 48) Learners in Grade 5/6

Word Attack R2 change


ENG ESL ENG ESLOLP 22.7 *** 10.4 * 52.5 *** 45.4 ***Cognitive Ability j 2 7 *** 218 *** 0 .6 2.4Total Variance 35.4 32.2 53.1 47.8* p < .05, ** p < .01, *** p < .001

Table 2. Summary of the Percentage of Variance Accounted for by the Composites on the Lower- and Higher-Level Writing Skills for ENG (n = 123) and ESL (n = 48) Learners in Grade 5/6


ENG ESL ENG ESLOLP 25.6 *** 1 0 .8 * 24 i *** 0 .8

Cognitive Ability g 7 *** 4.4 2 .1 1.1

Total Variance 35.2 15.2 26.2 1.9* p < .05, **p < .01, *** p < .001

Table 3. Summary of the Percentage of Variance accounted for by the Composites on Story Schema for ENG (n = 123) and ESL (n = 48) Learners in Grade 5/6

Story Schema R2 change

ENG ESLOLP 6.3 ** 3.0Cognitive Ability 7.2 ** 1.5Total Variance 13.5 4.5* p < .05, **p < .01, *** p < .001

123


These results indicate that there was basically no difference in the structure of the

predictors for word attack between grade 3 and 5/6 regardless of whether Cognitive

Ability or OLP was entered first. However, for reading comprehension, there was a larger

contribution of OLP in grade 5/6 than in grade 3, again, regardless of whether Cognitive

Ability or OLP was entered first. There were differences in the structure of the predictors

for both spelling, story construction and story schema, with no significant contributions

for ESL in grade 5/6 on these measures. It is important to note that caution must be taken

when interpreting these results given the small sample size when presented by grade.

124


The relation between reading and writing development in english and esl students

Education

esl reading

writing development

predictive of reading

story construction of

simple views of reading

esl language

copyright owner

authors permission