Effects of computer-assisted and teacher-led phonological awareness instruction for first-grade students at risk for rea...

MISSING IMAGE

Material Information

Title:
Effects of computer-assisted and teacher-led phonological awareness instruction for first-grade students at risk for reading failure
Physical Description:
ix, 275 leaves : ill. ; 29 cm.
Language:
English
Creator:
Travis, Penny Cynthia
Publication Date:

Subjects

Subjects / Keywords:
Special Education thesis, Ph.D   ( lcsh )
Dissertations, Academic -- Special Education -- UF   ( lcsh )
Genre:
bibliography   ( marcgt )
non-fiction   ( marcgt )

Notes

Thesis:
Thesis (Ph.D.)--University of Florida, 1997.
Bibliography:
Includes bibliographical references (leaves 262-274).
Statement of Responsibility:
by Penny Cynthia Travis.
General Note:
Typescript.
General Note:
Vita.

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 028638715
oclc - 38863581
System ID:
AA00014245:00001


This item is only available as the following downloads:


Full Text















EFFECTS OF COMPUTER-ASSISTED AND TEACHER-LED
PHONOLOGICAL AWARENESS INSTRUCTION FOR
FIRST-GRADE STUDENTS AT RISK FOR
READING FAILURE














By

PENNY CYNTHIA TRAVIS

















A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA

1997














ACKNOWLEDGMENTS

Many people contributed to the successful completion of

this study. The encouragement and support they offered were

invaluable. For this I am grateful and wish to express my

sincerest appreciation.

I want to thank Dr. Cecil Mercer, who chaired my

committee, and Dr. Mary Brownell, my cochair. Dr. Mercer

encouraged me to enter the doctoral program and provided

ongoing guidance and support. Dr. Brownell assisted me

throughout the study, offering creative ideas and valuable

resources. I thank Dr. David Miller for his good humor and

statistical expertise. I am grateful to Dr. Cynthia Griffin

and Dr. Cary Reichard for their advice and assistance,

especially at the proposal stage of this study. Special

appreciation must go to Dr. Holly Lane, who was, and is, a

mentor, a role model, and a friend.

I want to acknowledge the wonderful people at the

Multidisciplinary Diagnostic and Training Program (MDTP)

where I have held a graduate assistantship for more than 3

years. In particular, Dr. John Ross has given me continued

support and encouragement and provided resources that have

been invaluable in completing my research.

I could not have completed this study without the help

and cooperation of the administrators, teachers, and students

who were so much a part of it. Thanks go to Mr. Ken Blair,










Ms. Sandy Hollinger, and Dr. Fay Cake for allowing me to

conduct my research in their schools. Special thanks must go

to Ms. Terri Sapp, Ms. Sandy Medeiros, and Ms. Kathy Dixon.

Each served as my "contact person" at her respective school.

The first-grade teachers who allowed me to conduct my

research in their classrooms deserve acknowledgment: Ms.

Alford, Ms. Berry, Ms. Bruey, Ms. Collins, Ms. Emmanuel, Ms.

Green, Ms. Harner, Ms. Helm, Ms. Hunter, Ms. Mickle, Ms.

Rucker, Ms. Sapp, Ms. Travis, and Ms. Zant. I am grateful to

have worked with their students.

I wish also to express appreciation to the graduate

students who assisted me throughout the planning and

implementation of the intervention. Caroline Coxe and Alisa

Hadley developed an excellent instructional curriculum and

helped in many other areas too numerous to record here.

Thanks go to the students who assisted me in teaching the

interventions to children. I was continually impressed by

their professionalism, commitment, and caring: Ms. Andrea

Barron, Ms. Diane Coughlin, Ms. Jill Fuchs, Ms. Jannie Jones-

Wims, Ms. Holly Mallory, Ms. Cherie Manning, Ms. Joyce

Tullis, and Ms. Lynn Swartzwelder.

I am grateful for the financial support provided by the

Florida Educational Research Council (FERC). This grant,

awarded by the state of Florida, allowed me to pay a small

stipend to my assistants and to purchase materials and

supplies. Dr. Charlie Council, FERC Executive Director,


iii









helped guide me through the paperwork entailed in managing

this grant, and his assistance was most helpful.

There are others I want to thank. Dr. LuAnn Jordan was

always there to listen and offer help. Ms. Harriet Hayes

guided me through the IRB process with wit and humor. Ms.

Sybil Brown ably served as statistical consultant. Ms.

Barbara Smerage dispensed valuable advice along with

professional word processing. Dr. Joseph Torgesen was

generous with ideas and materials.

I thank Jeff for his patience and understanding and for

making me laugh. The good times we shared together helped

get me through the hard times.

Finally, I thank Stephanie and Justin, my most wonderful

and amazing children. Their love and support have meant

everything to me. I am as proud of their accomplishments as

I know they are of mine.


















TABLE OF CONTENTS

Daae

ACKNOWLEDGMENTS .......................................... ii

ABSTRACT .......................... ..................... vii

CHAPTERS

I INTRODUCTION TO THE PROBLEM ................... 1

Introduction .................................. 1
Rationale for the Study ....................... 5
Scope of the Study............................ 6
Definition of Terms............................ 7
Summary .. ................ ....................... 9

II REVIEW OF THE LITERATURE....................... 10

The Relationship of Phonological Awareness
and Reading Ability..... .................... 11
The Construct of Phonological Awareness....... 26
The Content of Phonological Awareness
Instruction.... ............................. 36
Direct Instruction and Phonological Awareness. 77
The Potential for Computer Assistance to
Overcome Barriers to Direct Instruction..... 88
The Purpose of the Study...................... 106
Summary of the Review of the Literature........ 107

III METHODS AND PROCEDURES........................ 109

Purpose ......... ...... ........................ 109
Research Questions and Hypotheses ............. 109
Methods and Procedures........................ 112
Summary ................ ....................... 143

IV RESULTS ................... .............. 146

Introduction .................. ... ............. 146
Reliability of Instructional and Measurement
Procedures................................... 147
Statistical Analysis of the Data.............. 150
Summary .................. ................... 165











V DISCUSSION .................................... 167

Summary of the Hypotheses and Results.......... 171
Discussion.................................... 175
Limitations of the Study....................... 190
Questions to Guide Future Research ............ 190
Summary ............................................ 196

APPENDICES

A PARENTAL INFORMED CONSENT...................... 198

B PHONOLOGICAL AWARENESS INSTRUMENTS ............ 200

C NONWORD READING EFFICIENCY TEST ............... 206
D WORD READING EFFICIENCY TEST.................. 210
E SCRIPTED LESSONS. ............................. 214
F REPRESENTATIVE PICTURES....................... 256

G REPRESENTATIVE GAME ........................ 260

H FIDELITY OF TREATMENT--OBSERVATION CHECKLIST.. 261

REFERENCES............................................... 262

BIOGRAPHICAL SKETCH .................................. 275
















Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment
of the Requirements for the Degree of
Doctor of Philosophy

EFFECTS OF COMPUTER-ASSISTED AND TEACHER-LED
PHONOLOGICAL AWARENESS INSTRUCTION FOR
FIRST-GRADE STUDENTS AT RISK FOR
READING FAILURE

By

Penny Cynthia Travis

December, 1997

Chairperson: Cecil D. Mercer
Cochairperson: Mary T. Brownell
Major Department: Special Education

Experimental, neurophysiological, and genetic research

support a phonological model of beginning reading. This

model provides a conceptual framework that demonstrates how

phonological processing affects reading and the importance of

phonological instruction to acquiring beginning reading

skills. The purpose of this study was to determine if

computer-assisted instruction in phonological awareness is as

effective as teacher-led instruction for first-grade children

with delayed development of phonological awareness. In an

experimental pretest-posttest-control group design,

participants (N=64) were randomly assigned to groups

receiving computer instruction, teacher instruction, or no

instruction in phonological awareness. At pretest, a series

of analyses of covariance (ANCOVAs) indicated no significant


vii









differences between the three groups on measures of blending,

segmentation, elision, and reading. At posttest, planned

orthogonal comparisons revealed no significant differences on

these measures between groups receiving computer-assisted

instruction and groups receiving teacher-led instruction.

The intervention groups made significantly higher pretest-to-

posttest gains than the control group on the measure of

segmentation. No significant differences between groups were

indicated on measures of blending, elision, or nonword

reading. These results were inconsistent with prior research

in which groups instructed in phonological awareness have

consistently demonstrated significantly higher increases in

phonological and reading skills when compared to control

groups receiving no phonological awareness instruction. The

results of the present study may be due to (a) the effects of

classroom reading instruction and (b) lack of transfer from

phonological to reading skills. Classroom reading

instruction that included blending skills may have been

sufficient for the control group to achieve progress

comparable to the intervention groups. Failure to transfer

their superior skill in segmentation to reading tasks may

have prevented the teacher-led group from significantly

increasing reading skills relative to the control group.

Implications for research include designing interventions to

determine the optimal duration and intensity of instruction

that facilitate transfer of phonological skills to reading.

The effects of metacognitive instruction in phonological


viii









awareness on transfer of skills should continue to be

investigated.















CHAPTER I
INTRODUCTION TO THE PROBLEM

Introduction

Two decades of research provide significant support for

the role that phonological awareness plays in the development

of reading ability (e.g., Adams, 1990; Stanovich, 1986, 1988;

Torgesen, Wagner, Simmons, & Laughon, 1990; Vellutino &

Scanlon, 1987; Wagner & Torgesen, 1987; Wagner, Torgesen, &

Rashotte, 1994). Correlational studies show phonological

awareness is positively related to reading ability; that is,

good readers have higher levels of phonological awareness

than poor readers (Bradley & Bryant, 1985; Calfee, Lindamood,

& Lindamood, 1973; Fox & Routh, 1975, 1980; Rosner & Simon,

1971; Tunmer & Nesdale, 1985; Vellutino & Scanlon, 1987).

Additionally, longitudinal studies demonstrate that

phonological awareness in young children predicts reading

ability (Blachman, 1984; Bradley & Bryant, 1985; MacLean,

Bryant, & Bradley, 1987; Mann, 1984; Mann & Liberman, 1984).

Moreover, a substantial number of studies have shown that

phonological awareness skills can be taught and that

increases in children's phonological awareness skills have

been consistently accompanied by improvement in reading

ability (Bradley & Bryant, 1985; Byrne & Fielding-Barnsley,

1991; Iverson & Tunmer, 1993; Lundberg, Frost, & Petersen,










1988; McGuinness, McGuinness, & Donohue, 1995; Williams,

1980).

Neurobiological findings support the existence of a

causal link between phonological awareness and reading.

Brain scans show tissue anomalies and differences in the

locus of metabolic activation during reading in normal

readers and persons with dyslexia (Galaburda & Geschwind,

1985; Lyon, 1995). Brain scan research is augmented by

genetic studies that provide a genetic explanation for

structural abnormalities of the brain (Cardon et al., 1994).

Genetic studies of persons with dyslexia suggest that

phonological awareness may be influenced by heredity (Lyon,

1995).

Taken together, experimental, intervention,

neurophysiological, and genetic findings support a

phonological model of beginning reading (Adams, 1990). This

model provides a conceptual framework that demonstrates how

phonological processing affects reading and suggests the

importance of phonological instruction to acquiring beginning

reading skills.

Because of the established relationship between

phonological awareness and beginning reading achievement,

instruction in phonological awareness is critical in the

primary grades. To design and measure effective instruction,

it is necessary to examine the construct of phonological

awareness (McBride-Chang. 1995; Yopp, 1988). Areas of










research suggesting that phonological awareness is not a

unitary construct but a multifaceted group of related skills

are (a) developmental evidence, (b) construct validation

evidence, and (c) training study evidence.

Developmental evidence indicates that younger children

have fewer phonological skills than older children. Tasks

are acquired in a predictable sequence based on the

complexity of the phonological task to be performed and the

level of the language component to be manipulated (Smith,

1995). Correlational studies provide further evidence of the

developmental nature of phonological awareness. These studies

show poor readers have fewer phonological skills than good

readers, regardless of age (Stanovich, Nathan, & Zolman,

1988).

Construct validation evidence (McBride-Chang, 1995;

Wagner, Torgesen, Laughon, Simmons, & Rashotte, 1993; Yopp,

1988) suggests that more than one phonological skill may

significantly affect reading performance. Two distinct but

correlated abilities composing phonological awareness have

been identified: analysis (segmentation) and synthesis

(blending).

Finally, training study evidence indicates both blending

and segmentation skills are necessary to affect reading. A

review of training studies that have had positive effects on

phonological awareness (for example, O'Connor, Jenkins,

Leicester, & Slocum, 1993; Slocum, O'Connor, & Jenkins, 1993;










Torgesen, Morgan, & Davis, 1992) suggests that tasks that

develop children's ability to segment and blend phonemes are

crucial to developing automatic and fluent word recognition.

To further determine effective intervention, it is

necessary to consider the method of instructional delivery.

Training study literature indicates that explicit or direct

instruction was most often used in studies resulting in

positive effects on phonological awareness and reading skills

(Fox & Routh, 1984; O'Connor, Notari-Syverson, & Vadasy,

1996; Weiner, 1994). This method of instruction has been

repeatedly demonstrated to assist high-risk learners in the

acquisition of basic skills.

Though direct, explicit instruction is necessary for

children who have not developed phonological awareness,

challenges exist to delivering this type of instruction in

regular classrooms. Providing explicit instruction that is

well planned, interactive, and closely monitored requires

teacher time spent in planning and managing instruction for

children who may represent only a small percentage of the

total classroom group. Even when time can be found to

implement such instruction, children do not necessarily

master skills as a group. Teachers may not wish to hold back

a group until every child has mastered each skill. To do so

is limiting to children who learn at a faster rate.

Providing for the needs of children who learn at slower rates

can be difficult in diverse classrooms. Furthermore,










paraprofessionals, volunteers, and other personnel who might

be available to provide help to teachers may not have

received training in this kind of intervention. Lack of

teacher time, insufficient training, and untested materials

can all affect the ability of the teacher to deliver

appropriate instruction to children at risk for reading

failure.

Rationale for the Study

Computer-assisted instruction in phonological awareness

may be a more viable alternative to classroom teachers than

direct instruction. Computer-assisted instruction has proven

effective in teaching a variety of reading skills (Jones,

Torgesen, & Sexton, 1987; Leong, 1995; van Daal & Reitsma,

1990; Wise & Olson, 1994), including those subsumed under

phonological awareness (Barker & Torgesen, 1995; Foster,

Erickson, Foster, Brinkman, & Torgesen, 1994; Uhry &

Shepherd, 1993). No research exists, however, on whether

computer-assisted instruction is as effective in increasing

phonological and reading skills as teacher-led instruction.

The purpose of this study was to compare the effects of

computer-assisted instruction, teacher-led instruction, and

no instruction on the phonological skills of children with

deficits in phonological awareness. The same phonological

skills were taught in each treatment group. Other variables

(such as duration, stimulus words, and reinforcement) were

identical. The salient difference between the two groups was










the medium of instructional delivery. By comparing the

results of computer-delivered and teacher-led instruction,

conclusions can be drawn about the relative effectiveness of

each form of intervention. If children learn skills equally

well or better when instructed by computer, teachers may be

justified in providing computer-assisted instruction as a

method for increasing phonological awareness.

Scope of the Study

In the following section, the delimitations and

limitations of the study are discussed. Delimitations and

limitations are important to consider when generalizing study

results to other populations.

Delimitations

Participants were first-grade students identified by a

phonological screening test as having deficiencies in

phonological awareness skills. Gender and ethnicity were not

considered in the selection process. Students identified by

the screening test whose parent or guardian gave permission

for participation in the study were pretested on measures of

phonological knowledge and word and nonword reading ability.

Participants were randomly assigned to one of three groups.

Two groups received instruction in phonological awareness

skills via the computer or in small teacher-led groups. The

third group served as a control group and received no extra

instruction. The research settings were three elementary

schools in north central Florida.










Limitations

This study was conducted with first-grade students

during the second half of the school year. The computer-

assisted instruction was conducted with the DaisvOuest (1993)

and Daisy's Castle (1993) software programs. Results of this

intervention may not be generalizable to other software

programs, especially if the content of such programs is not

based on prior research. The teacher-led instruction was

conducted by university graduate students outside the regular

classroom. Because this arrangement offered fewer

distractions to students, it may not be directly comparable

to small group instruction taking place within the classroom.

Thus, students in the direct instruction group may have

enjoyed a slight advantage over children instructed under

normal classroom conditions.

Definition of Terms

In this section key terms used in the chapters are

defined more fully, and others relevant to the conduction of

this study are discussed. Additional terms will be defined

in later chapters as they occur.

Metaohonoloav is the ability of an individual to reflect

on and consciously manipulate phonological segments and

phonemes and is also referred to as metaphonological or

metalinguistic skill (Hodson, 1994).

Onset and rime are defined as follows: The onset of a

syllable includes any consonant sounds that precede the










vowel, and the rime refers to the vowel and all succeeding

phonemes (Adams, 1990). In addition to analyzing and

synthesizing at the word, syllable, and phonemic level, the

tasks can be performed at the onset-rime level. For example,

the word cat segmented at the onset and rime level is /k/

(onset) and /at/ (rime).

A phoneme is the smallest linguistic unit. It refers to

one individual sound. The word cat has three phonemes, /k/--

/a/--/t/. The word see has two phonemes, /s/ and /e/. There

are 44 phonemes in the English language (Shaywitz, 1996).

Phoneme analysis is the ability to analyze words into

single sounds (Yopp, 1988). It includes the ability to

segment or separate linguistic elements such as words,

syllables, and phonemes. For example, to segment the word

cat, one could pronounce each of its phonemes separately

(/k/--/a/--/t/).

Phoneme synthesis is the ability to blend isolated

sounds into words (Yopp, 1988). It is demonstrated by tasks

in which linguistic segments are blended together to form

larger elements. Blending /k/--/a/--/t/ yields the word cat,

for example.

Phonics is used to describe a method of reading

instruction that teaches word recognition through learning

grapheme-phoneme associations (Adams, 1990).

Phonological awareness is one's sensitivity to, or

explicit awareness of, the phonological structure of the










words in one's language. It is demonstrated by tasks that

require children to identify, isolate, or blend the

individual phonemes in words (Torgesen, Wagner, & Rashotte,

1994).

Phonological Drocessing refers to using phonological

information to process oral and written language and

encompasses phonological awareness and metaphonology (Hodson,

1994).

Summary

The purpose of this study was to determine the effects

of computer-assisted instruction, teacher-led instruction,

and no instruction on the phonological awareness skills and

reading ability of first-grade students. Chapter II provides

the theoretical basis upon which interventions based on

phonological skills were grounded and contains a review of

the relevant literature on phonological awareness

intervention and computer-assisted instruction. Chapter III

describes the methods and procedures that were used to

conduct the study and analyze the resulting data. Chapter IV

reports the results of the study, and Chapter V contains a

discussion of the implications of the results for teachers,

teacher educators, administrators, and other educational

professionals.
















CHAPTER II
REVIEW OF THE LITERATURE

Chapter II provides the theoretical basis upon which

interventions based on phonological skills are grounded and

contains a review of the relevant literature on phonological

awareness intervention and computer-assisted instruction.

This chapter includes a summary and analysis of the research

examining the relationship between phonological awareness and

reading ability, the development of phonological awareness in

good and poor readers, and instruction in phonological

awareness. The literature on the efficacy of computer-

assisted instruction in phonological awareness is also

presented.

The chapter is divided into several sections. The

relationship of phonological awareness skills and reading

ability is discussed in the first section. The subsequent

sections provide a discussion of the specific skills that

make up the construct of phonological awareness, how

instruction can help students acquire these skills, and

challenges to implementing such instruction within the

general education classroom. The chapter concludes with a

summary of the research findings on the potential for

computer-assisted instruction in phonological awareness,










including the implications of previous research for this

study.

The Relationship of Phonolocical Awareness
and Reading Ability

In this section the relationship of phonological

awareness and reading ability is discussed. The section

begins with a short discussion of the three types of

experimental research that point to the causal relationship

of phonological awareness and reading ability. The results

of current neurophysiological and genetic research are

described, and phonological models of reading that interface

with this research are explained.

Evidence for a Causal Role

Phonological awareness is hypothesized by many

researchers to be a developmental skill that has a causal

relationship with reading ability. Two decades of

correlational, longitudinal, and training research provide so

much support to this hypothesis that the research findings

appear irrefutable.

Correlational studies. Correlational studies are based

on the view that phonological awareness is a developmental

skill (Bradley & Bryant, 1985; Calfee et al., 1973; Fox &

Routh, 1975, 1980; Rosner & Simon, 1971; Stanovich et al.,

1988; Stanovich & Siegel, 1994; Tunmer & Nesdale, 1985;

Vellutino & Scanlon, 1987). Thus, younger children should

have fewer phonological awareness skills than older children,

and poor readers perform similarly to younger children on










tests of phonological awareness. To test these hypotheses,

two types of studies were conducted.

One set of studies examined the correlations between

good and poor readers on tests of phonological awareness and

reading ability. For example, Calfee et al. (1973) tested

students from kindergarten through the 12th grade on their

ability to match phonemes using a sequence of colored blocks

to represent an auditory stimulus. A substantial correlation

at all grade levels was found between performance on this

auditory-phonetic test and reading ability as measured by the

Wide-Range Achievement Test (WRAT) (Jastak & Jastak, 1978).

In another correlational study (Tunmer & Nesdale, 1985),

first-grade children were given an assessment of phonemic

segmentation ability and reading achievement. Results

indicated that word segmentation ability was positively

related to reading achievement. Studies such as Calfee et

al. (1973) and Tunmer and Nesdale (1985) indicated a positive

relationship between the acquisition of phonological skills

and a child's age and reading ability.

A second set of correlational studies showed that older

and younger children matched on reading ability had similar

performances on tests of phonological awareness (for example,

Stanovich et al., 1988; Stanovich & Siegel, 1994). In the

reading-level matched design, an older group of children with

reading disabilities is compared with a younger group of

children without reading disabilities. Various tasks are

administered with the assumption that any differences that










are observed between groups cannot be the result of

differences in reading ability. For example, Stanovich et

al. (1988) tested groups of third-, fifth-, and seventh-grade

students who were matched on reading ability. The various

tasks administered included reliance on phonological

information in decoding. Results indicated that task

performances among the three grade level groups were similar.

In another study, Stanovich and Siegel (1994) used a

regression-based method analogous to the reading-level match

design. The researchers divided 907 children (ages 7-16

years) into three groups based on reading ability. The

children were compared on a set of tasks which included

phonological activities. The results generally supported the

phonological-core variable-difference model of reading

disability. Results of the study showed the critical

processing deficit impairing the word recognition process of

persons with reading disabilities lies in the phonological

domain, regardless of age or grade level.

Predictive studies. Predictive studies are based on the

assumption that children deficient in phonological skills at

the prereading stage will have difficulty acquiring reading

skills. Conversely, children who show early facility with

phonological tasks will more likely become good readers.

Mann (1984) showed that future reading performance could be

predicted by administering two tests of phoneme awareness,

phoneme segmentation and invented spelling. Each test was

given to 100 kindergarten children. One year later the same








14

children received standardized reading tests. Scores on each

test of phoneme awareness predicted between 30% and 40% of

variance in first-grade reading ability. Additionally,

Bradley and Bryant (1985) followed over 400 children on their

ability to detect rhyme and alliteration. Three and four

years later, the initial scores bore a consistent and

significant relationship to progress in reading and spelling.

These and similar studies (Blachman, 1984; MacLean et al.,

1987; Mann & Liberman, 1984) established a strong

relationship between phonological awareness and subsequent

reading skill.

Correlational and predictive studies by themselves,

however, are not conclusive of a causal relationship. Some

unknown factor that influences both phonological awareness

and reading ability may exist. Phonological skill could

influence reading ability, but the reverse might also be

true. Alternatively, the relationship might be reciprocal,

with each ability having an influence on the other. To

establish causality, intervention studies are needed.

Intervention studies. If phonological awareness causes

students to become proficient readers, then phonological

awareness interventions should improve reading skills. A

substantial number of studies have shown that phonological

awareness skills can be taught and that increases in

children's phonological awareness skills have consistently

accompanied improvement in reading ability. Lundberg et al.

(1988) used listening games, rhyming tasks, and segmentation








15

and blending of various linguistic components to increase the

phonological awareness skills of 235 Danish preschool

children. Bradley and Bryant (1985) trained children to

categorize words by their sounds. The training had positive

effects on reading and spelling skills. A host of other

studies that support these findings are discussed later in

this chapter.

The Neurobioloqg of Reading

In addition to the research supporting a causal link

between phonological awareness and reading, neurobiological

and genetic research support its existence. Although

neurophysiological and genetic factors have long been known

or suspected in cases of language and reading-related

problems, the means for studying brain structures were

lacking. With the advent of brain scan technology, research

examining the structural anomalies of dyslexics' brains and

differences in the locus of metabolic activity during reading

is possible and informative.

Neuroohvsioloaical research. When the brain tissue of

deceased dyslexic and nondyslexic individuals was examined

(Galaburda & Geschwind, 1985), the dyslexic brains were found

to have significantly more misplaced and unusually organized

nerve cells. The anomalous cells were more numerous in the

left hemisphere and included the language regions that border

the Sylvian fissure. The neural tissue present in the

temporal planum was symmetrical across the hemispheres of the










subjects with dyslexia, though the region is usually larger

on the left in the general population.

In a review of research currently in progress under the

aegis of the National Institute of Child Health and Human

Development, Lyon (1995) reported that brain scans showed

learning disabled groups tended to be more symmetrical. In

dyslexic individuals, the absence of the normal left-greater-

than-right asymmetry in the region of the posterior temporal

planum indicates atypical neural organization. Thus,

dyslexia is expressed in anomalous organization of the tissue

and processing systems of the posterior left hemisphere.

Shaywitz (1996) reported research employing functional

magnetic resonance imaging (fMRI) to measure changes in the

metabolic activity of the brain while an individual

identifies letters. Results showed that letter

identification activates sites located in the occipital lobe,

specifically the extrastriate cortex. Thus, it is believed

that phonological processing takes place within the inferior

frontal gyrus. Accessing the meaning of words appears to

activate the middle and superior temporal gyri of the brain.

Differences between males and females in the locus of

phonological representation were also reported. When

reading, males engaged the left inferior frontal gyrus,

whereas females activated not only the left but the right

inferior frontal gyrus as well.

As a result of the studies employing fMRI technology,

scientists have concluded that deficiencies in word










recognition skills are associated with less-than-normal

activation in the left temporal region and may be one

possible explanation for the predominance of reading

disabilities among men as compared to women. Additionally,

positron emission tomography (PET) studies indicate that

dyslexic adults have greater-than-normal activation in the

occipital and prefrontal regions of the cortex (Lyon, 1995).

New developments in brain scan technology have

facilitated the study of the structure and function of the

brain areas involving reading. These studies help us

understand the range of individual variability in brain

structure and function and aid in the identification of the

neurobiological causes of reading disabilities. In addition,

new technologies for examining an individual's genetic makeup

have allowed researchers to gather evidence about genetic

determinants of reading ability.

Genetic research. Scientists have long speculated about

a genetic relationship. For example, Locke (1994) proposed a

genetic explanation for the lack of planar asymmetry in the

brains of dyslexics. Locke's theory suggested that aspects

of language acquisition develop at critically timed phases.

These phases occur in sequence and are a result of stored

utterances which activate analytical mechanisms for

developing linguistic grammar. When the phases do not occur

on time, the critical period for optimal development passes.

Continued efforts to speak, however, stimulate atypical

allocations of neural resources into linguistic service.










Locke proposed that this activation leads to compensatory

growth in the right hemisphere which may account for the lack

of asymmetry characteristic of some dyslexic brains. Locke

speculated that genetic factors may influence the rate of

brain development in children and that it is slow

neurological maturation, not language development, that runs

in families.

Some support for Locke's (1994) theory of compensatory

growth in the right brain hemisphere came from investigations

of neural plasticity. Neural plasticity allows for the

modification of brain organization and function and is

observed in response to injury or damage to the brain. An

example of neural plasticity is the ability of

hemispherectomy patients to recover bilateral motor and

sensory functions. Recent experimental work has shown that

some of the mechanisms believed to operate in the recovery of

function can also be activated in the healthy brain during

development (Witelson, 1987). The implication for children

with reading disabilities is that intervention provided

during the crucial period of early development might be more

successful than remediation that occurs much later after the

child has failed to learn to read. The identification of a

genetic marker for reading disability might provide the means

for early diagnosis and intervention.

Scientists at the Institute for Behavioral Genetics at

the University of Colorado in Boulder have provided

convincing evidence for a genetic basis for reading










disabilities (Cardon et al., 1994). These researchers have

located a gene for reading problems on a small region of

chromosome 6, one of the 23 pairs of chromosomes that compose

the human genetic blueprint. The genetic studies in Colorado

were combined with a study of 19 families at the Center for

Hereditary Communication Disorders in Omaha. All of the

study participants had average or higher intelligence and

some type of reading disorder. The twins were ages 8 to 20,

and the families studied comprised three generations. The

twin study included one twin with reading problems and one

without. By comparing genes in blood samples taken from

participants with and without reading problems, the

researchers were able to link an abnormality in a segment of

chromosome 6 to reading problems. Strong evidence for

genetic etiology of reading disabilities points to

phonological awareness as reflecting the greatest degree of

heritability (Lyon, 1995).

Neurophysiological and genetic factors affecting reading

ability have long been hypothesized. Currently the means for

studying brain structures and function as well as the genetic

makeup of individuals with reading disabilities are

available. Lyon (1994) summarized the conclusions from this

research.

1. Dyslexia is related to anomalous organization of

tissue and processing systems in the posterior left

hemisphere.








20

2. Deficiency in word recognition skills is associated

with less-than-normal activation in the left temporal region.

3. Dyslexic adults have greater-than-normal activation

in the occipital and prefrontal regions of the cortex.

4. There is strong evidence for genetic etiology of

reading disabilities.

5. Deficits in phonological awareness reflect the

greatest degree of heritability.

6. At least one type of reading disability can be

linked to a region of chromosome 6.

The implications of these results are important for

children at risk for reading failure. Longitudinal studies

show that of those children who are reading disabled in the

third grade, approximately 74% remain disabled in the ninth

grade (Lyon, 1995). Because strong evidence now exists that

reading disabilities have a biological basis and represent a

persistent deficit rather than a developmental lag, early

identification and intervention are indicated. Brain scan

technology and genetic testing may eventually provide the

means to do so even before reading instruction begins.

Phonolocical Models of Reading

Taking into account results of experimental research and

current neurophysiological and genetic findings, researchers

have created phonological models of reading. These models

help us understand the crucial role of phonological awareness

in accomplishing the reading task and the importance of

phonological awareness instruction to acquiring beginning








21

reading skills. Reading is made up of more than phonological

awareness, however. In the following section, four reading

processors are described.

The four reading processors. Adams's (1990) model of

the reading processes involved in developing word knowledge

incorporated four processors: (a) the orthographic

processor, responsible for perceiving the sequences of

letters in text, (b) the meaning processor, containing

knowledge of word meanings, (c) the context processor, in

charge of constructing an on-going understanding of the text,

and (d) the phonological processor, which maps letters onto

their spoken equivalents. The processors work together in a

multidirectional manner, interacting by way of continuously

relaying information and feedback to each other.

The orthographic processor enables the reader to accept

information from print in the form of letter sequences. The

vast majority of print consists of a relatively small store

of very frequent words. These familiar words are processed

visually. Each time a particular letter-sequence is

encountered, the association between the letters is

strengthened. In time, after enough encounters with the

letter-sequence have occurred, the reader will perceive the

sequence holistically. The sequence will no longer have to

be processed letter-by-letter. The ability to process words

as wholes vastly improves reading speed and accuracy.

Familiar words and phrases can be processed automatically.

Unfamiliar words can be processed quickly by pronouncing the










familiar letter-strings. Even pseudowords, if made up of

familiar letter-sequences, can be easily and quickly decoded.

If the frequency of exposure to patterns of letter-

strings produces strong associations enabling automatic word

recognition, the opposite also holds true. Unfamiliar

strings are processed, if at all, slowly and sluggishly.

Less skilled readers who have not formed a store of

associative links between individual letter-recognition units

are slow and inaccurate decoders.

In order to make use of their store of letter

associations to decode words, letter-strings from the printed

page must first be encoded in their proper order. Knowledge

about letter sequences facilitates encoding. The visual

system can easily identify letters. Processing information

about their spatial locations is slow and difficult unless

they can be compared with previously formed letter sequence

associations in memory. These associative links aid in

encoding the spatial location of the identified letters.

Once the order of the letters is accurately encoded, decoding

of letter-sequences and whole words is facilitated.

There is at least one other service that letter-sequence

associations perform for the reader. These associations

allow longer words to be broken into syllabic units which can

be analyzed as wholes. Skilled readers break long words down

into syllabic units. They do so automatically and in the

course of perceiving them. It is the store of letter-

associated links that facilitates this automaticity.










In summary, the orthographic processor allows

automaticity with familiar words. It facilitates speed and

accuracy in decoding unfamiliar words. The inability of the

orthographic processor to recognize unfamiliar letter-

sequences explains the difficulty in decoding new or

irregularly spelled words. The orthographic processor makes

fast and accurate encoding possible and allows longer words

to be analyzed as syllables.

The orthographic processor interacts with the meaning

processor reciprocally. The meaning processor is activated

when recognition of whole words serve to turn on sets of

associated meaning elements, or concepts. Concepts are

formed by experience, and though differing among various

individuals, given concepts share common elements. Over

time, as an individual accumulates more and more experiences,

concepts change and expand. Activation of associated meaning

elements causes reciprocal activation of the spelling

patterns they require. In this manner, the association

between letter patterns and their meanings is strengthened

and reinforced.

The context processor interacts with the orthographic

processor and the meaning processor to construct an on-going

understanding of the text. In so doing, it activates all

meaning elements in the meaning processor that are compatible

with its ongoing interpretation. If the context is strongly

predictive of the word to follow, that word's meaning element

is strongly activated. If weakly predictive, the activation










is dispersed over many units in the meaning processor. The

context also plays a part in selecting among several

alternatives the appropriate meaning for any given word.

The phonological processor accepts information in the

form of speech. Less frequent words are analyzed by syllable

and translated automatically to their phonological

equivalents. The phonological translations serve both to

turn on the word's meaning and its spelling, and the more

such encounters occur with a word, the stronger direct

spelling to meaning connections become.

How the phonological model works. The language system

may be conceptualized as a hierarchical series of modules

(Shaywitz, 1996). The upper levels involve semantics (word

meanings), syntax (grammatical structure), and discourse

(connected sentences). The lowest level, the phonological

module, processes the sound elements (phonemes) that make up

language.

The phoneme is the fundamental element of the

linguistic system. For example, the word cat is made up of

three phonemes (/k/--/a/--/t/). To identify words, the

reader must break (or segment) them into individual sounds.

Once the word is in its phonological form, it can be

identified. Only after the phonological module identifies a

word can it be understood. Thus, word identification (or

decoding) is the fundamental task of the reader.

Both speech and reading rely on phonological

processing. In contrast to spoken language, which is








25

apparently acquired instinctively when humans are exposed to

speech, reading is an invention and must be learned at a

conscious level. Learning to speak gives no clue to the

segmental nature of language. In speaking, the individual

phonemes in words are coarticulated or merged into a single

unit of sound and automatically parsed back into words for

the listener. In reading, the individual phonemes must be

segmented for a word to be identified. The reader is faced

with the task of transforming printed symbols into linguistic

ones.

Though the orthographic, meaning, and context

processors play important parts in the reading process, it is

the phonological processor that plays the crucial role in

beginning reading. In order to decode unfamiliar words, the

phonological processor must analyze them by syllable. To do

so requires the understanding that words are made up of sound

units that can be separated (segmented) or merged (blended).

It is this knowledge that we call phonological awareness.

Similarly, what facilitates the strong direct spelling to

meaning connections that allow the orthographic processor to

decode words rapidly is knowledge of the mapping system, or

grapheme-phoneme correspondences in the alphabetic system.

Making sense of letter-sound instruction requires the

knowledge that words are made up of individual sound units

that can be segmented and blended. Therefore, the ability to

acquire the alphabetic system rapidly is highly affected by

phonological knowledge.











Summary of the Relationship of Phonolocical Awareness and
Reading Ability

The relationship of phonological awareness to reading

ability is documented by experimental research that shows its

causal role in the development of reading skills. Adding to

our understanding of this relationship is current research on

the neurobiological and genetic aspects of reading disorders.

Phonological models of reading help us understand the crucial

importance of phonological awareness to the reading task.

Because deficits in phonological knowledge represent the core

deficit in reading disability (Lyon, 1995; Perfetti, 1986),

phonological awareness is a crucial aspect of instruction for

beginning readers.

The Construct of Phonological Awareness

Because instruction in phonological skills is an

important part of beginning reading instruction, careful

consideration must be given to the construct of phonological

awareness. This issue is important because it affects

decisions about the content of interventions. If

phonological awareness represents a single ability, one

intervention task should suffice to improve global ability.

If the construct is actually a collection of several distinct

abilities, then more than one skill may need to be taught.

Three areas of research suggest that phonological

awareness consists of more than one skill. First,

developmental evidence supports the concept that different

phonological tasks present different levels of difficulty










(Smith, 1995) and that poor readers have fewer phonological

skills than good readers, regardless of age (Stanovich et

al., 1988). Second, construct validation evidence indicates

that more than one phonological skill may significantly

affect reading performance (Wagner et al., 1993). Finally,

training study evidence suggests that the most effective

interventions for increasing phonological awareness were

those that consisted of multiple tasks (Davidson & Jenkins,

1994; Hurford, 1990; O'Connor et al., 1993; Slocum et al.,

1993; Torgesen et al., 1992).

Development Evidence

A phonological task consists of manipulating sound units

(or language components) based on a given criterion. The

ability to perform the manipulation is assumed to demonstrate

phonological awareness. Adams (1990) identified five levels

of phonological tasks in order of their increasing

difficulty: (a) rhyming, (b) oddity tasks, requiring the

identification of rhyme or alliteration, (c) blending and

syllable splitting, (d) phoneme segmentation, and (e) phoneme

deletion. In the word oddity task, children are asked to

choose one word from a set of three or more that does not

belong according to a stated criterion. Word oddity tasks

may be based on rhyme (example set--get, mat, or pet) or

alliteration (example set--cat, cup, Dal). Syllable

splitting is performed at the onset-rime level. On hearing a

word, the child must say either the onset or the rime (hear

pair, say onset--/p/) or (hear pair, say rime--air). In










phoneme segmentation, individual phonemes must be isolated.

In phoneme deletion, one of the isolated phonemes must be

dropped (Say cat without /k/). The degree of difficulty of

a phonological task is affected by two variables: the

language component with which it must be performed and the

complexity of the task. Smith (1995) developed a

hierarchical framework of phonological tasks based on these

two variables.

Lanauaae components. In Smith's (1995) developmental

hierarchy of phonological tasks, the largest sound unit used

to perform most tasks is the word. Phonological tasks at the

word level do not require segmentation and, therefore, appear

to be easiest. Tasks become increasingly more difficult when

they must be performed at the syllable, onset-rime, and

phoneme levels, respectively.

Smith's (1995) hierarchy indicated that the easiest

divisions into which words can be separated are syllables and

that phonological tasks at the onset-rime level involve a

greater degree of abstraction, perhaps because, unlike

syllables, the division between onset and rime units cannot

be perceived in the normal pronunciation of a word. Adams

(1990) and Wagner et al. (1993) suggested a different

progression. They asserted that, in general, prereaders can

perform tasks that require segmentation of words based on the

intrasyllable units of onset and rime before they can perform

these tasks with syllables. These experts agree, however,

that the phoneme level requires the greatest level of










abstraction and, thus, is the most difficult level at which

to perform phonological tasks.

Task complexity. Tasks representing simple phonemic

awareness and compound phonemic awareness have been proposed

(Clarke-Klein, 1994; Yopp, 1988). Clarke-Klein described a

simple phonemic task, such as blending, as a task requiring

only one phonological manipulation. In blending, individual

phonemes must be coarticulated. A compound phonological task

consists of two manipulations. Phoneme reversal is a

compound task because it requires first that individual sound

units be isolated and then that they be manipulated. For

example, to perform the reversal task, one must say a word

backwards (pal becomes lap). To accomplish this, the

individual phonemes must first be segmented. Then the order

of the phonemes must be reversed.

In Smith's (1995) developmental hierarchy, easier

phonological tasks are designated as appropriate at the

preschool/kindergarten level. In order of increasing

difficulty, these tasks are (a) word oddity, (b) sentence

segmentation by words, (c) word segmentation by syllables,

and (d) sound matching. Sentence segmentation by words

involves counting words heard in a sentence. Word

segmentation by syllables consists of clapping or counting

syllables in a word. Sound matching may follow several

formats such as matching initial phonemes, matching final

phonemes, and matching medial phonemes.










More difficult tasks are appropriate at the

kindergarten/first grade level. These include (a) blending,

(b) word manipulation, (c) syllable splitting, (d) phoneme

segmentation, and (e) phoneme manipulation. Blending may be

performed at the syllable, onset-rime, or phoneme level.

Word manipulation, sometimes called elision, is a deletion

task that can be performed at the word level (Say baseball

without saying base), at the syllable level (Say winter

without saying ter), and at the onset-rime level (Say mat

without saying /m/). In phoneme manipulation, a deletion or

reversal task must be performed at the phoneme level.

Correlational studies such as those previously discussed

(Calfee et al., 1973; Stanovich & Siegel, 1994; Stanovich et

al., 1988; Tunmer & Nesdale, 1985) provide further evidence

of the developmental nature of phonological awareness. In

these studies, typically performing younger children had

fewer phonological skills than older readers. Older and

younger children matched on reading ability had similar

performances on tests of phonological awareness, and poor

readers had fewer phonological skills than good readers.

In summary, developmental evidence seems to suggest that

phonological awareness is not a unitary construct but may

consist of multiple skills representing distinct aspects of a

complex ability. In the following section, the existence of

phonological awareness as a unitary construct or a group of

distinct abilities is further explored.










Construct Validation Evidence

Clarke-Klein's (1994) description of simple and compound

phonological awareness was based on a principal factor

analysis of phonological awareness tasks (Yopp, 1988) in

which the two highly related factors emerged. Factor 1 tasks

require one operation (i.e., segment, blend, or isolate a

given sound). Factor 2 tasks require more steps to

completion. These tasks require that children perform one

operation (isolate a given sound) and then hold the resulting

sound in memory while performing yet another operation. In

the elision test, for example, the child must recall the

remaining sounds and blend them (Say "frog" without /r/).

Confirmatory factor analysis has been used in several

studies investigating the question of whether phonological

awareness is a unitary construct or consists of multiple and

distinct skills. Confirmatory factor analysis is an

application of structural equation modeling that combines a

structural model linking latent variables (unobservable

constructs) with observed variables providing reliable

measures of the latent variables (Moore, 1995). The

constructs are arranged in a model that shows how they are

related in a given theory. The hypothesized factor model

implies a structure in the data that may be compared to

actual data. To the degree that these two structures match,

a factor can be "confirmed."

Wagner et al. (1994) used confirmatory factor analysis

to compare alternative models of phonological processing










abilities in 184 kindergarten and grade 1 children. There

were five latent variables: (a) synthesis, (b) analysis, (c)

memory, (d) serial naming, and (e) isolated naming. The

synthesis and analysis tasks assessed phonological awareness.

Synthesis and analysis as two separate but correlated factors

were confirmed.

A longitudinal study of the phonological processing and

reading abilities of 244 kindergarten children (Torgesen et

al., 1994) supported a different conclusion. Torgesen et al.

used the five assessment tasks validated by Wagner et al.

(1994) to measure phonological processing skills. These

children were given a battery of 22 tasks that assessed the

five phonological abilities, reading and prereading skills,

and general verbal ability. The tasks were similar to those

in the Wagner et al. study. All tasks were readministered to

the same children at the beginning of first and second

grades. Performance on each phonological task had a

statistically significant influence on first-grade word-

reading skill when they were considered in isolation.

However, when the five phonological tasks were analyzed as

simultaneous causes, only one of the phonological variables

(phonological analysis, a measure of phonological awareness)

was significant. Performance on each phonological task had a

statistically significant influence on second-grade word-

reading skill also. However, when the five phonological

tasks were analyzed as simultaneous causes, only phonological

synthesis was significant. Torgesen et al. (1994) readily








33

conceded that had they included a more fine-grained analysis

of different types of reading skill (e.g., accuracy as

opposed to fluency), they might have been able to show that

their five phonological skills contributed differentially,

depending on the type of reading skill to be explained.

In McBride-Chang's (1995) analysis, phonological

awareness was measured with three separate tasks. These

tasks were position analysis, phoneme deletion, and phoneme

segmentation. The position analysis task required children

to identify the position of a phoneme in a nonsense word (Say

nelf. Does /f/ come before or after /1/?). Results showed

all three tasks were significantly related to phonological

awareness. Position analysis loaded highest (.84) on the

phonological awareness construct, and phoneme segmentation

loaded lowest (.78).

In summary, evidence exists for a number of correlated

factors that make up the construct of phonological awareness.

Results of research attempting to isolate those factors that

contribute significant variance to reading ability have been

inconclusive. The factors most often emerging in these

studies are phonological analysis and synthesis.

Training Study Evidence

Several studies have manipulated the combinations of

phonological awareness skills taught and compared groups in

an effort to determine which skills produced the greatest

effects on measures of phonological awareness and reading

(Davidson & Jenkins, 1994; Hurford, 1990; O'Connor et al.,










1993; Slocum et al., 1993; Torgesen et al., 1992). The

Torgesen et al. (1992) study included 48 kindergarten

children randomly assigned to three groups. The children

were matched on age and verbal ability. One group of

children learned blending and segmenting skills, while the

second group practiced blending only. The third group served

as a control group and received equal instructional time in

reading and writing activities similar to the Language

Experience Approach. Results indicated that instruction in

blending alone was not sufficient to affect reading as

measured by a reading analog test (using letter-like forms).

The group taught blending plus segmenting performed

significantly better than the control group. One limitation

of this study was the lack of a segmentation only group.

Davidson and Jenkins (1994) studied 40 kindergarten

children who were divided into four groups. Group A received

segmentation training only; group B received blending only;

and group C received training in both skills. The fourth

group served as a no-treatment control group. Results showed

that training in one skill did not seem to improve the other

skill. Significant effects on an analog word reading test

were noted for the segmentation-only group and the

segmentation-plus-blending group, but not for the blending

alone group.

Slocum et al. (1993) examined the effects of various

sequences of skill training with 35 prekindergarten children.

Those children taught blending and then segmenting were










compared with those who received segmenting training first

and then blending. As in the Davidson and Jenkins (1994)

study, Slocum et al. noted no indication of transfer across

phonological tasks. An analysis of the effects of sequence

of training showed no significant differences as measured by

onset-rime segmentation and blending.

In contrast, Hurford (1990) taught children auditory

discrimination of phonemes and found that the training

improved segmentation ability. The study suggests that

phonemic discrimination is a necessary yet insufficient skill

for phonological awareness. A weakness of the study was the

lack of a nonphonemic training group.

Summary of the Construct of Phonological Awareness

In summary, three sources of evidence provide support

for the conclusion that phonological awareness consists of

more than one skill. Evidence of the developmental nature

of phonological skills suggests the existence of a hierarchy

based on the level of difficulty of each skill. Construct

validation evidence indicates that more than one phonological

skill significantly affects reading performance. Finally,

training study research indicates that teaching one skill

does not necessarily result in an increase in other

phonological skills. Best practice seems to require that

interventions designed to increase phonological awareness

include a combination of phonological skills.










The Content of Phonolocical Awareness Instruction

If interventions including a combination of phonological

tasks are warranted, there is a need to determine those tasks

that have been most successful in increasing phonological

awareness skills. The following section includes a review of

the research on phonological training. The purpose is to

examine the results of such research to determine which

phonological skills should be included in an intervention to

increase phonological awareness.

Selection of Studies

Research published in refereed journals from 1970 to the

present was identified through the ERIC information resource

and by hand searching current issues of relevant journals.

Reference lists of selected articles were examined to

identify additional training studies to be included in this

review. Studies meeting the following criteria were

reviewed.

1. Participants and settings were thoroughly described.

2. Descriptions of treatments were detailed enough to

permit replication.

3. The experimental designs and data analyses were

sufficiently explained.

4. Interpretations were consistent with the results

displayed.

In accordance with these criteria, 45 articles

containing one or more intervention studies in the area of

phonological awareness were reviewed.










Early Studies

In the 1970s, evidence began to accrue that some

combination of phonological skills was necessary to enable

children to acquire basic literacy skills. Table 1

summarizes the research in phonological awareness that took

place in the 1970s.

Rosner (1974) investigated whether preschoolers could

learn auditory word analysis skills. Positive results from

this work were followed by other investigations in which

researchers sought to isolate the variables that affect the

acquisition of segmentation and blending skills. These

studies examined the effects of sequential memory skill

(Goldstein, 1976), phoneme position (Marsh & Mineo, 1977),

unit of analysis (syllable or phoneme level) (Liberman,

Shankweiler, Fischer, & Carter, 1974), and visual phoneme

representation with tiles or letters (Haddock, 1976; Marsh &

Mineo, 1977) on blending or segmenting training. Other

aspects of this research attempted to establish whether

segmentation skill is prerequisite to blending or if the

relationship is reciprocal (Fox & Routh, 1976;). The issue

of transfer to new word contexts was also investigated (Marsh

& Mineo, 1977). Significant increases in phonological skills

(Liberman et al., 1974; Marsh & Mineo, 1977; Rosner, 1974) or

reading ability (Fox & Routh, 1976; Goldstein, 1976; Haddock,

1976) were found in these early studies. The following is a

summary of their conclusions:










Table 1

Early Research in Phonological Awareness (1970s)


Study Subjects Dependent Measures Intervention Duration


40 Prekindergarten
Age 4
(average learners)


Reading
analog tasks


Proficient segmenters
Group 1. Blending instruction
Group 2. No blending


Nonsegmenters
Group 3. Blending instruction
Group 4. No blending instruction

Results: Significant positive gains in reading only for those children already proficient in segmentation.
No evidence of generalization from blending to segmentation.


Goldstein 23 Prekindergarten Sequential memory Group 1. Blending, segmenting 10 minute
(1976) Age 4.5 skill task & letter sound instruction sessions
Word analysis and Group 2. Sequencing tasks & 5 sessions
synthesis skill naming letters per week
tasks 13 weeks
Reading achievement
tests

Results: Group 1 produced significantly higher gains on reading measures.
Limitations: Small N, nonequivalent groups, and nonrandom assignment weaken the design of this study.


Fox &
Routh
(1976)


To criteria










Table 1--continued.


Study Subjects Dependent Measures Intervention Duration


Haddock 80 Prekindergarten
(1976) Age 5


Blending test
(word reading)


Group 1. Auditory blending &
letter-sound instruction
Group 2. Auditory blending using
manipulative letter tokens
Group 3. Letter sound instruction
only


Results: Significantly greater gains for Groups 1 and 2. Group 2 was significantly higher than Group 1.
Limitations: Unequal cell sizes. Did not assess segmentation skills. Results may have been due to
blending plus segmentation, not blending alone.


Liberman, 135 Preschool to Segmentation task Group 1. Tapping out number of 4 sets of
Shankweiler, grade 1 syllables and phonemes trials to
Fischer, & (average learners) (segmentation) criterion
Carter Group 2. No intervention
(1974)

Result: Fewer younger than older children reached criterion on segmentation task. Younger children
required more trials than older children to reach criterion.


10 minute
sessions
5 sessions
per week
3 weeks










Table 1--continued.


Study Subjects Dependent Measures Intervention Duration


Marsh & 64 Prekindergarten Phoneme identification Group 1. Phoneme identification 8 days
Mineo Age 5 during training given visual cueing (letters)
(1977) (average learners) Phoneme identification Group 2. Phoneme identification
during transfer to with colored cards represent-
new words ing letters

Results: There were no significant effects for phoneme position, transfer for visual (letter) cue, or
phoneme contrast. Continuants (/m/ /f/ /sl) were easier to learn than stop sounds (/p/ /t/ /d/).
Limitations: Short duration.


Rosner 145 Prekindergarten Criterion referenced Auditory analysis training As part of
(1974) & kindergarten test for each objective program (oral phonological reading
Age 4 & 5 in program tasks) curriculum
Group 1--inner city 2 school
Prekindergarten years
Group 2--Suburban kindergarten
Group 3--Inner city
kindergarten-control
(no intervention)
fReults: Significantly more objectives were mastered by Group 1 than Group 3. Significantly more
objectives were mastered by Group 2 than Group 1.
Limitations: Unequal cell sizes. Subjects were not randomly assigned or matched on ability.










1. Preschool children can be taught segmentation and

blending skills.

2. Some combination of blending, segmentation, and

letter-sound knowledge appears to enhance reading skills.

3. Segmentation at the phoneme level is more difficult

than at the syllable level and develops later.

4. There may be no transfer effects for training one

phonological skill to the acquisition of a second

phonological skill.

Components of Effective Interventions

Beginning in the 1980s, studies began to focus more

clearly on the components of effective intervention. The

most frequently included tasks in these studies were

segmentation, blending, and rhyming. Less frequently

included tasks were deletion, phoneme identification, and

phoneme substitution. The following is an overview of the

phonological training studies conducted from 1980 to the

present.

Longitudinal studies. Table 2 summarizes the

longitudinal training studies that have been conducted with

phonological awareness. Longitudinal studies are important

because they indicate effects of long-term training and show

whether training effects are maintained over time. These

studies (Bradley & Bryant, 1985; Byrne & Fielding-Barnsley,

1993; Foorman, Francis, Novy, & Liberman, 1991; Kozminsky &

Kozminsky, 1995; Lie. 1991; Lundberg et al., 1988; McGuinness

et al., 1995; Olofsson & Lundberg, 1985; Williams, 1980) have










Table 2

Longitudinal Training Studies


Study Subjects Dependent Measures Intervention Duration


Bradley & N = 65
Bryant Age 6
(1985) (average learners
with deficient
phonological
discrimination)


Sound categorization
(oddity task)
Neale Analysis of
Reading
Schonell Test of
Reading
Schonell Test of
Spelling


Group 1. Rhyme, alliteration,
global phonological awareness
(PA) activities
Group 2. Above combined with
manipulative letters
Group 3. Conceptual categori-
zation of words
Group 4. No intervention


Results: Group 2 made significant gains over Groups 3 and 4 on reading and spelling measures. No
significant difference between Groups 1 and 2 on reading measures. Group 2 significantly surpassed Group 1
in spelling.


N = 19
Age 6
(Australia)
(average learners)


Phoneme identity
Elision
Alphabet knowledge
Word Identification
(WRMT-R)
Pseudoword
Identification


Group 1. Identification of
phonemes and letter-
sound practice.
Group 2. Story reading,
semantic categorization
games


Results: Children with advanced levels of phonemic awareness (regardless of initial group membership)
scored significantly higher on each of the dependent measures. The authors concluded that some aspects of
phonemic awareness may be a consequence of literacy instruction rather than a cause.


1 session
per week
2 years


Byrne &
Fielding-
Barnsley
(1993)


One year
follow-up
of Byrne &
Fielding-
Barnsley
(1991)











Table 2--continued.


Study Subjects Dependent Measures Intervention Duration


Foorman, N = 80 Gates-MacGinitie Group 1. Letter-sound instruction 45 minutes
Francis, Grade 1 Reading Test w/segmentation & blending daily
Novy, Age 6 Segmentation test training 1 year
& Liberman (average learners) Researcher-constructed Group 2. Whole word in context
(1991) reading, spelling instruction
tests
Results: Growth curve analyses (using hierarchical linear model) showed Group 1 improved at a faster rate
than Group 2 for phonetically regular and exception words.


N = 150
(70 Kindergarten
53 Grade 1
27 Grade 3)
(Israel)


Lindamood Auditory
Conceptualization
(LAC)
Phonological Awareness
Test (PAT)
Reading Comprehension
Test


Group 1. Listening games, rhyming,
segmentation, blending,
deletion, letter-sound
instruction.
Group 2. Visual-motor integration
training.


Results: Significantly higher scores on the LAC for PA for the PA-trained group at all grade levels.
Initial sound isolation and sound deletion explained 70% of the reading comprehension score variability in
grade 1 subjects.


Kozminsky
&
Kozminsky
(1995)


20 minute
sessions
2 sessions
per week
8 months











Table 2--continued.


Study Subjects Dependent Measures Intervention Duration


Lie N = 208
(1991) Grade 1
Age 7
(Norway)


Reading test
Spelling test


Group 1. Segmentation--sequential
training (identification of
sounds in sequence in words)
Group 2. Segmentation--positional
training (identification of
beginning-middle-ending sound)
Group 3. Look at illustrations,
other "neutral" activities


Results: Those taught segmentation by identifying sounds in sequence made significant gains in reading at
the end of Grade 1. There was a significant main effect for IQ. Those with lower ability showed greater
effects on dependent measures.


Language completion
task
Vocabulary
Letter knowledge
Reading
Metaphonological tasks
(rhyme, segmentation,
synthesis, deletion)


Group 1. Various games such as
listening, rhyming,
segmentation, blending, &
letter-sound instruction
Group 2. Regular program--no
letter-sound instruction


Results: Significant gains for group 1 relative to control group on reading and spelling at grades 1 and 2.
Limitations: Nonrandom assignment. Effects on reading were only marginally significant.


Lundberg,
Frost, &
Petersen
(1988)


10-15
minute
sessions
5 sessions
per week
1 school
year


N = 390
Kindergarten
Age 6
(Denmark)


15-20
minute
sessions
5 sessions
per week
8 months
(follow-up
testing at
grades 1
and 2)











Table 2--continued.


Study Subjects Dependent Measures Intervention Duration


N = 42
Age 5-8
(above average
verbal ability)


LAC
WRMT-R-word attack,
word identification


Group 1. Auditory Discrimination
in Depth (ADD) program
(phoneme isolation and
identification) implemented
within whole-language
classroom
Group 2. ADD within traditional
classroom


Results: Comparable gains on the LAC and WRMT-R subtests were reported for both groups taught with ADD
method. ADD was found to be an effective means of delivering PA and letter-sound instruction in both whole
language and traditional classrooms.
Limitations: Small N for multiple regression analysis. Nonrandom assignment.


Olofsson & N = 83
Lundberg (follow-up of
(1985) 1983 study)
(Sweden)
(below average
phonological
awareness and
reading)


Synthesis
Analysis
Spelling
Reading


Group 1. Ryhming,
segmentation, phoneme
identification, deletion,
blending
Group 2. Nonverbal sound
discrimination
Group 3. Regular classroom
preschool program


Results: Long-term effects of PA training were demonstrated on PA tasks. Transfer to reading and spelling
was not demonstrated.


McGuinness,
McGuinness,
& Donohue
(1995)


20-30
minutes
daily
5 sessions
per week
8 months


Longitudinal
analysis
1 year
after
interven-
tion











Table 2--continued.


Study Subjects Dependent Measures Intervention Duration


Williams N = 127
(1980) Age 7-12
(students with
specific learning
disabilities)


Syllable analysis
Syllable blending
Letter-sound
correspondence
Decoding
Phoneme analysis
Phoneme blending


Group 1. Segmentation, blending,
letter-sound practice
Group 1. Traditional classroom
reading program


Results: At the end of year 1, significant gains were reported for Group 1 on measures of phonological
awareness. At the end of year 2, significant gains were reported for every measure, and significant
positive gains were noted for Group 1 versus the control group on all but one measure.


30 minutes
daily
26 weeks
2 years










included a wide array of phonological skills taught to

children over an extended period of time. The Lundberg et

al. study is representative of many of these longitudinal

studies. In this research, 390 Danish kindergarten children

(age 6 in Denmark) were trained over a period of 8 months.

The training program consisted of metalinguistic games and

exercises to stimulate children to discover and attend to the

phonological structure of language. These games and

exercises consisted of the following activities:

1. Listening games (nonspeech and speech sounds).

2. Identifying the rhyme and creating rhymes.

3. Identifying repeated words, dividing sentences into

their component words, and reflecting upon the length of

words.

4. Segmenting words into syllables or into subsyllabic

units.

5. Blending syllables into words, blending syllables

from different words, and blending subsyllabic units into

words.

6. Phoneme segmentation (initial, middle, and/or final

phoneme in a word).

7. Counting syllables, subsyllabic units, or phonemes

in a word.

8. Blending phonemes into a word.

9. Sounding the remaining part of a word following a

phoneme deletion.










Results of the Lundberg et al. (1988) study indicated

significant positive effects for the treatment group versus

the control group on phonological measures of rhyming,

blending, segmentation, and deletion. The difference at

grade 1 and grade 2 on reading skills between the

experimental and control groups was significant. A weakness

of the study was the nonrandom assignment of participants to

classrooms and the nonrandom assignment of classrooms to

groups.

Kozminsky and Kozminsky (1995) replicated the Lundberg

et al. (1988) study with 150 age 6 Hebrew-speaking children.

Children were randomly assigned to classrooms and groups.

Children in the experimental group engaged in activities

similar to those in the Lundberg et al. study. Children in

the control group were given activities promoting visual-

motor integration. Results showed children receiving the

phonological awareness training had significantly higher

scores than the control group on the phonological tasks of

the Lindamood Auditory Conceptualization (LAC) test. Initial

sound isolation (phoneme identification) and sound deletion

explained 70% of the reading comprehension score variability

in grade 1.

Longitudinal studies showed children could be taught a

broad array of phonological skills, that these skills

positively affected reading skill, and that the effects were

maintained over time. Other studies focused on teaching the

same phonological awareness tasks over a shorter period of










time. These studies can be organized into two categories:

(a) interventions consisting of phonological awareness

skills alone and (b) interventions consisting of phonological

awareness skills taught in combination with letter-sound

practice.

Teaching phonological skills alone. Teaching

phonological tasks without practice with letter-sound

correspondence has consistently resulted in increased

phonological awareness (O'Connor et al., 1993; Slocum et al.,

1993; Torneus, 1984). Other studies have shown that teaching

phonological skills alone has resulted in both increased

phonological awareness and reading skill (Brady, Fowler,

Stone, & Winbury, 1994; Cary & Verhaeghe, 1994; Davidson &

Jenkins, 1994; Fox & Routh, 1984; O'Connor et al., 1996;

Torgesen et al., 1992; Treiman & Baron, 1983). Table 3

summarizes this research.

An important aspect of this research focused on

determining which skills should be taught and in what

combinations. Fox and Routh (1984) compared children taught

segmentation alone to a group taught segmentation plus

blending. The segmentation plus blending group was

significantly better on measures of phonological awareness

and reading. A weakness of the study was the lack of a

blending only group.

Torgesen et al. (1992) compared a group taught

segmentation and blending to a group taught blending alone

and a control group. The children were matched across groups










Table 3

Studio P;


Training Phonolo l


Study Subjects Dependent Measures Intervention Duration


N = 42
Age 5-4
("at risk"--inner
city schools)


Achievement measures Group 1. Oral PA activities
Letter identification Rhyming
Word Identification Segmentation
(WRMT-R) Phonemic categorization
Word Attack (WRMT-R) ADD program
Spelling Segmentation using manipulative
Phonological measures tokens
Rhyme Group 2. Regular classroom
Segmentation instruction (whole language
Deletion approach)
Basic phonological
processes
Perception
Memory span
Speech production
Boston Naming Test


Results: No significant differences on achievement measures. Significantly higher gains for Group 1 on
rhyming, segmentation, and speech production. At follow-up (1 year later), Group 1 performed significantly
better on WRMT-R Word Identification and Word Attack.


Brady,
Fowler,
Stone, &
Winbury
(1994)


20 minute
sessions
3 sessions
per week
18 weeks
Follow-up
1 year
later


Awa -es Wi-l-t T----CmdPrc


nw~r~np~ Withnll+ T.P++Pr-9nrlnA Dr~r~;r~











Table 3--continued.


Study Subjects Dependent Measures Intervention Duration


Cary & Experiment 1
Verhaeghe N = 32
(1994) Age 5-6
(low SES)










Experiment 2
N = 47
Age 5-6
(low SES)


Rhyme
Segmentation of
sentences
Syllable deletion
Syllable synthesis
Syllable segmentation
Phoneme deletion
Phoneme segmentation
Awareness of
articulatory
gestures (position
of articulators)

Rhyme detection
Syllable deletion
Syllable segmentation
Phoneme deletion
Phoneme segmentation


Group 1. Games involving rhymes,
syllables, and phonemes
Group 2. As above, but without
games involving phonemes
Group 3. No intervention









Group 1. Phoneme analysis only
Group 2. Syllable analysis only
Group 3. Activities similar to
Groups 1 and 2 but with visual,
nonlinguistic material


Results: Group 1 displayed significant performance progress in all subtests. Group 2 displayed significant
performance progress in rhyme and syllable level tasks only. The results seem to indicate that facility
with tasks at the rhyme and syllable level does not generalize to phonemic tasks. In contrast, analysis
training with phonemes can enhance awareness at the rhyme and syllabic level. Thus, phonemic awareness at
the phoneme level must be explicitly instructed.


30 minute
sessions
2 sessions
per week
7 weeks









30 minute
sessions
2 sessions
per week
8 weeks










Table 3--continued.


Study Subjects Dependent Measures Intervention Duration


Davidson & N = 40
Jenkins Kindergarten
(1994) Age 6


Phonemic Group 1. Blending only
generalization tests Group 2. Segmentation only
(blending, Group 3. Both blending &
segmentation) segmentation
Transfer tasks Group 4. No intervention
(reading & spelling)


Results: Blending test: Group 1 and Group 3 made significantly higher gains than Groups 2 and 4.
Segmentation test: Group 2 and Group 3 made significantly higher gains than Group 1 and Group 4.
Reading and spelling measures: Group 2 and Group 3 made significantly higher gains than Group 1 and 4.
This study found transfer on a word reading analog test for Groups 2 and 3. This study found no evidence of
generalization from blending training to segmentation or vice versa.


Fox & N = 41
Routh 31 Kindergarten
(1984) "nonsegmenters"
Age 5-6
10 Kindergarten
"segmenters"
(average learners)


Segmentation
Blending
Reading analog tasks


Group 1. Segmentation practice
Group 2. Segmentation plus
blending practice
Group 3. No intervention


Results: Group 2 was significantly higher on all tasks. No generalization from segmenting to blending was
indicated.
Limitations: Study lacked a blending only group.


10 minute
sessions
5 sessions
per week
8-12 weeks
(to cri-
terion)


15 minute
sessions
4-5 sessions
per week
5 weeks











Table 3--continued.


Study Subjects Dependent Measures Intervention Duration


N = 47
Prekindergarten
Age 4-6
(special education
students)


Rhyming
Blending
Segmenting


Group 1. Rhyming practice
Group 2. Blending practice
Group 3. Segmentation practice
Group 4. No intervention
(proficient PA skills)


Results: Significant gains for Group 2 and Group 3 on all measures compared to Groups 1 and 4. Little
transfer between blending and segmentation tasks.


N = 107
Kindergarten
Age 5-6
31 with disabilities
57 regular education,
low SES
19 Retained


Sound repetition, Groi
letter naming, p1
Syllable deletion, (
rhyming, blending r
identify first sound, s<
segmentation, Woodcock- Grol
Johnson Tests of
Achievement-Revised
(WJ-R)
Letter-Word, Identification
subtest, WJ-R Dictation
subtest


up 1. Combination of
honological activities
word & syllable awareness,
hyming, blending,
segmentation)
up 2. No intervention


Results: Significant positive effects for Group 1 for blending, segmenting, reading, and writing measures.
Children with disabilities made fewer gains than retained kindergartners.
Limitations: Retained kindergartners' gains may be due to small amounts of reading instruction and second
year of exposure to kindergarten objectives. Ceiling effects for PA measures were reported.


O'Connor,
Jenkins,
Leicester,
& Slocum
(1993)


O'Connor,
Notari-
Syverson,
& Vadasy
(1996)


10 minute
sessions
4 sessions
per week
7 weeks


5-15 minutes
daily
6 months










Table 3--continued.


Study Subjects Dependent Measures Intervention Duration


N = 35
Prekindergarten
Age 5
(low receptive
language)


Segmentation
Blending


Phase 1
Group 1. Blending then segmenting
instruction
Group 2. Word manipulation then
segmenting


Phase 2
Group 1. Segmenting then blending
instruction
Group 2. Word manipulation then
blending.

Results: Phase 1: Group 1 performed significantly better on the blending test only. Phase 2: Group 1
performed significantly better on the segmentation test only. No indication of transfer across phonological
tasks was noted.
Limitations: Floor effects prevented detection of small gains so effect sizes could not be established.


Segmentation
Blending
Letter-sound
Word-learning
synthesis (reading)


Group 1. Blending alone
Group 2. Segmentation + blending
Group 3. Language Experience
Approach


Results: Group 1 was significantly higher than Group 3 in blending only. Group 2 was significantly higher
than Group 3 in segmentation and blending. Group 2 was significantly higher than Group 3 on reading.
Learning to blend did not transfer to skill in segmentation.
Limitations: The blending plus segmenting group received an extra week of the intervention. The study
lacked a segmentation only group.


Slocum,
O'Connor,
& Jenkins
(1993)


10 minutes
sessions
5 sessions


Torgesen,
Morgan, &
Davis
(1992)


N = 48
Age 6
(at risk)


20 minute
sessions
3 sessions
per week
7-8 weeks











Table 3--continued.


Study Subjects Dependent Measures Intervention Duration


Torneus N = 38
(1984) Grade 1
Age 7
(average learners)
(Sweden)


Segmentation
Sound blending
Reading task
Spelling task


Group 1. Phonological awareness
activities (rhyme, alliteration,
segmentation, blending)
Group 2. General language
activities


Results: Group 1 improved relative to Group 2 on all measures. The lowest skilled children in Group 1 at
pretest improved in PA and spelling but not in reading.
Limitations: Ceiling effects on segmentation test. Short duration.


Treiman & Experiment 1
Baron N = 8
(1983) Kindergarten
Age 5

Experiment 2
N = 20
Kindergarten
Age 6


Reading Task


Experiment 1
Group 1. Analysis training
Group 2. No analysis training


Experiment 2
Group 1. Analysis + blending
Group 2. No analysis or blending


Results: In both experiments 1 and 2, a significant interaction was found between group and item type in
the reading task, suggesting a causal link between the ability to analyze spoken syllables and the ability
to benefit from spelling-sound relations in reading.
Limitations: No synthesis (blending) alone condition.


20 minute
sessions
1 session
per week
8 sessions


5 days
(to
criterion)










on age and verbal ability. Measures of segmentation,

blending, and reading indicated significant positive effects

for the blending-only group versus the control group on the

blending test only. Significant positive effects were found

for the segmentation-plus-blending group on all three

measures. The researchers concluded that teaching blending

plus segmentation was sufficient to significantly affect

reading skills. As previously discussed, a limitation of the

study was the lack of a segmentation-only group.

A similar result was obtained in a study that taught a

different skill--blending, segmentation, or rhyming--to each

of three groups (O'Connor et al., 1993). They were compared

with a control group receiving no intervention. Based on

measures of segmentation, blending, and rhyming, only the

segmentation and blending groups made significant gains in

all areas.

The Davidson and Jenkins (1994) study discussed

previously included three groups--segmentation only, blending

only, and segmentation plus blending. The study also

included a no treatment control group. The segmentation-

plus-blending group made significant gains on both measures

of phonological awareness. On the word reading measures, the

one group failing to make significant increases was the

blending-only group.

In their examination of the importance of sequencing in

phonological training, Slocum et al. (1993) taught blending

and then segmenting to one group of at-risk prekindergarten










children and segmenting and then blending to the second

group. Those taught blending performed significantly better

on the blending test. Those taught segmenting performed

significantly on the segmenting test. The order of

presentation of the skills had no effect. The researchers

reported floor effects that may have prevented detection of

small gains. They suggested that low verbal skills and lack

of motivation of participants may have affected results.

Summary of interventions emploving Phonological

instruction alone. Interventions instructing children in

phonological awareness have been successful in increasing

phonological and reading skills. The majority of significant

effects on phonological and reading skills have been found in

groups taught segmentation-plus-blending skills.

Phonological skills combined with letter-sound training.

Further training study evidence suggests that phonological

awareness instruction may be effective when combined with

letter-sound practice. Combining segmentation training with

instruction and practice in letter-sound correspondence has

been investigated in several studies (Ball & Blachman, 1988;

Bentin & Leshem, 1993; Blachman, Ball, Black, & Tangel, 1994;

Byrne & Fielding-Barnsley, 1991; Defior & Pio Tudela, 1994;

Tangel & Blachman, 1992; Vellutino & Scanlon, 1987). Table 4

summarizes this research.

In two studies representative of this type of

intervention, Blachman et al. (1994) and Tangel and Blachman

(1992) taught segmentation in combination with letter-sound











Table 4

Studies Training Phonological Awareness and Letter-Sound Practice


Study Subjects Dependent Measures Intervention Duration


WRMT-R
Word Identification
Word Attack
LAC


ADD program
(phoneme isolation and
identification)


Results: Significant gains on both measures of reading and the measure of phonological awareness.
Limitations: This study lacked a control group but is often cited because of the severity of participants'
reading disability.


Ball & N = 90
Blachman Kindergarten
(1988) Age 6
(average learners)


Segmentation test
Letter naming
WJ-R Word
Identification
Word List


Group 1:
Segmentation &
letter-sound practice
Group 2:
Language
awareness activities
Group 3:
No intervention


Results: Significant gains for Group 1 on segmentation and word reading measures. No differences in letter
naming.


Alexander,
Anderson,
Heilman,
Voeller, &
Torgesen
(1991)


N = 10
Age 8-13
(severe
dyslexics)


Total 38-124
hours


20 minute
sessions
4 sessions
per week
7 weeks










Table 4--continued.


Study Subjects Dependent Measures Intervention Duration


Ball & N = 89 Segmentation Group 1. Segmentation, letter- 20 minute
Blachman Kindergarten Letter-names sound practice, alliteration, sessions
(1991) Age 6 Letter sounds spelling by sounds, rhyming 4 sessions
(average learners) WRMT-R Word Group 2. Letter-sound practice per week
Identification combined with language 7 weeks
Word reading activities
Spelling Group 3. No intervention

Results: Significant gains for Group 1 on segmentation and reading tasks.
Limitations: Subjects may have received blending training in classroom program.


Bentin & N = 508
Leshem Age 5.5
(1993) Israel
(lower quartile
on PA)


Phoneme isolation
Phoneme matching
Phoneme addition
Phoneme deletion
Reading tests


Group 1. Phonemic segmentation
with colored tokens
Group 2. Phonemic segmentation
with letter shapes
Group 3. General language skills
Group 4. (Low PA at pretest)
No intervention
Group 5. (High PA at pretest)
No intervention


Results: Reading performance assessed 4 and 9 months after training showed similar results for Groups 1 and
2. These two groups performed significantly better than Groups 3 or 4 on all measures and similar to Group
5. The authors concluded that learning to read is the most important factor affecting the acquisition of
phonemic segmentation skills in first grade. The emergence of phonemic awareness requires a previously
developed sensitivity to phonology, and this development by explicit training can prevent reading failure.


30 minute
sessions
2 sessions
per week
10 weeks
Follow-up at
4 and 9
months










Table 4--continued.


Study Subjects Dependent Measures Intervention Duration


N = 159
Kindergarten
(average learners)


Segmentation Group 1. Segmentation and
Letter names and sounds letter sound practice
WJ-R Word Group 2. No intervention
Identification
Word Attack
Word Reading
Spelling


Results: Group 1 gained significantly over Group 2 on all measures except the WJ-R Work Attack Subtest
(Group 2 read more nonwords than Group 1).
Limitations: Assignment to groups was not strictly random.


Byrne & N = 126 Phoneme identification Group 1. Identification of 25-30 minute
Fielding- Age 5 tasks phonemes and letter-sound sessions
Barnsley (average learners) Letter-sound recognition practice 1 session
(1991) Word reading Group 2. Story reading, semantic per week

categorization games 12 weeks
Results: Significant overall increases in PA and reading. Group 1 had greater gains than control.


Blachman,
Ball,
Black,
& Tangel
(1994)


15-20 minute
sessions
4 sessions
per week
11 weeks










Table 4--continued.


Study Subjects Dependent Measures Intervention Duration


Castle, Experiment 1 Experiment 1 Experiment 1 Experiment 1
Riach, & N = 30 Phonological Awareness Group 1. Segmentation, deletion, 20 minute
Nicholson Age 5 tasks, WRAT, substitution, rhyme, letter- sessions
(1994) spelling words, sound practice 2 sessions
dictation, written Group 2. Process writing per week
vocabulary, letter with invented spelling 10 weeks
knowledge

Experiment 2 Experiment 2 Experiment 2 Experiment 2
N = 51 Phonological Awareness Group 1. Segmentation, blending, 20 minute
Age 5 tasks, letter alliteration, rhyming, sessions
knowledge, word deletion, letter-sound practice 1 session
(both groups were reading, pseudo Group 2. Meaning-based per week
below average in words, dictation, instruction (categorizing 15 weeks
phonological written vocabulary, semantic cues, letter
awareness) print concepts names only, listening to
stories

Results: Experiment 1: Greater effects for Group 1 on PA measures and WRAT spelling.
Experiment 2: Significant Group 1 effects for all measures.
Limitations: Control group had higher cognitive skills. Effects were small.










Table 4--continued.


Study Subjects Dependent Measures Intervention Duration


Phoneme discrimination
Reading
Writing
Mathematics


Group 1. Categorization of words
based on phonemes
Group 2. As above but including
the use of plastic letters
Group 3. Categorization of words
based on concepts
Group 4. As for Group 3 but
including written words as
and aid
Group 5. No intervention


Results: Significant effects on reading, decoding, and writing measures for Group 2 when compared with all
other groups. Effects were maintained at follow-up assessment 2 months later.


N = 48
Grade 1
(average learners)


Word lists


Group 1. PA activities at phoneme
level
Group 2. PA activities at onset-
rime level
Group 3. Whole-word instruction
(meaning based)
Group 4. No intervention


Results: Those taught PA at the onset rime and phoneme level performed significantly better on reading
measures than those taught with the whole word method or those given no intervention. There was no
significant difference in performance between Group 1 and Group 2.


Defior
Pio
Tudela
(1994)


N = 55
Age 6
Spain
(low PA)


90 minute
sessions
1 session
per week
20 weeks


Haskell,
Foorman,
& Swank
(1992)


20 minute
sessions
2 sessions
per week
6 weeks










Table 4--continued.


Study Subjects Dependent Measures Intervention Duration


Hohn & N = 24 Segmentation Group 1. Segmentation using 20 minute
Ehri Kindergarten Deletion letter tokens sessions
(1983) Age 6 Decoding Group 2. Segmentation using 5 sessions
(nonsegmenters) blank tokens per week
Group 3. No intervention

Results: Significant effects for Groups 1 and 2 on all PA measures. No significant differences were noted
on reading measures.


O'Connor & N = 10 Blending, segmentation Group 1. Segmentation and letter- 10 minute
Jenkins Kindergarten Spelling sound practice (spelling with sessions
(1995) Age 5-6 WRMT Word magnetic letters and writing 20 sessions
(developmental delays) Identification words)
WRMT Word Attack Group 2. Reading words only

Results: Group 1 made significant gains over Group 2 in word identification and spelling real words and
nonwords. No differences in segmentation or blending.
Limitations: Small N. Subjects probably had segmenting and blending ability greater than usual for
kindergarten due to phonics-based reading curriculum.










Table 4--continued.


Study Subjects Dependent Measures Intervention Duration


N = 67
Kindergarten
(language delays,
Exceptional Student
Education (ESE)
identified and ESE
referred)


Rapid letter naming
Syllable deletion
Rhyme production
Segmentation, Blending
Letter-naming
Word identification
LAC, Reading
analog trials


Group 1. Segmentation and
blending combined with
letter sound practice
Group 2. Global PA activities
combined with letter sound
practice
Group 3. Letter-sound practice
Group 4. No intervention


Results: No differences were reported between Groups 1 and 2 on PA skills or a reading analog test. Both
performed better than the letter-sound only group.


Olofsson & N = 95 Researcher-developed Group 1. Rhyming, segmentation, 30 minute
Lundberg Kindergarten phoneme synthesis and phoneme identification, sessions
(1983) Age 6-11 phoneme analysis, and deletion, blending, letter-sound 3-4 sessions
(Sweden) reading measures practice per week
(below average Group 2. Nonverbal sound 6-8 weeks
in phonological discrimination
awareness and Group 3. Traditional preschool
reading) program

Results: Only Group 1 showed significant improvement across tasks.
Limitations: Nonrandom assignment. Normality assumption violated. Ceiling effects.


O'Connor,
Jenkins, &
Slocum
(1995)


15 minute
sessions
2 sessions
per week
10 weeks










Table 4--continued.


Study Subjects Dependent Measures Intervention Duration


Tangel & N = 149
Blachman Age 5-6
(1992) (average learners)


Phoneme segmentation Group 1. Say it/move it (Elkonin
Letter names task) segmentation, other
WJ-R Word Identification segmentation, letter-sound
Reading phonetically practice
regular words Group 2. No intervention
Reading phonetically
regular nonwords
Developmental spelling


Results: Group 1 significantly outperformed the control group on phoneme segmentation, letter names, sound
knowledge, word and nonword reading, and developmental spelling.


Vellutino N = 300 Pseudoword decoding Group 1. Phonemic segmentation, 30 minute
& Scanlon Grade 2 (average test letter-sound practice sessions
(1987) learners) and Phonemic segmentation Group 2. Response association 2 weeks
Grade 6 (those with ability test Group 3. Both treatments combined
Specific Learning Response acquisition Group 4. No intervention--Grade 2
Disabilities) measures (recall) Group 5. No intervention--Grade 6
Word identification
(code acquisition)

Result: Groups 1 and 2 made significantly higher gains than Groups 2 and 4 on all measures.


15-20 minute
sessions
4 sessions
per week
11 weeks










instruction to a total of more than 300 kindergarten

children. Each intervention took place 4 times a week over a

period of 11 weeks. These children were compared with those

who received no specialized training in phonological

awareness. The researchers used a "say-it-and-move-it"

segmentation task, modeled after Elkonin's (1973) similar

activity. In this segmentation activity, children moved

colored disks from the top of a laminated card to the bottom

half to represent the sounds they heard in words. Letters

were put on the tiles of those children who had mastered both

the name and sound of the letter. Children could use a

combination of letter tiles and blank tiles, or they could

use all blank tiles to segment a word. Other segmentation-

related activities included categorizing groups of pictured

words by rhyme or alliteration. Cards were placed on a

table, and the children pointed out the one that did not

rhyme with the others or did not begin with the same initial

sound. Results of the Blachman et al. (1994) study indicated

significant positive effects for the segmentation group on a

test of segmentation and positive effects for reading on the

Word Identification subtest of the Woodcock-Johnson

Achievement Tests-Revised (Woodcock & Johnson, 1989, 1990).

Results of the Tangel and Blachman (1992) study indicated

significant positive effects for the segmentation group on a

test of segmentation, two measures of reading (phonetically

regular words and nonwords), and a test of developmental

spelling.










Though positive effects on phonological awareness and

reading were found in both of these studies, the Ball and

Blachman (1991) study has been criticized because the

blending training some of the children may have received in

their regular classroom instruction would have confounded the

effects of the phonological intervention (Davidson & Jenkins,

1994). If such were the case, the positive effects of the

intervention may have been the result of blending-plus-

segmentation training rather than segmentation alone. Such a

limitation was avoided by Byrne and Fielding-Barnsley (1991),

who taught phoneme identification to 126 kindergarten

children for 12 sessions. Reading was measured by matching

an auditory stimulus to a printed word ("Does this say sat or

mat?"). In such a task, blending skill would not be a factor

because only the correct identification of the initial

phoneme would be required. Significant overall increases

were reported in phoneme identification and reading.

Other interventions have consisted of teaching more than

a single phonological task combined with letter-sound

training (Alexander, Anderson, Heilman, Voeller, & Torgesen,

1991; Ball & Blachman, 1991; Castle, Riach, & Nicholson,

1994; Haskell, Foorman, & Swank, 1992; Hohn & Ehri, 1983;

O'Connor & Jenkins, 1995; O'Connor, Jenkins, & Slocum, 1995;

Olafsson & Lundberg, 1983). These studies included

segmentation and blending instruction, sometimes combined

with instruction in other phonological skills. For example,

O'Connor et al. (1995) sought to determine the breadth of








68

phonological awareness instruction necessary to increase the

phonological skills of kindergartners at risk for reading

failure. The study included a low phonological awareness

group (language delayed, special education, or special

education referred students) and a high phonological

awareness group (average learners). The 67 children in the

low phonological awareness group were divided into three

treatment groups. One group received blending and

segmentation plus letter-sound instruction. The second group

received the same instruction with the addition of tasks such

as syllable deletion and rhyming. The third group received

letter-sound training only. Results indicated no significant

differences for the group taught phonological skills only and

the group taught phonological awareness skills plus letter-

sound correspondence on the Lindamood Auditory

Conceptualization (Lindamood & Lindamood, 1979) test and a

reading analog test. Both performed significantly better

than the letter-sound-only group. The authors concluded that

learning segmentation and blending provide enough base to

transfer to broader phonological awareness. Teaching

additional phonological awareness skills may not be

necessary.

A major weakness of most of the studies combining

phonological training and letter-sound practice is that none

included a group that received phonological training alone.

No conclusion about the relative effectiveness of

phonological training with and without the letter-sound










component can be drawn from this research. A well-designed

study that did include a phonological-skill-only group (Hohn

& Ehri, 1983) was conducted with 24 kindergarten children.

One group used letter tokens to practice segmenting words. A

second group practiced the same skill with blank tiles. The

study resulted in significant effects for the letter token

group for segmentation skill but no significant differences

in the two groups for decoding nonwords. The authors

concluded that using letters facilitates segmentation

instruction but does not enable kindergartners to decode

nonwords (blend). The Defior et al. (1994) study also

compared groups taught phonological awareness with and

without letter-sound instruction. The study took place in

Spain with 55 age 6 subjects. Significant effects for the

letter-sound group on all measures were noted.

Summary of studies combining phonoloaical and letter-

sound training. Overall, the results of studies that have

used the phonological plus letter-sound combination training

have been positive. Significant increases in phonological

awareness skills have been reported for children in these

treatment groups when compared with children engaged in other

language activities or those given no extra instruction.

Because few of these studies have included a phonological-

skills-only treatment group, it is difficult to draw

conclusions about the relative effectiveness of phonological

skills training with and without the addition of letter-sound

practice. A study that did make such a comparison (Hohn &










Ehri, 1983) found no significant differences between groups

on reading measures. A second study making the same

comparison (Defior et al., 1994) found significant effects

for the letter-sound group.

Contextual application of phonolocical skills. There is

evidence suggesting that phonological skills taught within

the context of reading-connected text is effective for some

children (Cunningham, 1990; Hatcher, Hulme, & Ellis, 1994;

Lane, 1994; Vandervelden & Siegel, 1995; Weiner, 1994).

Table 5 summarizes this research. In Cunningham's (1990)

study, children in kindergarten and first grade received

explicit instruction in the application of phonological

awareness skills and discussed how and when to use the skills

while reading stories. They were compared with a group

taught phonological awareness skills by the isolated skills

practice method. These children received no instruction

within the context of reading. Both groups displayed a

significant improvement in reading achievement; however, the

degree of improvement of the first-grade participants

depended strongly upon the type of instruction received. The

children who were helped to think about the value of

phonological awareness skills and their application to

reading performed better on a transfer measure of reading

achievement than those who received no contextual

application.

The relative effects of phonological awareness skills

training combined with reading instruction were compared with






Table 5

Studies Training Contextual Application of Phonological Skills


Study Subjects Dependent Measures Intervention Duration


Cunningham N = 84
(1990) 42 Kindergarten
Age 6
42 Grade 1
Age 7
(average learners)


Phoneme deletion
Phoneme oddity
LAC
Metropolitan
Achievement Test
(MAT)


Group 1. Metalevel segmentation
& blending with explicit
application (discussed how
and when to use skills in story)
Group 2. Skill and drill
segmentation and blending
(implicit application-no
application in reading context)
Group 3. No intervention


Results: Significantly greater positive effects for Group 1 when compared to Groups 2 and 3. For Group 1,
grade 1 subjects were significantly higher than kindergarten subjects on deletion and a transfer measure of
reading achievement.


15-20 minute
sessions
2 sessions
per week
10 weeks






Table 5--continued.


Study Subjects Dependent Measures Intervention Duration


N = 124
Age 7
(United Kingdom)
(below average
reading skills)


Sound deletion
Nonword segmentation
Sound blending
Sound categorizing
Word & text reading
Spelling


Group 1. Reading with letter
practice and PA activities
Group 2. Reading with letter-
practice--no PA activities
Group 3. PA activities alone
letter-sound practice or
reading application
Group 4. No intervention


Results: Those taught PA in the context of reading (Group 1) performed significantly better than all other
groups on reading and spelling measures. At a follow-up assessment 9 months later, reading gains were
maintained, but spelling gains tended to diminish.


Lane N = 43
(1994) Grade 1
(below average
phonological
awareness)


Elision
Sound Matching
Blending
Segmentation
Decodable words
and nonwords
Reading fluency
Reading comprehension


Group 1. Instruction in PA
skills with contextual
application
Group 2. Instruction in PA
skills without contextual
application


Results: When Group 1 and Group 2 were divided into levels according to reading fluency pretest scores,
significant differences were noted on measures of total phonological awareness and reading comprehension for
those children with very low fluency on the pretest.


Hatcher,
Hulme, &
Ellis
(1994)


sound

sound

-no


30 minute
sessions
40 sessions
25 weeks


30 minute
sessions
18 sessions
8 weeks






Table 5--continued.


Study Subjects Dependent Measures Intervention Duration


Vandervelden N = 30
& Siegel Kindergarten
(1995) Age 5-6
(22 with below
average phonological
awareness)


Phonological recoding
tasks
Speech-to-print
matching
Invented spelling
Nonword reading
PA and letter tasks
Phoneme awareness
(auditory
recognition)
Deletion
Substitution
Letter naming
Letter-sound
correspondence
Reading tasks
Word learning
Word reading
Finger-point reading
Key-word reading


Group 1. Segmentation activities
Spelling practice
Reading practice
Guidance in using letter-sound
& PA knowledge to read
and spell
Group 2. As above but no guidance
in application of PA or letter-
sound knowledge to read and
spell


Results: Group 1 gained significantly compared to Group 2 on all measures. Group 2 had significant gains
on only one measure (finger-point reading of text).


30 minute
sessions
2 sessions
per week
12 weeks






Table 5--continued.


Study Subjects Dependent Measures Intervention Duration


Weiner N = 75
(1994) Grade 1
(low achieving &
middle achieving)


Segmentation Group 1. Segmentation, blending,
Deletion substitution--no letter-
Substitution name or sound practice
Reading decoding Group 2. As above but followed
Gates-MacGinitie Reading by decoding practice (no
Test explicit training for con-
Oral reading assessment necting PA skills with
decoding activity)
Group 3. Conceptual (Group 2
intervention plus explicit
instruction on how to apply
skills in text reading)


Results: No significant differences among groups on measures of phonological awareness or reading. For the
lowest-ability subjects only, conceptual training (application in the context of reading as for Group 3) was
the least effective intervention.


25 minute
sessions
2 sessions
per week
6 weeks










the effects of reading instruction alone and phonological

awareness instruction alone (Hatcher et al., 1994). Groups

of children (N-124) matched on age, reading ability, and IQ.

Training of phonological awareness in the context of reading

produced strong positive effects on measures of spelling and

word reading. Nine months later, the effects on reading were

still in evidence, though beginning to show a diminishing

tendency. Vandervelden & Siegel's (1995) study with children

with low phonological awareness resulted in significant

progress for children who received contextual instruction for

one group but no practice in context for the second group.

Results indicated significant progress for the group taught

within the context of reading when compared to the group who

received no contextual instruction.

In Lane's (1994) study, data analysis revealed that for

the lower achieving children, those who practiced skills

within the story context made significant gains on two

measures, total phonological skills and reading

comprehension.compared to those taught with no contextual

application.

Weiner (1994) compared 23 low-achieving and 52 middle-

achieving first-grade students to determine their responses

to training that included contextual application of

phonological skills. In contrast to the findings of

Cunningham (1990) and Hatcher et al. (1994), Weiner's










research indicated that training providing children with a

conceptual connection between phonemic skills and reading was

generally ineffective for low readers.

Summary of the Content of Phonoloaical Awareness
Instruction

Phonological awareness training has been implemented

with beginning readers in an attempt to make them consciously

aware of the segmental nature of language and to help them

gain some facility in the manipulation of the sound

components that make up language. A review of the literature

reporting these studies suggests the most frequently included

tasks were (a) segmentation, (b) blending, and (c) rhyming.

Less frequently included tasks were (a) deletion, (b) phoneme

identification, and (c) phoneme substitution.

Phonological awareness training has consistently

increased phonological awareness skills and reading

performance. This result has held constant whether the

phonological tasks were taught alone or in combination with

letter-sound instruction. Though some studies (Ball &

Blachman, 1991; Hohn & Ehri, 1983) indicated the importance

of letter-sound instruction, such training does not seem to

be crucial to successful intervention. Clearly, phonological

instruction alone can be effective in increasing phonological

skills and reading skills measured by real or nonword reading

(Davidson & Jenkins, 1994; O'Connor et al., 1993; Slocum et

al., 1993; Torgesen et al., 1992). The children in the

present study demonstrated letter-sound knowledge on an










invented spelling task administered prior to the pretest

battery, and letter-sound instruction was not considered

necessary to the design of this intervention.

In this review of phonological training studies,

critical skills that must be taught to increase phonological

awareness and reading skill have been identified. There is

additional evidence suggesting the method of instructional

delivery that has been most effective in teaching these

skills. In the following section the characteristics of

direct instruction are discussed.

Direct Instruction and Phonological Awareness

A previous review of phonological training studies

(Smith, Simmons, & Kameenui, 1995) reported that conspicuous

strategies (sequences of teaching events and teacher actions

that make explicit the steps required to hear and manipulate

sounds) were present in all interventions. Smith et al.

suggested the following components of successful instruction

in phonological awareness skills:

1. Explicit instruction in which phonological skills

are directly and overtly taught and practiced. Instruction

in phonological skills must be systematic, structured, and

taught to mastery.

2. Demonstration, modeling, and guided practice. Tasks

should be modeled by the teacher and practiced over time to

make the blending and segmenting processes obvious and

automatic. Corrective feedback should be provided.










The instructional variables described by Smith et al.

(1995) are characteristic of direct, systematic instruction.

Results of the more exhaustive review undertaken for the

present study corroborate these findings. In all studies

where instruction in phonological awareness resulted in

significant increases in phonological and reading skills,

elements of direct, explicit intructi-n were present.

Furthermore, several experimental studies have compared the

effects of the direct instruction method with another method

of instruction on the phonological awareness skills of

children. In the following section, these studies are

reviewed.

Experimental Evidence for the Efficacy of Direct
Instruction

Additional evidence supporting the efficacy of the

direct instruction method in teaching phonological skills was

provided by several studies comparing other methods of

instruction with the direct instruction method (Castle et

al., 1994; Felton, 1993; Foorman et al., 1991). Castle et

al. (1994) compared children taught segmentation, deletion,

substitution, rhyming, and letter-sound correspondence with

those taught a process writing activity using invented

spelling. The writing activity included planning, editing,

and publishing creative compositions. The invented spelling

component involved using letters to represent the sounds

heard in words. Results indicated greater effects for the

direct instruction group on phonological awareness measures










and spelling measured by the WRAT (Jastak & Jastak, 1978).

In a second experiment, children were explicitly taught

segmentation, blending, alliteration, rhyme, deletion, and

letter-sound correspondence and compared to those taught by a

meaning-based approach (categorizing, semantic cues, letter

names, listening to stories). Results showed significantly

superior effects for the direct instruction group on measures

of phonological awareness, letter knowledge, word reading,

pseudowords, dictation accuracy, writing vocabulary, and

print concepts.

In a 2-year study, Felton (1993) sought to determine how

children with phonological processing deficits respond to

code versus context reading instruction. Children at risk

for reading failure (N=48) were randomly assigned to

treatment groups. In a second experiment children at risk

(N=65) were placed in treatment groups by matching for the

extent of their phonological processing deficits. Results of

both experiments indicated higher means for the code group

for all measures (including the Word Attack and Word

Recognition subtests from the Woodcock Reading Mastery Test-

Revised (Woodcock, 1987).

Training study reviews provide evidence that the direct

instruction approach has been successful in increasing

children's phonological awareness and reading skills.

Within-study comparisons of direct instruction in

phonological awareness to other instructional methods

corroborate these findings. Such evidence supports










interventions that provide explicit instruction in

phonological awareness to children deficient in these skills.

In the following section, the characteristics of direct

instruction are discussed.

Characteristics of Direct Instruction

Direct instruction variables include engaged time,

frequent responding, immediate feedback, and error correction

(Rosenshine, 1976). As practiced, direct instruction often

incorporates principles of systematic instruction, including

identifying goals and objectives, breaking down tasks into

simplified units, carefully sequencing units to accomplish

the learning goal, and providing systematic assessment to

evaluate instruction (Spiegel, 1992). In direct instruction,

students are focused on what is to be learned and explicitly

taught how to do a particular process through teacher

modeling and explanation. This instruction is followed by

guided practice as students are given assistance in

performing skills. Independent practice takes place with

relevant materials until mastery of the skill is achieved.

Throughout direct instruction, assessment provides data that

inform decision-making for instructional change. Direct

instruction is based on providing enough sustained, focused

practice to enable learners to use skills and strategies

effectively.

Direct instruction is conceived as a carefully sequenced

curriculum (Sexton, 1989). Teachers present information,

usually guided by a prepared script. Students repeat the








81

information. Teachers ask questions. Students respond. The

teacher praises appropriate responses or offers corrective

feedback for incorrect responses. The process is repeated

until the children can respond correctly. At that point,

instruction advances to the next sequence.

Summary of Direct Instruction and Phonolooical
Awareness

Training study evidence provides support for including

elements of direct instruction in interventions designed to

increase phonological awareness skills. The components of

direct instruction have been well defined in the professional

literature. Implementing such instruction may be difficult

for teachers in the general classroom, however. Challenges

teachers face in providing such instruction are discussed in

the following section.

Barriers to Delivering Direct Instruction in General
Classrooms

Though research supports the use of direct instruction

methods to teach phonological awareness, barriers exist to

providing such instruction in the general classroom. First,

conflicts in philosophy lead to disagreement about the way

instruction should be delivered in elementary classrooms.

Second, lack of teacher time, insufficient training, and

untested materials limit the ability of the teacher to

deliver direct instruction. Finally, the influence of school

culture affects teachers' planning of instructional










accommodations for individuals and subgroups at risk for

school failure.

Philosophical conflicts. Conflicts in philosophy

concerning reading instruction have centered around the

debate concerning whole language versus code emphasis

approaches to literacy. Given the generally acknowledged

fact that phonological awareness facilitates acquisition of

the alphabetic code and that phonological knowledge can be

assessed accurately, it appears logical to pursue

interventions to increase the phonological skills of students

at risk for reading failure. Much controversy exists,

however, over how these skills are to be taught to young

students. Though the emphasis in whole language classrooms

on the integration of speaking, listening, reading, and

writing has increased early literacy experiences, a

significant number of students do not learn to read simply by

being exposed to language. The missing element is direct

instruction in phonological knowledge and decoding skills

(Blachman, 1993; Mather & Roberts, 1994). While whole

language proponents may advocate phonics instruction on an

"as needed" basis, a systematic, ongoing assessment of these

skills is often absent from whole language classrooms.

Teachers may depend on their observational skills (Adams,

1990). Thus, it is improbable that every student gains

adequate instruction in the alphabetic code. Those students

already at risk because of lack of early literacy experiences

or learning disabilities may fail to learn to read.










Pressley and Rankin (1994) compared learning to read

with learning to play baseball. Students can learn by

playing the game, but honing individual skills is best

accomplished by direct instruction during practice sessions.

It is difficult to imagine learning any activity involving

discrete skills without the benefit of direct instruction.

Though possible for some learners, those without natural

aptitude will be at a disadvantage. This is precisely what

happens when we fail to provide instruction to students about

the segmented nature of spoken words and how sound segments

in spoken words map onto individual letters of the alphabet.

This is not to say whole language methods are bad or that

whole language classrooms should be abolished. On the

contrary, research supports a balanced approach which

combines both whole language and code-oriented instruction

(Vellutino, 1991).

In support of the balanced approach to instruction, a

continuum of teaching methods was proposed (Mercer, Lane,

Jordan, Allsopp, & Eisele, 1996). This model illustrated how

instruction can be conceptualized as a range of choices from

explicit, teacher-controlled learning to implicit, student-

regulated approaches. The authors suggested that methods of

instruction ranging from those based strictly on behavioral

principles to the discovery learning characteristic of whole-

language classrooms can be appropriate. Educators must

consider the needs of students and the nature of content

before making instructional choices for individual learners.








84

Implications of the balanced approach for instruction in

diverse classrooms include a rich variety of literacy

experiences combined with explicit instruction in letter-

sound correspondence. Within this context, instruction in

phonological skills should be provided for children who lack

sufficient phonological awareness to make sense of letter-

sound instruction. Direct instruction in both phonological

awareness and phonics must be provided on a consistent basis.

Ongoing assessment must ensure that skills are taught to the

mastery level.

Issues of time. training, and materials. Torgesen and

Barker (1995) discussed the difficulties teachers may

encounter in providing appropriate instruction in

phonological awareness for children at risk for reading

failure. Providing consistent direct instruction that is

well planned, interactive, and closely monitored requires

teacher time spent in planning and managing instruction for

children who may represent only a small percentage of the

total classroom group. Even when time can be found to

implement such instruction, children do not necessarily

master skills as a group. Teachers may not wish to hold back

a group until every child has mastered each skill. To do so

is limiting to children who learn at a faster rate.

Providing for the needs of children who learn at slower rates

can be difficult in diverse classrooms. Furthermore,

paraprofessionals, volunteers, and other personnel who might










be available to provide help to teachers may not have

received training in this kind of intervention.

There are other problems in providing direct instruction

on a consistent basis. Lessons must encourage frequent

responding and provide adequate amounts of feedback for each

child. Motivational factors play an integral role in

children's learning and must be considered. Direct

instruction in basic skills can seem repetitive and boring

unless lessons are engaging, interactive, and varied. Such

lessons require planning time for gathering and preparing

appropriate materials as well as consistent enthusiasm on the

part of the teacher. An effective behavior management system

must often be in place with adequate planning devoted to

providing meaningful activities for the students in the

classroom who do not need instruction in phonological

awareness.

An additional factor is the necessity for effective data

management. The effectiveness of instruction in phonological

awareness must be observed ongoingly; data must be recorded;

and progress must be assessed. From this assessment,

individualized instruction that matches the learning needs of

each child can be implemented. A system of data management

requires time to design, implement, and maintain. Yet, it is

essential to track children's progress in order to provide

effective and efficient instruction.

Such factors are principles of effective instruction

that most teachers are aware of and try to include in their










classroom instruction. However, implementing such

instruction for a small group of children within the regular

classroom requires an additional expenditure of time and

materials that the already overcommited teacher may not be

willing or able to invest.

The influence of school culture on teacher planning. In

a special issue of Effective School Practices, the Joint

Committee on Teacher Planning for Students with Disabilities

(1995) summarized 4 years of research across four projects

sponsored by the Office of Special Education and

Rehabilitative Services under the initiative of "Research on

General Education Teacher Planning and Adaptation for

Students with Handicaps." The results of this research on

teacher planning in America's schools explain some of the

problems associated with meeting the needs of students with

disabilities in general education classrooms. These problems

are applicable to planning instruction for students with

deficient phonological awareness skills. The following

statements from the Joint Committee report describe how

teachers plan for student learning in diverse classrooms.

1. Planning is focused on group learning.

Instructional approaches most often center around whole-class

activities.

2. Individual accommodation frequently conflicts with

the goals of group instruction. Teachers do not have time to

plan or implement accommodations for individuals or subgroups

in the classroom.










3. Teachers often plan for the group by planning for

the "one." Teachers plan for a "meta-student" who represents

the characteristics of many students encountered across their

careers. This hypothetical student is often an above-average

performer.

4. Teacher planning is characterized by efficiency and

economy. Plans focus on what must be accomplished rather

than the more difficult question of learning failure.

5. Teachers are not sure how to plan for students with

disabilities. Teachers believe they do not have the necessary

competencies to plan for and teach students with

disabilities.

6. Activities are chosen for their attractiveness to

students. Methods adopted to meet the needs of diverse

learners are frequently abandoned when they are perceived as

nonmotivating to higher-achieving students, even when such

methods have been effective in enhancing the learning of

these students.

7. Planning centers on how students will spend their

time. Planning books contain references to activities to be

conducted, pages to be covered, and assignments to be given.

References to accommodations for individual students are

rare.

8. Most teachers plan units rather than lessons.

Teachers talk most about unit goals. Less attention is given

to individual lesson goals. Lesson plans are rarely altered

to accommodate individuals or small groups.








88

Summary of barriers to delivering direct instruction in

general classrooms. Barriers exist to providing direct,

explicit instruction in general classrooms. Conflicts in

philosophy lead to disagreement about the way instruction

should be delivered. Lack of teacher time, insufficient

training, and untested materials limit the ability of the

teacher to deliver such instruction. The influence of school

culture affects teachers' planning and instruction of

individuals and subgroups at risk for school failure.

Fortunately, there is potential for computer-assisted

instruction to help teachers overcome some of the

difficulties in delivering direct instruction to children at

risk for reading failure because of delayed development of

phonological skills.

The Potential for Computer Assistance to Overcome
Barriers to Direct Instruction

Though problems exist in providing direct, explicit

instruction in general classrooms, computer-assisted

instruction can help teachers overcome these barriers.

Computers can provide explicit instruction in whole language

classrooms for children who have difficulty learning skills

inductively. Computer assistance offers alternatives when

lack of time, training, and appropriate materials present

difficulties in delivering instruction. Computer instruction

can also help ameliorate the influence of school culture that

affects teachers' planning and instruction for individuals

and subgroups at risk for school failure. The following










sections highlight the characteristics of computer-assisted

instruction that can help teachers overcome barriers to

delivering direct instruction in diverse classrooms.

The Match Between Direct Instruction Methods and the
Characteristics of Effective Computer Proarams

The characteristics of effective computer programs

interface well with principles of direct instruction. In

direct instruction, teachers use modeling, explanation, and

guided and independent practice to make explicit to students

how a particular process can be performed. Direct

instruction incorporates ongoing assessment and instructional

change based on assessment data. Instructional variables

include engaged time, frequent responding, and immediate

feedback and error correction (Rosenshine, 1976).

Characteristics of effective software programs are (a)

sufficient degree of learner control, (b) adequate and

appropriate feedback, (c) on-going data management, (d)

effective use of graphics, and (e) integrated motivation

enhancement (Larsen, 1995). These characteristics interface

with the characteristics of direct, explicit instruction in

the following manner.

Sufficient degree of learner control. The ability of

the student to interact with the computer ensures "engaged

time" and frequent responding, two important direct

instruction variables. In an interactive program, each

stroke of the keyboard or click of the mouse will elicit as

few events as possible in order to keep the student










frequently and actively engaged (Geoffrion & Geoffrion,

1983). Ideally, students also will have some control of the

"directive" force. For example, the DaisyOuest (1993) and

Daisy's Castle (1993) computer programs require frequent

clicks of the mouse to choose areas and rooms of the castle

where different skills will be practiced. If students are

unresponsive to cues for their input, the wizard/instructor

begins to snore audibly. The student is thus prompted to

attend to the task. As often as needed, students can click

on stimulus pictures for pronunciation by the computer. In

the tutorial and practice format, student responses elicit

positive reinforcement or corrective feedback. In effective

programs, control is balanced between the program and the

learner.

Adequate and appropriate feedback. Immediate corrective

feedback is an important component of direct instruction and

an integral aspect of computer learning (Balajthy, 1989).

Larsen (1995) asserted that feedback must consist of more

than whether responses are correct or incorrect. In order to

include a directional dimension of learner control, feedback

should consist of corrective cues and information about why a

response is wrong and how it can be changed. Larsen

suggested that one of the greatest advantages of using a

computer program is its ability to provide consistent and

immediate directional and corrective feedback.

On-coina data management. A salient aspect of effective

direct instruction is frequent assessment of progress and








91

changes in instruction based on such data. Computer programs

can store performance or evaluation data necessary for the

documentation of student progress (Balajthy, 1989;

Hasselbring, 1991). Longitudinal student-performance data

allow teachers to make informed decisions based on an

individual student's progress. Programs based on the concept

of mastery learning can provide instantaneous feedback and

identify and reward correct responses while also maintaining

complete records of pupil performance. Such information

informs teachers who wish to select appropriate teaching

sequences for individuals (Barker & Torgesen, 1995; Spencer,

1996).

Effective use of graphics. Direct instruction for

prereaders must necessarily rely on illustrations to

reinforce auditory learning. These elements can be an

important factor in delivering instruction that is engaging

and motivating. Computer graphics presented as an integral

part of the learning content also reinforce learning and

enhance retention.

Integrated motivational factors. In direct instruction,

positive feedback for correct responding ensures built-in

motivation for students. Computer-assisted instruction is

characterized by three kinds of motivational factors: (a)

unintegrated extrinsic rewards, (b) integrated extrinsic

rewards, and (c) intrinsic interest (Larsen, 1995). Examples

of unintegrated extrinsic rewards are smiling faces or points

given to the student or appearing on the screen when correct




Full Text
xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID EGR1WBVHS_G9P1N4 INGEST_TIME 2014-04-14T18:59:04Z PACKAGE AA00014245_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES