The Effects Of
Computer-Based Instruction On The Phonemic Awareness Development Of Children
With Specific Learning Disabilities
Éamonn Murtagh
ABSTRACT
This study focuses on the phonemic awareness
development of children with specific learning disabilities, (SLD) sometimes
referred to as dyslexia. A sample of children with SLD (N = 138) was assigned
to two treatment conditions. One group was taught phonemic awareness using
traditional methods while a computer-based approach to phonemic awareness
development was used with the other group. Data about the children’s ages,
reading ability and phonemic awareness development were collected at the
pre-treatment stage and an identical set of data was collected after the two
groups had been in the respective treatment conditions for a six-month period.
Phonemic awareness and reading achievement data were also collected from a
group of pupils in first class (N = 109), primarily for the purpose of validating
the instrument used to test the phonemic awareness development of the SLD
pupils, but the first class data also provided a useful benchmark of the
phonemic awareness development of the children in the SLD group at the
pre-treatment stage.
It was found that the first class group
(mean age = 7.6 yrs) achieved a significantly higher phonemic awareness
development score than the SLD group (mean age = 10.0 yrs.) at the
pre-treatment stage, largely because the first class children were better able
to pronounce nonsense words. The SLD group (mean age =10.0yrs.) was estimated to have a mean deficit of
approximately 2.5 years in phonemic discrimination and a significantly greater
deficit in phonemic blending. More than half of the SLD group achieved a reading
percentile of 5 or less.
Following a comparison of the pre-treatment
data with the post treatment data for the two treatment groups, it was evident
that the mean reading score and the mean phonemic awareness development score
of both groups had increased significantly during the treatment period. There
was a small but significantly greater increase in the reading achievement
scores of children with very low scores in reading when the computer-based
approach was used to teach them phonemic awareness, rather than traditional
methods. These results and other conclusions from the study are discussed.
This
work is dedicated to the many children who have a specific learning difficulty
and to the teachers who strive to find the key to unlocking the children’s
potential.
1.2 The importance
of phonemic awareness
1.3 Children with
a specific learning disability
1.5 Using
Computers to Develop Phonemic Awareness
2.3 Early research
in phonemic awareness development
2.4 Levels of
phonological awareness development
2.5 The nature of
phonemic awareness
2.6 Assessing
phonological awareness
2.7 Phonemic
awareness and early reading acquisition
2.8 Teaching
phonemic awareness development
2.9 Phonemic
awareness development of children with SLD
2.11 Computers and
literacy instruction
2.12 Computers and
phonemic awareness development
3.1 Introduction:
Context of the Study
3.2
Characteristics of the subjects of the study
3.4 Assessment of
pupils prior to enrolment
3.7 Administration
of the phonemic awareness test
3.8 Validation of
the phonemic awareness test
3.10 Profile of
the phonemic awareness development test
3.11 Summary of
the result of the analysis of the data for first class pupils
3.12 Assessment of
children with specific learning disabilities at enrolment
3.13 Assignment of
pupils to treatments groups, OM and US. 49
3.14 ‘Units of
Sound’ treatment
3.15 ‘Other
Methods’ treatment
3.16 Data
collection and analysis
4.4 Analysis using
continuum model
4.5 Analysis of
the phonemic awareness development of SLD pupils
4.6 Importance of
phonemic awareness in reading development
Chapter 5: DISCUSSION OF
RESULTS
5.2 Results of the
tests of first class children
5.3 Results of
data analysis of the SLD group
Appendix A:
Phonemic Awareness Development Test
Appendix B: Letter
to Remedial Teachers
Appendix C:
Phonemic Awareness Test: Record Sheet 103
Appendix D: Data
for first class and SLD pupils
Chapter 1: INTRODUCTION
1.1 Background
The debate about methods of teaching
reading has seldom been out of the educational limelight since the publication
of "Why Johnny Can't Read" (Flesch, 1955). Chall (1967) succeeded to
some extent in casting oil on trouble waters with her seminal publication in
which she concluded that, for most children, the best approach to teaching
reading was to be employ a combination of methods or the so-called 'eclectic'
approach. The 'reading wars' broke out again in the 1980's between the
proponents of a ‘whole-language’ approach using ‘real books’ and the advocates
of traditional approaches. Many of the issues involved continue to be hotly
debated in academic journals and other publications and the controversy has
been an important element in the design of curricula in the UK and in the focus
of contemporary research on the teaching or reading. Rays of clarity have
recently begun to penetrate the smog of what has, at times, been a destructive and rancorous debate. The importance
of phonemic awareness for success in learning to read is one of the areas of
significant interest that has been brought to light in more recent times.
1.2 The importance of phonemic awareness
Phonemic awareness is critical for the
development of reading skills in an alphabetic language, such as English. It is
defined as “the awareness that phonemes exist as abstractable and manipulable
components of spoken language. It is the ability to reflect on speech and
experiment (play) with its smallest components (phonemes). Phonemic awareness
is not phonics and not auditory discrimination.” (Yopp, 1992).
A number of researchers have concluded that
phonemic awareness development is the single most effective predictor of young
children's success or failure in learning to read. Measures of phonemic
awareness can more accurately forecast success or failure in reading than
measures of intelligence, for example. Some of the more significant findings of
this research will be reviewed in later chapters. It has also been revealed that lack of phonemic awareness is a
common variable among the seemingly disparate groups of children who fail to
learn to read. Children with specific
learning disabilities (SLD) represent a small subset of poor readers that has
puzzled researchers in the medical, educational and psychological disciplines
for over a century.
1.3 Children with a specific learning disability
Children with a specific learning disability (SLD) are defined in the report of the Special Education Review Committee (Ireland, 1993) as having
“… impairments in specific areas such as reading, writing, spelling and arithmetical notation, the primary cause of which is not directly attributable to assessed ability being below the average range, to defective sight or hearing, emotional factors, a physical condition or to any extrinsic adverse circumstances.” (pg. 86)
The Review Committee recommended that an
assessment by a psychologist on a standardised intelligence test should
establish that the pupil’s intelligence is at least average and that their
performance in basic literacy skills in relation to objective criteria should
be at a very low level. The Review Committee did not define precisely what was
meant by a ‘very low level.’
One of the defining characteristics of
children with SLD is very poor reading ability. Children with SLD, in common
with poor readers generally, have been shown to display difficulty with
phonologically regular nonsense words when matched with controls of an
equivalent literacy level. (Pumfrey, 1991). Evidence of their limited grasp of
phonemic awareness is frequently revealed in their poor, often bizarre spelling
which remains a lifelong characteristic of people who have specific learning
difficulties.
Children who are assessed as having average
intelligence and whose reading ability, as measured by a standardised test, is
at or below the second percentile are automatically eligible to be enrolled in
special schools or support units for children with specific learning
disabilities at ordinary schools (Department of Education and Science, 1999).
Development of children’s phonemic awareness is one of the main strategies
employed by all of the schools where there is special provision for children
with SLD to help them to overcome their literacy difficulties.
1.4 The Research
Since research has established that
phonemic awareness is critical to learning to read, it is of interest to
establish how the phonemic awareness development of children with SLD differs
from that of children whose reading ability is in the average range. It is also
of significance to establish whether there are aspects of phonemic awareness
that are a source of particular difficulty for children with SLD, or whether
their difficulties in manipulating speech sounds are more global.
Are there approaches to developing phonemic
awareness that work better for SLD pupils than the approaches that are used
with children generally? Can the results of different approaches to developing
phonemic awareness over a given period of time be quantified? If, as recent
research appears to suggest, there is a causal relationship between phonemic
awareness development and learning to read (Bradley & Bryant, 1983), does
teaching phonemic awareness to children with SLD result in improved reading?
If the latter proposition proved to be
true, then it might also indicate
strategies for helping SLD pupils overcome their learning difficulties. In order to attempt to answer these
questions, it would be necessary to create a profile of the phonemic awareness
development of children with SLD and to compare it with that of children at the
early stages of learning to read. To compare the phonemic awareness of the two
groups, it would be necessary to devise a test that would assess the phonemic
awareness development of both SLD pupils and pupils at the early stages of
learning to read. This approach could reveal a continuum of difficulty in the
phonological tasks presented to both groups. It would also reveal the
particular tasks that presented the greatest difficulty for SLD pupils.
1.5 Using Computers to Develop Phonemic Awareness
Pumfrey (1991) reflects on the possibility
that the technological revolution might have on teaching and learning. In
considering the educational potential of the microcomputer, Pumfrey
surmises as follows:
“Children's learning can be influenced by IT applications. The microcomputer is to the educational practitioner what the telescope and microscope were to the astronomer and the biologist, with at least one important difference: the possibility for developing theory and practice that it opens up are even greater.” (pg. 2)
Traditionally, workbooks, games, tapes and
specialised materials have been used to support the teaching of phonemic
awareness to children with SLD. Computers have been used to a limited extent,
but not in any systematic way. The British Dyslexia Institute published a
software program in 1998 entitled ‘Units of Sound.’ The main function of the
software was to teach phonemic awareness to children who have SLD and it was primarily intended for use
with children in the 9 - 12 age group.
The present research project sets out to examine the impact of the classroom
application of 'Units of Sound' software on the phonemic awareness development
of children with SLD, both in special schools and in support units attached to
ordinary schools.
1.6 Research Design
The complementary hypotheses for this
research are that pupils with SLD develop phonemic awareness differently to
children generally and that information and communications technology can
provide a solution for such pupils. The primary objective of the research is to
test these hypotheses through a combined correlational and experimental
research design and to establish the extent to which the children with SLD
assigned to different experimental conditions make progress in reading and in
the subsidiary skills of reading and writing,
phonological awareness and spelling.
In Ireland, pupils who meet the Department
of Education and Science criteria for SLD may be enrolled in special provision
for SLD for a two-year period. A new cohort of 140 pupils was enrolled in
special schools and special units in ordinary schools in the Dublin region in
September 1999. The phonemic awareness profile of this group prior to enrolment
and the subsequent phonemic awareness development of selected pupils under
specified experimental conditions are the main areas of investigation of this
study.
1.7 Aims of the research
The aims of the research are twofold. The first is to measure the phonemic awareness of children with SLD and to compare their phonemic awareness development with that of a normal cohort of children who are learning to read. This approach can provide an insight into the phonemic awareness profile of good, average and poor readers at a particular point in learning to read and it may also facilitate a comparison of the younger cohort of children with children who have SLD and are considerably older. The research is designed so that different dimensions of phonemic awareness can be measured, as well as the ability of the children to apply their knowledge of phonemic awareness to carry out decoding tasks similar to those they would use if they encountered an unfamiliar word while reading.
The second aim is to examine the extent to which the phonemic awareness of children with SLD develops under specified treatment conditions. One of the treatment conditions is the traditional approach that is employed in the special schools where teachers instruct children orally and use games and written activities to promote phonemic awareness. The second treatment condition involves providing approximately half of the newly-enrolled children with daily access to a computer-based software package called ‘Units of Sound’ to complement the other strategies that are used to develop their phonemic awareness.
Chapter 2: Literature Review
2.1 Introduction
The present study involves a comparison of two groups of children with specific learning difficulties, where one group was taught phonemic awareness using computers running a specified software program and another group was taught using other methods such as direct instruction, games and reinforcement activities. In preparation for the study, studies of phonemic awareness were reviewed, research about the phonemic awareness development of children with specific learning difficulties was investigated and literature about the use of computers in the teaching of reading was examined.
2.2 Definition
Terms such as ‘phonological awareness,’ ‘phonemic awareness’ or even ‘phonics’ are frequently confused. Phonological awareness can be defined as conscious knowledge that words are divisible into smaller units of sound According to Reynolds (1999) phonological awareness manifests itself in the ability to count, isolate, remove, recombine and otherwise manipulate phonological units. These units make up a hierarchy that includes sentences, words, syllables, onset- rimes and phonemes.
Phonemic awareness is the ability
to “recognize that a spoken word consists of a sequence of individual sounds”
(Ball and Blachman, 1991) or to
“manipulate individual sounds in the speech stream” (Yopp 1988). A phoneme is a
member of the set of the smallest units of speech that serve to distinguish one
utterance from another in a language or dialect. A unit of speech is considered
a phoneme if replacing it in a word results in a change of meaning. Phonemic
awareness, therefore can be described as the conscious understanding that
spoken language is composed of a set of discrete sounds and that manipulation
of phonemes can change the meaning of a word. Some phonemic sounds correspond
to individual letters (e.g. /a/ in apple), and other phonemes are made up of
combinations of letter (e.g. /ch/ in chin, /ea/ in eat).
2.3 Early research in phonemic awareness development
Interest in the development of phonemic
awareness is relatively recent. In 1971, the then Department of Education
placed little importance on children's phonological development, judging by the
guidelines to teachers that accompanied the ‘new’ curriculum (Ireland, 1971).
The curriculum guidelines informed teachers that ‘phonic techniques require the
teaching of the sound equivalent of the 26 letters (of the alphabet).’ The
guidelines mistakenly informed teachers that ‘there are about 60 sounds in
all’ - there are, in fact, only 44
discrete sounds in the English language. The prevailing attitude to teaching
phonics at the time is revealed in the cautionary sentence ‘since phonic
methods are slow and often uninteresting, they may be discouraging and tiresome
if introduced too soon….’ The guidelines guardedly proceed to point out that
phonics make pupils self-reliant in reading and less dependent on context or
other aids and are a valuable aid to spelling.
During the 1950s and 1960s research was
largely concerned with visual processes in learning to read (Williams, 1984).
Bruce (1964) published a phoneme deletion test and used it to demonstrate that
children younger than seven were unable to perform the tasks required by the
test. However, he did not correlate the test with reading, nor is it clear
whether the children had any previous training in phonemic tasks.
The work of Jean Chall (1967) remains of
seminal importance in any discussion about relationships between letters and
sound. The ‘great debate’ about which Chall wrote was sparked by Flesch’s
(1955) assertion that the ‘look-and-say’ approach to teaching reading was
causing large scale reading failure among children in the United States.
Over a three-year period, Chall undertook a
remarkably comprehensive review of the texts used in the teaching of reading, she
visited the classrooms of over 300 children in the United States, England and
Scotland and she reviewed all of the available research on the effectiveness of
‘look and say’ versus phonic approaches to the teaching of reading. From her
review of the research, she concluded that children taught through the look-say
method demonstrated an early advantage in fluency, comprehension and
silent-reading ability. On the other hand, children taught through phonic
approaches had an early advantage in independent decoding of unknown words
which they maintained over the comparison group. Moreover, there was evidence
to suggest that the children taught through phonic methods surpassed the
look-say children in reading vocabulary, reading comprehension and silent reading
by the end of second grade.
Summarising Chall’s work in one of the
other highly acclaimed works of twentieth century literature on early reading
development, Adams (1990) writes as
follow:
“The observations and data she amassed seemed inescapably to suggest that – as a complement to connected and meaningful reading – systematic phonic instruction is a valuable component of beginning reading instruction. Its positive effects appeared both strong and extensive. Yet the reader is left with the impression that these findings took Chall by surprise. Tucked away amidst the scientific discipline of her writing, she alternatively apologises, rationalises and philosophises: how could we not have known?” [pg. 39]
In the 1970’s, inspired by Chall’s work and
“faced with disillusionment about the benefits of visual training, the research
field turned its interest to the auditory modality” (Williams, 1984). Calfee, Lindamood and Lindamood (1973) gave
a test of auditory conceptualisation to 660 children aged between 5 years and
18 years. They also administered the reading and spelling subtests to the
same children. The average correlation
of the reading test scores of the 660 children with the scores achieved on the
auditory conceptualisation test was found to be .73. At about the same time,
Isabelle Liberman and Donald Shankweiler (1973) were analysing phonological
processing in an attempt to discover how it contributed to reading. They tested
children’s awareness of rhyme, their ability to segment words into separate
phonemes and to remember the order of sounds in words. Commenting on the
Liberman and Shankweiler’s work, McGuinness (1997) remarked that “their results
revealed that children do not spontaneously learn to segment words into their
constituent parts just because they are exposed to an alphabetic writing
system.” It was found that many
children were not aware even of words as independent units of speech. On
the other hand, it was found that children could segment syllables more easily
than phonemes. Significantly, it was also found that the ability to segment
phonemes was the best predictor of subsequent reading skills.
More recently, however, research has
revealed that phonemic awareness is a much more important skill in learning to
read than was previously acknowledged.
One of the most consistent findings to emerge from the studies based on
the teaching of reading in the past decade is the relationship between phonemic
awareness, reading acquisition and later reading success (Adams, 1990; Brady
& Shankweiler, 1991; and Torgesen, 1993). The findings strongly suggest
that children who lack phonemic awareness skills as a precursor to learning to
read are at risk of developing reading disabilities and need explicit
instruction in phonemic awareness if they are to become skilled readers and
spellers (Alexander, Anderson, Heilman, Voeller, & Torgesen, 1991).
Stanovich (1993) wrote
"In the last 10 years, researchers have come to a strong consensus about the cognitive processes that best predict reading progress. These cognitive processes have been called phonological awareness and are measured by tasks such as [phonemic deletion, blending, segmentation and counting]." (pg. 283)
Once beginning readers have some awareness of phonemes and their corresponding graphic
representations, research has indicated that further reading instruction
heightens their awareness of language, thereby assisting in developing the
later stages of phonological awareness mentioned above. Phonemic awareness,
therefore, is both a prerequisite for and a consequence of learning to read
(Yopp, 1992).
2.4 Levels of phonological awareness development
Phonological awareness development begins
at a very early stage in a child’s life and is an integral part of the language
acquisition process for children whose hearing is within the normal limits.
Snow and Burns (1998) summarise the sequence from speech perception to
phonological awareness as follows:
Practically from birth, infants are able to distinguish all the sounds of any human language, and within a short time their perceptual abilities become tuned to their native language, even though their productive repertoire remains limited to nonspeech sounds and babbling for much of the first year of life (e.g., Werker and Lalonde, 1988). Phonological development continues well beyond the first year and probably continues to be refined even in the early school years (e.g., Nittrouer, 1992; Gerken et al., 1994; Fowler, 1991). [pg. 47]
Studies reviewed in Gerken et al., (1994)
argue that children’s perception of speech undergoes a shift from holistic
awareness of words to perception of
smaller speech units, such as syllables and phonemes during the late
preschool period.
Children acquire the concept that one object or event may stand for another at a very early age (Marzolf and DeLoache, 1994).Young children playing with dolls or cars are symbolically representing their experience. In the field of literacy development, learning that the alphabet is a symbol system for sounds fits into this continuum of development. The ability to use symbols is gradually acquired during the first years of life as children progress from interpreting first pictorial, then iconic and finally graphic representations. Very young children, will recognise, for example, pictures of Santa Claus or Ronald McDonald. They may also recognise that golden-arched letter M stands for MacDonald's. But the fact that most 3-year-olds are able to recognise symbols in one context does not mean that they can apply this ability across all contexts without explicit teaching and specific practice.
The term ‘phonological awareness’ is often used as a global description of understanding that spoken language can be subdivided into words, and that words can be further divided into syllables. The latest acquired stage in phonological awareness is phonemic awareness (Yopp, 1992). A more recent elaboration of this notion, posited initially by Treiman (1985), is that there is more than one level of phonological awareness. Treiman’s research challenged the traditional view that spoken words could be segmented only into either syllables or phonemes. For example, a word such as ‘carpet’ can be divided into two syllables, ‘car-‘ and ‘-pet’ and the word ‘mat’ can be divided into three phonemic units /m/ /a/ /t/. Treiman proposed that there is another level between the syllable and the phoneme level where words such as ‘mat’ are divided into two units. The first unit, called the onset corresponds to the consonant cluster before the initial vowel sound and the second unit called the rime corresponds to the vowel-consonant cluster at the end of the word. Rimes can consist of just vowels, such as the word ‘see’ where /s/ is the onset and /-ee/ rime. Some words such as the word ‘at’ have a rime, but no onset.
Goswami
(1995) citing research by Bradley and Bryant (1978 & 1983) stated that
research has shown that onsets and rimes are ‘psychologically distinguishable
units for children long before they begin to learn to read.’ Bradley and
Bryant’s research had found that 4-year-olds could succeed on an odd-one-out
test of rhyming and alliteration. For
example, children were asked to say which word from a group of three has a
different first sound (hill, pig, pin) or which has a different rhyme
sound (cot, pot, hat). They also found that children’s ability as
preschoolers to categorise words on the basis of onset and rime was a strong
predictor of their reading and spelling ability four years later.
Thus, a clear progression of phonological
development was identified. Children’s awareness of onset and rime develops
before their awareness of phonemes. Goswami (1995) concludes that if children come to school with a good
phonological foundation already present, they can be taught to read immediately
via analogies and word families. For children with poor phonological skills,
extra work with rhyme and alliteration would be necessary so that they would
develop the phonological skills necessary to work with rhyme and alliteration. However, the question of the relationship
between onset-rime and reading is, as yet, unclear - see McGuinness (1998)
below.
2.5 The nature of phonemic awareness
Research is still grappling with the
fundamental question as to whether phonemic awareness is a unitary construct or
whether it is a combination of different constructs. The consistently wide
range of scores that children have
achieved on subtests of phonemic awareness appears to suggest that, either
there is more than one fundamental skill involved, or there is a continuum of
ability that may evolve over time, in tandem with oral language and perhaps,
reading development. In their position statement on phonemic awareness, the
Board of Directors of the International Reading Association (International
Reading Association, 2000) reflect the continuing uncertainty about this
question as follows:
Researchers are still working to sort out the kinds of tasks and what aspects of phonemic awareness they might tap. It appears from the research that the acquisition of phonemic awareness occurs over time and develops gradually into more and more sophisticated levels of control. Some research suggests that there is a diversity of developmental paths among children. How much control is necessary for the child to discover the alphabetic principle is still unclear. (www.readingonline.org )
There is general agreement among authorities on phonemic awareness that it can be difficult to learn (Adams, 1990). Reviewing a wide body of research Adams concludes that phonemic awareness eludes 25 per cent of middle-class American first graders and substantially more of those who come from a less linguistically-rich background. The reasons are complex. Adams summarised it as follows in Adams et al. (1998)
The problem, in large measure, is that people do not attend to the sounds of phonemes as they produce or listen to speech. Instead, they process the phonemes automatically, directing their active attention to the meaning and force of the utterance as a whole. The challenge, therefore, is to find ways to get children to notice the phonemes, to discover their existence and separability. Fortunately, many of the activities involving rhyme, rhythm, listening, and sounds that have long been enjoyed with preschool-age children are ideally suited for this purpose. (Chapter 1)
The processes
involved in phonemic awareness appear to centre on identity (also referred to
variously as ‘awareness,’ or ‘discrimination’) (Adams, 1990) and manipulation (also
referred to ‘isolation’ and ‘segmentation’ and includes ‘blending’ or ‘coding’)
(Yopp, 1988). Identity is quite problematic, because in speech, phonemes are
not acoustically discrete (Liberman, Cooper, Shankweiler &
Studdert-Kennedy, 1967). When a person’s mouth is shaped prior to saying the /s/
sound in the word ‘so’ for example, it forms a different shape to
when he / she shapes his/her mouth to
say the /s/ sound in the word ‘see,’ even though the /s/ sound in
both words is acoustically identical. This is because the words ‘see’ and ‘so’ are acoustically
indivisible. One of the questions that remains unresolved in the literature, is
how, if phonemic segmentation cannot be explained in terms of division of the
acoustic signal, phonemes come to be recognised.
For convenience, the two dimensions of
phonemic awareness are referred to as ‘discrimination’ and ‘blending’
henceforth in this text. Of the two dimensions
Yopps’ research (1988) confirmed the findings of earlier studies showing
that blending tasks were consistently found to be more difficult than
discrimination tasks.
2.6 Assessing phonological awareness
Adams (1990) suggests that phonemic
awareness is not a single, but a multi-dimensional trait. As discussed earlier,
stages of phonemic awareness development were identified by Bradley and Bryant,
(1983) and Treiman (1985). Testing for these stages of development was
considerably refined by Yopp (1988). She reviewed a variety of tests that had
been developed to operationalise the concept of phonemic awareness. The tests
she reviewed included rhyming tasks (Calfee, Chapman & Venezky, 1972),
counting syllables in words (Liberman, Shankweiler, Fischer & Carter,
1974), phoneme segmentation (Williams, 1980), phoneme blending and phoneme deletion
(Helfgott, 1974; Calfee, Chapman & Venezky, 1972) and identifying
alliteration (Bradley & Bryant, 1978).
Yopp (1988) undertook a study to compute
the reliabilities of the tests that operationally define phonemic awareness and
to select the test or combination of tests of phonemic awareness that best
predict success in the early stages of learning to read. She administered ten phonemic awareness
tests in three elementary schools to a cohort of 104 kindergarten children who
were aged from 5 years, 4 months to 6 years, 8 months. She also administered a
‘learning test’ which assessed children’s ability to “use sound-symbol
correspondence to decode printed artificial words” (pg. 167). She remarked that artificial words, or “pseudo-words”
were frequently used in reading research in order to eliminate the effects of
word familiarity or prior learning, citing the work of Stanovich et al.(1984)
and Zinna, Liberman and Shankweiler, 1986. (Words of this nature are
referred to hereafter in this text as ‘nonsense words.’) Underlining the
importance of the ability to decode nonsense words, Yopp (1988) asserts the
following:
“This ability, which distinguishes the code from the cipher stage of reading acquisition, requires the child to be able to use the systematic correspondence between sound and print. Eventually, regardless of the method of instruction, this ability must be acquired in order to achieve independence in reading.” (pg. 163).
The results of Yopp’s analysis of the results of the phonemic awareness subtests revealed a wide variation in the mean subtests scores. The rhyming task was found to be easiest where the mean score achieved by the kindergarten children was .714. Tasks involving manipulation of phoneme sounds proved to be most difficult with mean scores ranging from .584 for the phoneme counting task to as low as .251 for the phoneme deletion task. These findings are consistent with the findings of similar studies by Perfetti et al. (1981) and Stanovich et al. (1984).
2.7 Phonemic awareness and early reading acquisition
A strong connection has been clearly
established be phonemic awareness and reading (Bryant, MacLean, Bradley, &
Crossland, 1990). The relationship between phonemic awareness and reading
acquisition is interactive and complex.
Some aspects of phonemic awareness precede reading and other aspects are
the result of reading. (Stahl & Murray, 1994 ) This reciprocal relationship
is exhibited throughout the studies that were reviewed. The apparent importance
of early rhyming skills was demonstrated by Bryant, MacLean, Bradley &
Crossland (1990). They concluded that there is a direct connection between
ability to recognise or supply rhyme and the ability to read. Sensitivity to
rhyme and alliteration are precursors to phoneme deletion skills, which in turn
underpin the ability to isolate and recognise phoneme sound in words.
However, there is by no means universal agreement among researchers about the importance of rhyme or about its relation to reading. McGuinness (1997) reporting on several recent studies in the US, UK and Sweden concluded that children “do not use rhyming endings to decode words, they hardly ever decode words by analogy to other words and that ability to dissect words into onsets and rimes has no impact whatsoever on learning to read and spell.”(pg. 141) She cites a study by Nation and Hulme ( 1997) in which they tested seventy five children in Years 1, 3 and 4 in UK schools (aged approximately 5 to 9 years) on various phonemic awareness tests. Onset-rime tasks were found to be difficult: the mean score achieved across the three age / grade levels of children was consistently in the 55 per cent region. The authors concluded that children’s ability to perform onset-rime segmentation was not related to literacy.
These apparent contradictions in the literature about the role of
rhyme and analogy are potentially
confusing for curriculum developers and for teachers. The argument that rhyme
and alliteration both raise children’s awareness of pattern in spoken language
and that this maps to awareness of letter-sound relationships at a later stage
is a compelling one. Analogy may even
play a part in children’s perception of phonemes.
Liberman (1973) found that at the end of first grade, when children were aged 7, nearly 30 per cent of them had no concept that words could be segmented into phonemes. When Yopp (1988) correlated several phoneme awareness subtests with nonsense word reading, she found the highest predictive correlation between the phoneme subtests and the nonsense word reading test involved two tasks that required children to segment words phonemically. In both tasks, the predictive correlation exceeded .70. The ability to rhyme had the lowest predictive correlation at .47.
As well as correlational evidence indicating that phonemic awareness
is strongly predictive of reading attainment, a number of more recent studies
has shown that phonemic awareness has a causal influence on reading
development. This finding is of great significance, for without it one could
argue that phonemic awareness is purely a consequence of reading development,
or alternatively merely related to a third causal variable such as language
development, intelligence, or social class.
Bradley and Bryant (1983) studied the predictive ability of
sensitivity to rhyme and alliteration. They were interested in whether high
levels of sensitivity were associated with later reading success and low levels
with reading difficulty over the next four years. They screened 118 three year
olds, and 285 five year olds for reading ability, and none was able to read any
words on a reading test. This is an important safeguard because of the known
influence of reading ability on phonemic awareness. They were able to demonstrate
high correlations between the original sound categorisation scores and
children’s reading and spelling over three years later. They selected 65 of the
children with low (below 2 SD from the mean) phonemic awareness scores, and
randomly assigned them to a training, and a non-training group. The first group
was taught (in 40 sessions over two years) to attend to the sound structure of
words, while the second was taught to categorise words in terms of their
meaning. The children received normal reading instruction in school and at the
end of the project were re-assessed. The group that had received training in
the sound structure of words made significantly more progress in reading than
the group taught the meanings of words - an effect specific to reading as the
two groups were similar in a standardized maths test.
Bradley (1990) retested the original experimental and control groups
five years after the training was completed. The differences, quite remarkably,
were still present in all four reading and spelling tests administered. The
value of early intervention in sound categorisation is obvious. Those children
in the experimental group who were also taught letter-sound correspondences,
and how sound and letter patterns are connected, performed far better than all
other groups.
2.8 Teaching phonemic awareness
development
Byrne and Fielding-Barnsley (1989, 1995)
conducted studies in which the teaching of both phonemic awareness and
letter-sound correspondence to pre-readers was necessary to establish the concept
of phonemic awareness. Using a program they had developed, Sound Foundations
(Byrne and Fielding-Barnsley (1991), they examined whether young pre-school
children could be taught the phonemic awareness by emphasising how different
words may begin or end with the same sound (phoneme discrimination). It was
their belief (as earlier Bradley & Bryant, 1983 had argued) that phoneme
discrimination was the ideal means for promoting more global phonemic awareness
skills, including blending, as its
attainment implies the presence of segmentation skills. The question for them
was whether a well aimed, but minimal intervention could achieve an outcome
similar to a programme that undertook to teach children to recognise all 44
English phonemes. Their minimalist approach extended to teaching only a
representative range of sounds (seven consonants and two vowels) over a
twelve-week period. Compared to a control group who were involved in a global
phonemic awareness training regime, their children gained in phonemic awareness,
which extended even to sounds that they had not been taught.
Follow-up research (Byrne and
Fielding-Barnsley, 1993, 1995) indicated that this advantage extended to
superiority in nonsense-word decoding in Years One and Two, and to reading
comprehension in Year Two. They concluded that children who enter school with
some understanding of the structure of words prior to significant print
experience find it easier to master the task of decoding written language.
Other research has also demonstrated that
phonemic awareness can be taught (Ball & Blachman, 1991). Much of the
research reviewed asserts that unless it is taught, many children will fail to
learn to read. Griffith and Olson
(1992) suggest that teachers use literature that makes playful use of the
sounds of language (e.g., makes use of rhyme, alliteration, or assonance) and
offer several examples of titles of children’s literature that support such an
approach. Hallie Yopp (1995) presents
an annotated bibliography of read-aloud books that are especially useful for
developing phonemic awareness in children and she provides suggestions for ways
in which teachers can best use the books. The books all involve explicit play
with language, are appropriate for young children, and lend themselves to
creative extension.
2.8.1 Direct or indirect instruction in phonemic awareness?
Adams’ (1990) and Chall’s (1967) studies
converge to conclude that teaching phonemic awareness benefits all children
learning to read, but that children who can be identified as ‘at risk’ on
predictive variables benefit especially from instruction in phonemic awareness.
They also emphasise the importance of structured, organised, direct
instruction, because it encourages children to use the alphabetic system to
decode an unknown word rather than to guess. Adams concludes that research
indicates that direct
instruction in the alphabetic system is more effective than an indirect
approach which, she contends, lacks precision, order and clarity. Some children
will, according to Adams, intuitively work out the code for themselves, but all
will learn faster and better by receiving organised and explicit
instruction using decodable text mostly comprised of words that contain the
sounds or symbols being taught.
By direct instruction or explicit instruction, commentators usually mean teaching the phonemes that correspond to letter combinations directly, rather than encouraging children to infer the letter sounds from words that they recognise by sight. Structured, organised instruction usually implies teaching all 44 sounds and decodable text suggests text artificially created for the purpose of teaching reading. The Byrne and Fielding-Barnsley (1991) and subsequent follow-up studies suggest that such a structured approach may not be necessary for all, or even for most learners. Also, using artificially created texts to teach reading, rather than children’s literature, has been the subject of vociferous debate for the last three decades and compelling arguments against such an approach abound in the literature. (Smith, 1978, 1992; Goodman, 1986)
Other activities that can be used as the
basis for teaching phonemic awareness which children enjoy are activities
involving rhyming, riddles, guessing games, storytelling and word games. Some of
the leading researchers in phonemic awareness, including Adams, Foorman,
Lundberg, and Beeler (1998) have produced a book containing guidelines for a
phonemic awareness curriculum. It contains dozens of carefully sequenced
activities for teaching phonemic awareness. The curriculum includes teaching
objectives and lesson plans, ideas for listening games and sound manipulation
exercises and an informal screening device. The recommended activities can be
completed in approximately 15 minutes per day.
2.9 Phonemic awareness development of children with SLD
Specific learning difficulties as defined
in the Special Education Review Committee Report (1993) (see Chapter One above)
is closely related to the definition of developmental dyslexia offered by Harm
and Seidenberg (1999). They define developmental dyslexia as ‘failure to
acquire age-appropriate reading skills, despite adequate intelligence and
opportunity to learn.’ (pg. 507). Foorman and Fletcher (2000) cite the
following ‘research-based’ definition of dyslexia:
Dyslexia is one of several distinct learning disabilities. It is a specific language-based disorder of constitutional origin characterized by difficulties in single word decoding, usually reflecting insufficient phonological processing abilities. These difficulties in single word decoding are often unexpected in relation to age and other cognitive and academic abilities; they are not the result of generalized developmental disability or sensory impairment. Dyslexia is manifest by variable difficulty with different forms of language, often including, in addition to problems with reading, a conspicuous problem with acquiring proficiency in writing and spelling.
This latter definition differs mainly from other definitions by the inclusion of language and phonological processing abilities as causal factors in the disability.
The causes of developmental dyslexia has
been the subject of considerable debate ranging over almost a century. Interest
has focused on areas as diverse as genetics, word blindness, distortion in the
visual field, mixed laterality and impaired function in the language area of
the brain. More recently, research has focused on the lack of phonemic
awareness as possibly the main cause of developmental dyslexia or specific
learning disability.
Poor readers generally have been identified
as having poor phonemic awareness skills. Those designated as having a specific
learning difficulty have shown lower use of rhyme in a cued recall task when compared
with younger readers (Rack, 1985 cited in Pumfrey & Reason, 1991). There
have been concerns expressed in the literature that children with specific
learning difficulties may be irreversibly insensitive to phonemes (Torgesen, Wagner, & Rashotte, 1994),
and thus unable to derive any significant benefit from a phonemic awareness
programme. However, a study by Alexander, Anderson, Heilman, Voeller, and
Torgesen (1991) involving a group of
children with SLD (aged 7 years 9 months to 12 years 10 months) noted
significant improvement in phonemic awareness and phonological recoding
following instruction in the Auditory Discrimination in Depth Program
(Lindamood & Lindamood, 1969).
Swan and Goswami (1997) conducted an interesting experiment in which they tested the picture- and word-naming performance of developmental dyslexics and compared the results with the picture and word naming performance of non-dyslexic poor readers, and with other control groups that were matched to the experimental group for reading age and chronological age. The stimulus list used for both tasks was manipulated for word length and word frequency. In order to examine picture-naming errors in more depth, an object-name recognition test assessed each subject's vocabulary knowledge of those names which they were unable to spontaneously label in the picture naming task.
Findings indicated that the dyslexic and
non-dyslexic poor readers exhibited a picture naming deficit relative to both
chronological and reading age-matched controls. Findings also indicated that
both groups of impaired readers obtained better scores in the word- naming task
than in the picture naming task, while both groups of controls showed no
difference in performance across tasks. The dyslexics' picture naming errors,
but not those of the non-dyslexic poor readers, were particularly marked on
polysyllabic and/or low frequency words, indicating a possible phonological
basis to the picture naming deficit of the dyslexic children.
Foorman & Fletcher (2000) summarise the
findings of important interventions in helping children with dyslexia. Firstly,
they suggest that explicit instruction in word recognition skills emphasising
the alphabetic code results in more favourable outcomes than does a
context-emphasis or embedded approach (Smith, 1978). Secondly, they conclude
that the type of direct instruction alphabetic approach is less important than
intensity, duration and explicitness, so long as it is structured and provides
opportunities to engage in reading and writing. Thirdly, they report a
University of Texas study (author not cited) which found that 15 minutes of
instruction in the alphabetic code as part of a standard kindergarten
curriculum led to significant gains in phonological analysis skills relative to
children in the same curriculum who did not receive this training. Only an explicit instruction approach was
associated with average levels of reading proficiency after one school year of
intervention. They conclude that instruction for children with dyslexia needs
to be more explicit and intensive and that the children involved need to be
supported emotionally as well as cognitively. Finally they suggest that the
following quotation from Hinshelwood, an ophthalmologist who was among the
first to study dyslexia, is as true today as it was almost a century ago:
It is a matter of the highest importance
to recognize the cause and the true nature of this difficulty in learning to
read which is experienced by these children, otherwise they may be harshly
treated as imbeciles or incorrigibles and either neglected or flogged for a
defect for which they are in no wise responsible. The recognition of the true
character of the difficulty will lead the parents and teachers of these
children to deal with them in proper way, not by harsh and severe treatment,
but by attempting to overcome the difficulty by patient and persistent training
(Hinshelwood, 1902).
2.10 Computers and Learning
Before turning to the question of literacy
specifically, what is the evidence generally that computers can make teaching
and learning more effective? Research has provided some convincing arguments
for the use of computers in teaching and learning. One of the most
comprehensive compilations of studies on the effectiveness of computers in education
has emanated from the University of Michigan, where several meta-analyses of
research on computer-aided instruction conducted over a number of years have
been undertaken. The meta-analyses involved the use of a statistical technique
called effect size (difference
between means of control and experimental group divided by standard deviation
of the control group). The effect size is calculated by using a common
measurement scale across a number of studies and it facilitates an estimate of
the size of the effect of an experimental treatment. An effect size of 0.3 is
considered to be moderate, but significant (Kulik & Kulik, 1991) and is
equivalent to learning what would normally take 15 months in 12 months.
During the 1970s and early 1980s, many assessments were conducted
of the effectiveness of computer-assisted instruction (CAI) compared with the
traditional forms of teaching. During this period, CAI was used primarily to give students drill-and-practice exercises
and feedback in basic knowledge and skills, such as number facts in arithmetic
and sight word recognition in reading. Assessments of the learning outcomes of
CAI compared with traditional methods of instruction have consistently shown
the superiority of CAI (Kulik, Kulik and Bangert-Drowns, 1984).
The main finding of these studies was that
students learned more effectively and were better motivated when computers were
used in the delivery of instruction (Kulik, Bangert, & Williams, 1983;
Kulik, Kulik, & Bangert-Drowns, 1984). For example, Kulik, Bangert, and
Williams (1983) found one study that recorded an 88% saving in learning time
using computerised instruction (90 minutes) versus classroom instruction (745
minutes) and another study that recorded a 39% savings in learning time (135 minutes
for computerised instruction versus 220 minutes for classroom instruction).
Both studies involved computer simulation instruction in physics. Kulik, Kulik,
and Schwalb (1986) identified 13 studies in which students using computers
mostly for tutoring learned in 71% less time than students in traditional
classroom instruction.
Following a later meta-analysis of almost 200 studies (Kulik &
Kulik,1987) it was concluded that as well as learning more in classes where there was computer based instruction, students
learned more quickly, liked classes more and developed positive attitudes to
computers.
A note of caution in interpreting the
findings is warranted. The studies involved a very wide range of learners,
including elementary pupils, high school students and adults in industry and in
the military. Results were consistently stronger in published studies than in
those that were not published, effects were greater when different teachers
taught experimental and control groups and effects tended to be greater in more
recent and in short studies. Nonetheless, the studies do provide reasonably
strong evidence that computer-based instruction can yield results that are
significantly better than traditional classroom-based instruction. It is not yet clear whether this is true for
all types of learning or for most students.
A report of an expert committee to the
President of the United States (United States, 1997) offers the following words
of caution about the reliability of meta-analytic studies:
While the preponderance of evidence would seem to argue for the efficacy of traditional computer-assisted instruction, some researchers have raised questions related to the methodology employed in these studies, or to the interpretation or import of the results they yielded. In particular, issues have been raised regarding the size and experimental designs of many of the underlying studies, the amenability of these studies (which often differ significantly in multiple dimensions) to meta-analytic aggregation, the robustness (after controlling for various contextual factors) and temporal persistence of the measured effects, the independence of those responsible for evaluating efficacy, and the possibility of systematic bias against the publication of negative results.(www.ed.gov.ie)
2.11 Computers and literacy instruction
If computers are therefore capable of
assisting instruction, what kinds of knowledge and skills can be more
effectively developed using computers rather than traditional classroom
approaches? Can computers help to develop the skills that are specific to
learning to read? Knowledge that is comprised of facts, procedures and rules of
discourse is usually taught and learned in ways where a teacher plays a
significant role in content delivery (Duffy, Lowyck, & Jonassen, 1993). Contemporary educational
theory suggests that higher order learning that involves developing declarative
knowledge is best achieved through teaching processes that support
learner-centred, collaborative and generative activity. Such instruction places
the learner at the centre of the teaching and learning process and active in
constructing a personal meaning of the content being delivered.
Lower order skills, such as the development
of phonemic awareness, acquisition of a sight-vocabulary and development of
literal comprehension skills can easily be mediated through computer-assisted
instruction. Such software is now frequently produced to accompany published
reading programmes. Examples such as The
Oxford Reading Tree, published by Oxford
University Press and Reading 360 published
by Ginn come to mind.
On
the other hand, higher order skills such as inferential and evaluative
comprehension that would be traditionally thought to be dependent on facilitation
by the teacher have been shown to be amenable to development through the use
of computers. Lehrer (1993) describes
the development, use and results of a hypermedia construction tool called HyperAuthor that 14-year old children
used to design their own history lessons about the American Civil War. This
approach is based upon the constructivist view that knowledge is a process of
design and not something to be transmitted from teacher to student. Lehrer's
students were engaged in a process whereby they were communicating their
interpretation of material they had read by designing their own hypermedia. In
this way, they were determining what was significant and what was peripheral in
the material they had read. Thus, they were demonstrating the sophisticated
reading comprehension skill of selecting main ideas and separating them from
supporting details.
According to Lehrer, "The most
striking finding was the degree of student involvement and engagement" (p.
209). At the end of the study, students
in the hypermedia group and a control group of students, who had studied the
American Civil War via traditional classroom methods during the same period of
time, were given an identical teacher-constructed test of knowledge. No
significant test differences were found.
However, a year later, when students in the experimental and control
groups were interviewed by an independent interviewer unconnected with the
previous year's work, important differences were found. Students in the control
group could recall almost nothing about the historical content, whereas
students in the experimental group displayed elaborate concepts and ideas that
they had extended to other areas of history.
Therefore, the students who had used computers to support their learning
displayed much greater long-term recall of what they had read, in Lehrer's
opinion had "knowledge that was richer, better connected and more
applicable to subsequent learning and events" (p. 221).
An American
study (Schultz, 1995, quoted in Software Publishers' Association, 1996)
examined the outcomes of an early literacy approach that combined Talking Books with activities to support
listening, speaking, reading and writing. Children in first grade who used the
technology system were found to have made greater gains in basic language
skills and reading comprehension than children who received traditional
instruction. More recently, McKenna (1997) described the use of electronic Talking Books with beginning and
disabled readers. The extent to which
children used the available supports to help them to read the text was
examined. Students could choose to hear
digitised pronunciations of individual words by clicking on the words. However,
it was found that they sometimes clicked on words they knew and sometimes
failed to click on words they did not know. A possible explanation for this
slightly confusing finding could be that children at the early stages of
mastering reading who lack the metacognitive awareness of older children may
have clicked on words they knew to confirm their knowledge. Similarly, it is
also a possibility that they thought they knew words that they did not in fact
know.
2.12 Computers and phonemic awareness development
Taylor (1980) proposed three roles for
computers in education, tool, tutor and tutee. A computer can be defined as
tutor where experts in a given area of knowledge, such as science, mathematics
or aspects of literacy skills are involved in defining the scope of the knowledge
to be taught and sequence in which it will be presented to the learner. It is
then possible for the computer to presents concepts and information to the
learner in a framework defined by the programmer. The learner can sometimes be
placed at a point on the learning continuum based on an initial assessment of
their knowledge of the topic in question. The computer also keeps records of
students' performance, it can be programmed to allow them to skip material that
they have already mastered or to repeat tasks that they have failed to
accomplish at the first encounter.
A computer as tutor can readily be applied
to developing phonemic awareness. In a French study focusing on disadvantaged
immigrant children in Paris, Cohen (1995) found that the use of voice
synthesisers that enabled children to hear the sounds of letters and words as
they typed contributed to children's phonological and reading development. The
results of the study indicate improved reading and writing achievement levels
for immigrant children when computers are used.
A tutorial and practice software programme
called Daisy Quest was developed to
increase young children's phonemic awareness. It was designed to help children
to understand the relationship between letters and their sounds by recognising
identical sounds in different words, such as the 'f' sound in 'fish' and 'fat'
or the final sound combination in 'ham' and 'jam.' In two separate studies of
different measures of phonological awareness (Foster et al. cited in Bialo
& Sivin, 1996), the computer-based approach was found to be significantly
more effective than traditional classroom instruction. The effect size was 1.05, which was considered to be significant.
2.13 Summary
The literature review provided insights
into the evolution of the current state of knowledge about the relationship
between phonemic awareness development and reading. There appears to be
unanimous agreement among researchers that phonemic awareness is a vitally
important correlate of reading development. Some authorities suggest that the
relationship between phonemic and reading development is causal. Not surprisingly, therefore, poor readers
are very frequently found to have very inadequate phonemic awareness development.
A widely accepted definition of developmental dyslexia reflects an acceptance
that lack of phonemic awareness is one of the underlying causes of dyslexia, or
specific learning disabilities, as it is referred to in Ireland.
The discovery that onset and rime is an
intermediate level of phonological awareness between syllabic awareness and
phonemic awareness is relatively recent. A relationship between onset-rime
awareness and reading has yet to be established, however. Assessments of
phonological awareness have found that rhyme is among the easier phonemic
awareness tasks and segmentation, manipulation of phonemes and blending phonemes are among the more
difficult. These latter skills are closely related to the ability to read. Most
authorities agree that phonemic awareness can be taught and assessment of
phonemic awareness has tentatively established
the order in which phonemic awareness should be taught. Interestingly,
one study showed that when a limited number of letter-phonemes had been taught,
many children appeared to grasp the alphabetic principle. The amount of
teaching required in order to establish phonemic awareness has yet to be fully
established. A number of studies have found that computer-based tutorial
programmes in phonological and phonemic awareness contributed positively to children’s
reading development.
Chapter 3: Methodology
3.1 Introduction: Context of the Study
The Department of Education and Science
established three special schools in Dublin for children with SLD, one in 1975,
one in 1981 and a third in 1994. Demand for places at the schools usually far
outstripped the number of places available, with the result that some pupils
who met the criteria for enrolment were unsuccessful in gaining a place at any
of the schools. Also, because of difficulties in transporting children to the
schools, about 25 per cent of parents whose children were referred to the
special schools did not subsequently follow through with an application for
enrolment.
In 1996, the first of a number of special
units for children was SLD was established, initially with an enrolment of 11
children. The unit was attached to a large Dublin national school and a teacher
from the staff was allocated to the unit.
This unit was designed to cater for pupils in north-eastern Dublin who
were unable to attend the special schools which were all located on the south
side of Dublin. The enrolment in this special unit has now increased to 33
children. Since 1996, special units were established at a further three schools
in the Dublin area and at a number of schools throughout the country. All of the pupils who were newly enrolled in
the Dublin special schools and support units for children with specific
learning disabilities in September 1999 are the subject of this research.
The
pupil-teacher ratio in these special schools is 11: 1 and the pupils follow the
normal school curriculum with the exception of the Irish language. Because
pupils do not learn Irish, they spend more time than pupils in ordinary
national schools on activities designed to help them to develop reading and
writing skills.
3.2 Characteristics of the subjects of the study
The subjects for the study were the 140
pupils enrolled in the special schools and support units for children with specific learning disabilities. The
Department of Education and Science provided revised guidelines to special
schools and mainstream schools that had a support unit for children with
specific learning difficulties in a letter in 1998 (Department of Education,
1998). These guidelines were already in place for the enrolment of children
with SLD in the special schools. The letter established three main criteria for
the enrolment of pupils with SLD in special provision, that is, either a
special school or a support unit.
General Ability
One of the defining characteristics of
children with SLD is that they are of average intelligence, or above.
Accordingly, the Department’s criteria for enrolment in special provision for
SLD pupils stipulates that an assessment by a psychologist on a standardised
test of intelligence should place general intellectual ability within the
average range, or above. This is equivalent to a standard score of 90, or
above, on an intelligence test. Where
it could be deemed unwise to calculate an overall I.Q. score because of
discrepancy between verbal and non
verbal abilities, one of either the verbal or performance scores should be
within the average range and the combined scores should indicate potential
ability within the average range.
This latter qualification leaves some discretion
to enrol children whose overall ability might be marginally outside the average
range, but who otherwise satisfy the criteria for enrolment. Therefore, some
pupils enrolled are marginally below the average range of intelligence.
Reading Ability
The Department guidelines clearly specify
that children enrolling in SLD provision should have very limited reading
ability in comparison to their peers. As not more than 2% of the overall pupil
population would be expected to have severe specific learning disabilities
(Ireland, 1993, pg 90), performance in basic literacy as measured by a
standardised test should be at a very low level. There must be an obvious
discrepancy between general intellectual ability and performance on a
standardised test of reading ability, the former being at or above the 25th
percentile for the overall pupil population and the latter being at or below the 2nd percentile
(Department of Education and Science, 1999).
Where children’s general ability is
assessed as above average, a child with a reading score above the 2nd
percentile could be enrolled. Thus, it can be inferred that the criteria for
enrolment are based on a discrepancy between reading achievement, as assessed
by a standardised test of reading ability, and intelligence, as defined by a
score on an intelligence test.
Age
Pupils transferring to a special school or
support unit for children with specific learning difficulties should have
completed Second Class in a primary school or be at least eight years old on the
first day of the school year. Most
pupils enrolling in special provision for SLD would therefore have spent a
minimum period of four years in mainstream school and most would have had the
benefit of remedial teaching for varying periods of time.
3.3 Design of the Study
The experimental hypothesis for the study
stated that children with specific learning disabilities would develop phonemic
awareness more effectively through an approach where computer software was used
to develop phonemic than they would if they were taught phonemic awareness
through other methods alone.
3.2.1 Development of the study
In
preparation for the study, the researcher
arranged a demonstration of the ‘Units of Sound’ software in December,
1998 by Mrs. Margaret Rooms of the British Dyslexia Institute for teachers from
all of the special schools and support units for children with specific
learning disabilities. The nature of the study was explained at the meeting and
it was agreed that each special school and support unit would use the ‘Units of
Sound’ software with half of the newly enrolled pupils for approximately 15
minutes per day for six months during the following year. The protocols for the
administration of the reading tests and the phonemic awareness tests were also
agreed at the meeting.
One copy
of the ‘Units of Sound’ software and the supporting documentation was supplied
to each of the schools at the meeting. Teachers were encouraged to install the
software and to become familiar with it
before the commencement of the experiment.
3.2.2 Assignment of pupils to treatments
As mentioned earlier, the subjects for the
study were the 140 pupils enrolled in three special schools and three support
units for children with specific
learning disabilities attached to ordinary schools in Dublin in 1999. All of the newly enrolled pupils in the
schools were involved in the study. These pupils were assessed prior to the
study using measures to assess their reading ability and their level of
phonemic awareness.
Pupils were assigned to either the ‘US’
(‘Units of Sound’ software) condition or to the ‘OM’(‘other methods’) condition
by the school principals in consultation with the researcher. As there were
constraints within the schools regarding the assignment of pupils to classes,
it was not always possible to achieve a random assignment of pupils to
treatments. The age of the pupils and their assessed reading ability were taken
into consideration in the assignment of pupils to classes and this had a
determining effect on the treatments in a significant number of cases. The
number of pupils in each treatment and their reading and phonological awareness
ability was therefore intended to be equivalent in both the US and the OM groups. Pupils were therefore assigned
on the basis of matched groups that took account of both reading ability and
phonemic awareness development at the pre-treatment stage. Gender was not taken
into account in the matching of the groups.
3.4 Assessment of pupils prior to enrolment
The procedure for enrolling pupils with
SLD in special provision is normally
initiated by either a psychologist who has carried out a psychological
assessment or by the principal teacher of the child’s school. Following
consultation with a child’s parents, the psychological report is sent to a
special school or to a school with a support unit. If the child is suitable for
enrolment, the child is placed on a waiting list. The psychological reports of
children on the waiting list are then examined by the principal teacher and an
inspector from the Department of Education and Science. Typically, psychological reports provide
information about the child’s intellectual ability, their scholastic
achievement in subjects such as reading and mathematics, their family
background and social and emotional adjustment.
Children who meet the Department of
Education and Science criteria for enrolment and are approved for enrolment by
the Department inspector are eligible for enrolment. When children have been
approved for enrolment by the inspector, the principal of the special school or
the school with a support unit then prepares to enrol pupils in accordance with
the enrolment policy of the school. At this stage, the principal usually
requests a report from the child’s school and invites child’s parents to visit
the special school or support unit to
discuss the option of enrolling the child there at the beginning of the
following school year. These interviews take place usually between the end of
February and the beginning of June. Following these interviews, pupils are
selected for enrolment in the schools the following September.
3.5 ‘Units of Sound’ software
The computer program chosen to support SLD
pupils’ development of phonemic awareness in this study was ‘Units of Sound’
which was developed by the British Dyslexia Institute in 1996. This program was
chosen because it was devised specifically for children with specific learning
disabilities. The software was designed with a single teaching purpose, that is, to assist children with
specific learning difficulties to acquire phonemic awareness. The programme
consists of three separate CD-ROMs, one CD-ROM for each stage of the programme.
Each CD-ROM is accompanied by a printed workbook which children use in
conjunction with the program. As ‘Units of Sound’ is a structured programme, it
is intended that children would move sequentially from Stage One to Stage
Three. However, each CD-ROM is a
‘stand-alone’ program and it is therefore possible to purchase each Stage
separately and to install each CD-ROM on separate computers.
To date, no independent assessment has been
made of the effectiveness of this software. Neither are any reviews of the
software available on educational websites in Ireland or the UK such as
Scoilnet or BECTA Although the programme
in its non-computerised format has existed for almost 30 years, no formal
evaluation of its effectiveness in developing phonemic awareness could be
found. According to a web page dedicated to ‘Units of Sound’ on
the British Dyslexia Institute website it is ‘a multimedia literacy development
programme suitable
for age groups 9 years through to adult.’
3.5.1 Origin of the program
Originally, Units of Sound was developed as
a tape-based resource for dyslexic children, but it does not appear to have
been well known or used very extensively. Writing about the project to convert
the taped version of the programme to CD-ROM format, Margaret Rooms, Education
Officer with the British Dyslexia Institute (1997) describes how she first
became of the Units of Sound programme:
I previously described it as “the best kept literacy secret in England” as it was never advertised or shown at exhibitions until the Dyslexia Institute took over the copyright in 1993. I myself came across the material by falling over a box in the Tonbridge Dyslexia Institute five years ago. I started using Units of Sound with dyslexic pupils at a nearby prep school. At the end of the first term I knew I had something that should not be hidden away. Pupils whose reading had ‘stuck’, in spite of regular specialist teaching, all made measurable reading gains after one term using Units of Sound.
Later in the same article, Rooms suggests that one of the main advantages of the CD-ROM based material over the taped material is one of accessibility, since ‘finding a particular point on a tape could waste a lot of time.’ She concludes that the main benefit in transferring the material to CD-ROM was to make it more accessible and because children like to work with computers. She add a cautionary note, however:
“Transferring that material to CD has not altered its main purpose. It is not 'edutainment'. It does not set out to coerce children into learning whilst kidding them that it is really fun. It is the achievement of learning to read that is enjoyable and we have found that once children realise they are making progress they are sufficiently motivated to continue.”
This may account for the sparse use of
graphics in the final design of the programme, a matter that was commented on
by several of the teachers who used the programme. There may have been a sound
pedagogical rationale for the omission of graphics, as it is widely recognised
that a high proportion of dyslexic children are easily distracted and any
gratuitous use of visually stimulating material could be counterproductive.
Thus, the simple on-screen layout may have positive benefits for pupils with
dyslexia and potentially increases the age appropriateness of the program for
older children.
3.5.2 Programme Content
‘Units of Sound’ provides a structured
sequence of phonemic awareness activities in multimedia format. Stage One
begins with initial consonant and short vowel sounds and the programme
progresses to Stage Three where children are taught to decode multi-syllabic
words. Each 'lesson' focuses on a
particular 'unit of sound', which is combined with a visual whole word
approach. Each sound is seen, heard, spoken and practised within blocks of
words. Children have to copy the phonemic sounds generated by the computer and
blend them to form words. Having practised the words on the computer, children
then read the words in continuous text from the book that accompanies the
program.
Stage One of the programme is described by
the Teachers’ Manual to be suitable for children with reading ages from 6 –8
years, Stage Two for children with reading ages from 8 – 10 years and Stage
Three for children with reading ages from 10 – 12 years.
The programme content and the approach to
instruction in ‘Units of Sound’ would appear to conform to what is believed to
be a sound pedagogical approach for children with specific learning
disabilities. It provides opportunities for ‘overlearning’ that is ‘considered
necessary in order to achieve the automaticity that is the hallmark of the
competent reader and writer.’ (Pumfrey & Reason, 1991, pg 118). It also combines structured, explicit teaching of
phonemes with systematic recording of children’s progress in acquiring phonemic
knowledge.
3.6 Test Instruments
Prior to their assignment to treatments,
all children enrolling in the special schools and support units were assessed
using a test of reading achievement and a test of phonemic awareness
development.
3.6.1 Reading achievement test: The Revised Neale Analysis of Reading Ability
The Neale Analysis Of Reading Ability:
Second Revised British Edition (Neale, 1997) consists of
a set of graded passages for testing the Rate, Accuracy, and Comprehension of
oral reading. It is widely used in Ireland by remedial teachers to assess the
reading proficiency of individual pupils. It is also used extensively in schools in Australia and the
UK.
It is both an attainment test and a
diagnostic test. It can be used to assess reading progress objectively and to
obtain structured diagnostic or clinical observations of an individual’s
reading behaviour and error patterns, thus contributing diagnostic information
that can be used to inform the development of a learning programme.
The score achieved on the Neale Analysis Of Reading Ability: Second Revised British Edition depends on the number of pronunciation errors that a child makes while reading aloud a passages of text. One of the administration protocols states that if a child makes more than a specified number of errors on a passage, the child should not be tested on the subsequent passage(s).
Scores on the test are obtained by reference to tables provided in the administration manual. Scores can be reported as either reading ages or percentile ranks. There are parallel forms of the test, so teachers can test the pupils reasonably frequently without the pupils becoming too familiar with the test.
3.6.2 Test of Phonemic Awareness
A test was devised in order to measure each
pupil’s mastery of a range of phonemic awareness tasks (Appendix A). The test was a two-part test consisting of
65 items. The first part of the test comprised 40 items that were designed to test awareness of rhyme,
and alliteration and to establish whether children could delete and substitute
phonemes in a word or count the number of syllables in a word. The second
section of the test 25 phonically
regular nonsense words to test their ability to successfully blend phonemes.
The design of the test was influenced
considerably by the work of Yopp (1992) and Reynolds (1997). Yopp defines
phonemic awareness as
"the awareness that phonemes exist as abstractable and manipulable components of spoken language. It is the ability to reflect on speech and experiment (play) with its smallest components (phonemes). Phonemic awareness is not phonics and not auditory discrimination."
Her research outlines a progression of phonemic awareness development from pre-school to the early stages of formal schooling as including the following abilities:
to hear rhymes or
alliteration
to isolate the first
sound in a word (e.g. what is the first sound in the word ‘five’)
to count phonemes in
words ( how many sounds do you hear in ‘knee’?)
to identify the
beginning, middle, and final sounds in words
to substitute one
phoneme for another (e.g., change the /c/ in ‘cat’ to /r/)
to delete phonemes
from words (e.g., omit the /c/ from ‘cat’)
Using a test based on the above model, Reynolds (1997) tested the English language phonemic awareness of Japanese High School students who were learning English as a second language.
3.6.3 Design of the phonemic awareness test
Forty items in the test comprised eight sets of 5 items that were intended to assess the pupils’ ability to perform phonemic discrimination tasks such as the above. These items followed the design and sequence in the Reynolds (1997) study, although the number of items in some sections of the test was changed and some of the items were changed in order to adapt the test to the target age group. A further twenty-five items in the test assessed the children’s ability to pronounce a graded series of phonically regular nonsense words.
3.6.4 Development of the phonemic awareness test
In order to establish the reliability and
validity of the test, it was decided to administer the test to a cohort of 100
pupils in First Class in primary schools. Pupils in First Class were chosen
because most pupils would be expected to have acquired a significant level of
basic phonemic awareness skills at that stage. Remedial teachers in 25 schools in the Dublin / Wicklow / Wexford
region were asked to supply data on 5 pupils in First Class in the school(s)
where they were teaching. The schools were chosen with the intention of
obtaining a sample that would be typical of schools generally. Thus, large
urban schools, small rural schools, single-sex and co-educational schools were
included in the selection. In the
explanatory letter to remedial teachers, (Appendix B) teachers were requested to select 5 children that would be
representative of the range of reading ability found in a typical First Class.
A standardised test score obtained in the past year was recommended as the
basis for selecting the children, but in the absence of a test score, the class teacher's judgement would be
acceptable.
The remedial teachers were asked to select
one pupil from the top 20 per cent of the class in reading achievement, one
from the bottom 20 per cent and 3 from the middle range. They were asked to indicate on the record
form (Appendix C) either the standardised score obtained on a reading test,
or the class teacher's assessment of
the child's reading ability (top 20%, middle 60%, bottom 20% of class). In the
event, standardised test scores were furnished in respect of all the pupils.
Instructions for the administration of the phonemic awareness test were
supplied to all teachers and they were asked to return the phonemic awareness
test record forms before the end of June, 1999. The timing of the issue of the
letter to remedial teachers ensured that all pupils would be tested during the
month of June.
3.7 Administration of the phonemic awareness test
The test consisted of 65 items and was
designed to test the children’s ability to carry out a number of phonemic
awareness tasks. There were two distinct sections to the test: the first part
consisted of 45 items to test phonemic discrimination and the second part consisted of 25 items to assess the children’s ability to
blend phonemes. Each section of the test was further divided into subsets of 5
items each in order to examine achievement in subcategories within both
phonemic discrimination and phonemic blending.
Phonemic discrimination
The phonemic discrimination section of the
test comprised eight subcategories of 5 items each. Before testing, children
were told that the purpose of the test was to find out how well they could hear
sounds in words.
Pattern test
In the pattern subtest, the children had to
respond to 5 items where they had to supply a pair word for a target word in
order to complete a pattern. They were given three sample items initially (cat,
bat; kite, bite; call, ball) and a practice item (cart, ?). This set of items
tested the children’s ability to recognise and pair initial sounds in words.
Onset test
On the onset subtest, the children were
told they would hear a word and would have to say the first sound in the word.
Two practice items, chair and no were provided. The teacher said
the word chair and then said the
first sound is /ch/ and repeated this procedure for the word no. The
purpose of this section of the test was to establish whether the children could
isolate the first sound in a word from the remainder of the word.
Missing sound test
The purpose of this subtest was to
establish whether the children could identify the onset of a word when they
were presented with the whole word and a partner word with the onset missing.
They were told that they would hear two words at a time. The second word was
the same as the first, except that the first sound was missing. The children
were to say the sound that was missing. Two examples chair, -air; no,
-o, were provided as practice
items. The five pairs of words included three words with single-consonant
onsets and two words with consonant blend onsets.
Onset deletion test
This subtest was similar to the previous
task, the ‘missing sound test,’ except in this case, the children had to listen
to a target word and say what it would be if the first sound was deleted. The
children listened to the example item chair and were told that if the
first sound /ch/ was deleted, /-air/ would be left. After a
further practice item, the children responded to the five test items, chicken,
sheep, flower, queen, mouse.
Odd-one-out rhyme identification test
In this test, children had to select the
one word from a group of five words that did not rhyme with the other four
words. They were told that words rhyme when the sound near the end of words is
the same. For example, man and pan rhyme and cook and book
rhyme. They were given an example go, snow, chair, blow, row and
told that the go, snow, blow and row ended in an /o/
sound. Chair ends with the sound /air/. The difficulty of the
task was increased by including words in the list that were similar, (e.g., play,
plate) but did not rhyme.
Provide a rhyming word for target word
Children were told that they would hear one
word at a time. They were asked to say a word that rhymes with it. If they
couldn’t think of a real word, they could make up a word that rhymes with the
given word. The example mouse, house was provided for them.
Syllable test
The children had to listen to a word and
say how many syllables were in it. They were told that a syllable is like a beat
and that some words have one beat and other words have more than one beat. The
teacher who was administering the test then demonstrated this by saying the
syllables distinctly and by tapping a beat for each of the sample words, understand and chair.
Phoneme identification test
The children were asked to say what
individual sounds made up the word said by the teacher. They were asked to
break the word down to the smallest sound unit possible. The word chair was
used as a practice item. Children were told that the word chair had
three sounds, /ch-/, /-ai-/, -/r/. They then had to say the individual
sounds in the words nose, hand, friend, foot and knee. If they
successfully identified all the individual phonemes in a word, for example /n/,
/o/, /z/ for the word nose, they were given a correct score. If any
of the phonemes were omitted or combined with another phoneme, for example /n/,
/oz/ the item was scored as incorrect.
Nonsense words tests
The nonsense words tests consisted of 25 nonsense words in 5 sets of five words each, arranged in the anticipated order of difficulty. The 5 subtest sets consisted of regular three-letter syllables (cvc), initial consonant blends (ccvc), initial consonant digraphs (th, wh, sh, ch), long-vowel syllables formed by final ‘e’ (cvc[e]) and regular bi-syllabic nonsense words. Nonsense words were chosen to eliminate the possibility that children would recognise the words visually, thus obviating the need to recognise and blend the phonemes.
3.8 Validation of the phonemic awareness test
The development of the phonemic awareness
test was described above. The test consisted of 65 items and was designed to
test the children’s ability to carry out a number of phonemic awareness tasks.
There were two distinct sections to the test: the first part consisted of 45
items to test phonemic discrimination and
the second part consisted of 25
items to assess the children’s ability to blend phonemes. Each section of the
test was further divided into subsets of 5 items each in order to examine
subcategories within both phonemic discrimination and phonemic blending. This
test was used to measure the phonemic awareness of the two treatment groups in
the study, the ‘Units of Sound’ group (US) and the ‘other methods’ group (OM)
at the pre-treatment stage and at the post-treatment stage.
A reliable test of phonemic awareness for
children who had begun to learn to read was not immediately available. It was
decided therefore to develop a test that would assess the important elements of
phonemic awareness suggested by the literature review. A test was therefore
designed to assess children’s ability to complete phonemic discrimination and
blending tasks. Because the test had not been used previously, it was also
necessary to establish the psychometric properties of the test. Arrangements
were therefore made to have the test administered to a random sample of
approximately 100 First Class pupils. The achieved sample was 109 pupils. The test administration procedures are
described in detail earlier in this chapter.
First Class is generally regarded as a very
important stage in the reading development of First Class children in Irish
primary schools. The vast majority of children enrol at school between their
fourth and fifth birthdays and spend two years in Infant classes. The Infant
classes are intended as preparatory phase for formal learning. The official curriculum in reading for
Infants (Ireland, 1971, pg. 82) recommends that reading activities should focus
on the ‘cultivation of reading readiness.’ The Teacher’s Handbook suggests pre-reading activities such as
‘building up a basic sight vocabulary….exercises and reading games…use of
suitable reading scheme…reading for pleasure.’ It does not make any
recommendation about phonemic development in the syllabus for the Infant
section, but refers to ‘word study, including elementary phonic training’ in
the syllabus for First and Second classes. It could be concluded, therefore,
that First Class pupils at the end of First Class would have established a limited
range of phonemic awareness skills. It would also become evident at the end of
First Class if children were having serious difficulties in reading. A survey
of remedial teaching in Irish primary schools (Ireland, 1998, pg. 20) revealed
that the proportion of pupils selected for remedial teaching in English was
greater in First Class than in any other class in the school.
3.9 The First Class sample
The
First Class sample involved 109 pupils in 23 schools. The number of pupils in
each school on whom data was provided ranged from 2 to 7 pupils, but the
majority of schools provided data on 5 pupils.
Table 3.1 Descriptive Statistics First Class Pupils
Variable
|
N |
Mean |
St Dev |
Minimum |
Maximum |
|
Age |
109 |
7.63 |
0.44 |
6.9 |
9.3 |
|
Reading
Percentile |
109 |
54.65 |
26.84 |
1.00 |
99.00 |
3.9.1 Age of pupils
As illustrated in Table 3.1, the mean
chronological age of the 109 pupils was 7.63 years, but the age range was between
6.9 and 9.3 years, a range of 2. 4 years. Eighty-six of the pupils were aged
between 7.0 years and 8.0 years and 20 pupils were aged between 8.0 and 9.0
years. Only 2 pupils were less than 7.0 years and only one was more than 9.0
years.
3.9.2 Distribution of reading ability in the sample
The mean reading percentile of the pupils was 54.65 on a nationally
standardised test of reading ability, suggesting that the mean reading ability
of the sample was above the average for pupils at the of First Class. Table 3.2
displays the results of a one-sample t-test of the mean reading percentile of the sample of First Class
pupils against the mean percentile
score for the reading test, i.e. the 50th percentile.
Table 3.2: Test for significance of the difference between scores of First Class pupils and 50th percentile
Variable
|
N |
Mean |
StDev |
T |
P |
|
Read. Percentile |
109 |
54.65 |
26.84 |
1.81 |
0.073 |
Although the probability value (p=0.073) is not
quite less than .05 (the chosen significance level) it is very close to
it. It can be
inferred from SD of 26.84 that there was considerable variability in the scores
recorded.
3.10 Profile of the phonemic awareness development test
The review of the literature on phonemic awareness (e.g. Yopp, 1998;
Stanovich, Cunningham, Cramer, 1984) suggested that phonemic awareness may not
be a single construct. Yopp’s analysis of several studies (Yopp, 1998)
suggested that at least two factors may underlie phonemic awareness which she
refers to as ‘simple phonemic awareness’ and ‘compound phonemic awareness.’ She
concludes that phonemic awareness tasks fall into one or other category,
depending on the complexity of the task. The first 45 items of the phonemic
awareness test developed for this study replicate a number of the tasks in the
Yopp(1998) analysis. The tasks in this section of the test will be referred to
as the phonemic discrimination tasks.
Because the subjects in Yopp’s study (mean age
5.83 years) were considerably younger than the subjects in the present study
(mean age 7.63 years), and because the children in this study were well
advanced in learning to read, it was decided to add an element of complexity to
the phonemic awareness test by requiring the children to apply their phonemic
knowledge to decoding nonsense words. The tasks in this section of the test are
referred to as the blending tasks.
3.10.1 Item statistics for the phonemic awareness test
Item statistics were computed using the
ItemAn item test and analysis program for Windows by Assessment Systems
Corporation. The test data obtained
from the administration of the phonemic awareness test to the First Class
children was returned to the researcher on the Phonemic Awareness Test Record
Sheet and was entered on a spreadsheet. It was then copied to the ItemAn
program for analysis.
3.10.2 Single scale analysis
The item analysis was approached in four stages. The initial approach to the
item analysis was to place all of the items on a single scale of 65 items. The
analysis yielded a point-biserial score and an item discrimination score for
each item on the scale. The point-biserial score is a Pearson’s correlation
coefficient that indicates how well an item discriminates between high
performing and low performing students. The normal range of biserial scores for
items is between 0 and 1, a 0 score indicating that this item has no
discriminative quality, and 1 indicating that the item discriminates well
between high and low performing students. The recommended minimum value for
point-biserial scores for a reliable test is above .20. Four of the items on
the scale had scores slightly below .20 and a further item had a score of .05,
indicating that it discriminated poorly between high performers and low
performers. The latter item was a very easy item which was correctly answered
by almost all of the children
A summary of the item statistics for the
single-scale analysis is presented in Table 3.3. The mean score of the students
on the test was 41.51 and SD was 11.8. For the purpose of establishing the
reliability of the test, the Alpha score is very important. Alpha is a
parameter that describes how well a group of items focuses on a particular
construct. It provides an estimate of the internal consistency or reliability
with which the test measures a construct. The complete results of the item
analysis for the 65-item scale are presented in Appendix D.
Table 3.3: Scales statistics for the 65-item scale
|
N of Items |
65 |
|
N of Examinees |
109 |
|
Mean |
41.514 |
|
Std. Dev. |
11.795 |
|
Minimum |
9.000 |
|
Maximum |
64.000 |
|
Median |
41.000 |
|
Alpha |
0.928 |
|
Mean Point Biserial |
0.566 |
Carefully developed standardised achievement
tests usually have reliability estimates around .90. On the 65-item scale,
an Alpha score of .92 for this test was obtained. Hills (1981) recommends that
a test should achieve an Alpha score of greater than .90 if it is to be used as
a basis for making significant decisions for educational provision for
individual students. Prima facie, the
high Alpha score would suggest that the phonemic awareness development test is
a highly reliable test as it is greater than .90 as Hills recommends. However,
the Alpha score assumes that a test is
a test of single construct. It could be argued that the phonemic awareness
development test assesses two separate constructs, phonemic awareness and
phonemic blending. Therefore, a further item analysis of the test data was
carried out using two separate items scales, phonemic discrimination and
phonemic blending.
3.10.3
Analysis of separate phonemic discrimination and phonemic blending
subscales
The phonemic discrimination tasks
comprised eight subtests with 5 items
in each subtest. Table 3.4(a) shows an
analysis of the mean scores of the 109 First Class pupils on each of the 8
phonemic discrimination subtests.
Table 3.4.(a): Mean scores of First Class pupils (N=109) on the 8
phonemic discrimination sub-tests.
|
Pattern test |
Onset test |
Missing sound
test |
Onset deletion |
Odd-one-out
rhyme identification test |
Provide a
rhyming word for target word |
Syllable test |
Phoneme
identification test |
|
4.07 |
4.57 |
3.48 |
3.73 |
3.91 |
4.39 |
3.56 |
2.88 |
The scores indicate that the First Class
pupils obtained the highest score on the onset test, where the children had to
say the first sound they heard in a given word, such as /ch/ for the word chair. Pupils
also found the task of providing a rhyming word for a target word relatively
easy. The relatively high scores achieved on the pattern, onset and rhyming
subtests suggests that there may have been a ceiling effect for these subtests.
The two tasks which pupils found most difficult were the missing sound test
(for example, hearing the words chair
and air and saying what is missing in
the second word) and identifying the individual phonemes in a word such as nose or friend. The overall mean score on the phonemic discrimination items
was 4.0, indicating that this section of the test was relatively easy for the
First Class pupils.
The mean scores on the 5 phonemic blending
subtest are presented in Table 3.4 (b). A comparison with the mean scores on the
discrimination tasks reveals that children found the blending tasks
considerably more difficult.
Table 3.4 (b): Mean scores of First Class pupils (N=109) on the
5 phonemic blending sub-tests.
|
short vowel |
initial cons blend |
cons digraph |
long vowel |
two-syllable |
|
3.10 |
2.50 |
1.65 |
1.54 |
2.13 |
The overall mean score on
the discrimination tasks for the 109 first class children was 4.0 whereas the
mean score on the blending tasks was 2.0. Children found three-letter cvc words easiest while they had
greatest difficulty decoding the long-vowel nonsense words, such as sile, nare and bute. All the long vowel nonsense words used in the test can be
decoded by analogy with real word rimes, for example, s-ile/sm-ile, n-are/d-are, b-ute/c-ute.
In the second stage of the item analysis,
the sixty-five items were separated into two-scales, a 40-item phoneme
discrimination scale and a 25-item phoneme blending scale. The results of the item analysis of the two
scales are presented in Table 3. 5.
Table 3.5: Item analysis of the phoneme discrimination subscale and phoneme blending subscale
|
Scale: |
phoneme discrimination |
phoneme blending |
N of Items |
40 |
25 |
|
N of
Examinees |
109 |
109 |
|
Mean |
30.596 |
10.917 |
|
Std. Dev. |
6.310 |
6.878 |
|
Minimum |
9.000 |
0.000 |
|
Maximum |
40.000 |
24.000 |
|
Median |
32.000 |
10.000 |
|
Alpha |
0.861 |
0.918 |
|
Mean Point Biserial |
0.569 |
0.751 |
In this representation of the analysis, the
Alpha score of the phoneme discrimination scale does not quite reach the level
of reliability recommended if it is to be used to make significant educational
decisions about individual pupils. Nonetheless, the Alpha score of .86 suggests
that is a reasonably reliable test. Also the mean biserial-point score for both
tests indicates that the individual items discriminate well between high and
low scorers on the test, especially in the case of the phoneme blending test
where the mean biserial-point score is .751. Interestingly, the mean biserial
point scores for all but 2 of the 25 individual items in this section of the
test were above .50.
3.10.4 Correlation of reading scale with phonemic
awareness development scale and subscales
Correlations were calculated between the reading achievement scale
and the phonemic awareness development scale and subscales. The results are
presented in Table 3.6. The correlation
of the phoneme discrimination with phoneme blending scale was calculated at
.599. This moderate correlation would tend to support the hypothesis that the
scales are measuring separate but related constructs. As would be expected, the whole scale, phonemic awareness
development, is very highly correlated with the subscales. More interesting is
the relationship between the reading scale and the phonemic awareness
development scale and subscales. The correlation with the reading achievement
scale was 0.674 while the correlations with both phonemic blending and phonemic
discrimination were slightly lower at 0.636 and 0.566, respectively. All
correlations were found to be significant at the <.01 level.
Table 3.6: Correlations (Pearson) of phonemic awareness development scale and subscales with reading
|
|
Phonemic
Discrimination |
Phonemic
Blending |
Phonemic
Awareness Development |
|
Phonemic
blending |
0.599 |
|
|
|
Phonemic
awareness development |
0.884 |
0.904 |
|
|
Reading
percentile |
0.566 |
0.636 |
0.674 |
3.10.5 Nine subscale analysis
The next phase of the item analysis was carried
out by placing the 40 phoneme discrimination items on eight separate subscales
with 5 items in each scales and by putting the 25 phoneme blending items on a
singles 25-item scale. The results of this analysis are presented in Table 3.7.
Table 3.7: Item analysis of eight phoneme discrimination subscales *
|
|
Pattern test |
Onset test
|
Missing sound test |
Onset deletion |
Odd-one-out rhyme
identification test |
Provide a rhyming word for
target word |
Syllable test |
Phoneme identification
test |
|
N of Items |
5 |
5 |
5 |
5 |
5 |
5 |
5 |
5 |
|
N of First
Class pupils |
109 |
109 |
109 |
109 |
109 |
109 |
109 |
109 |
|
Mean |
4.073 |
4.569 |
3.477 |
3.734 |
3.908 |
4.394 |
3.560 |
2.881 |
|
Std. Dev. |
1.412 |
1.078 |
1.268 |
1.524 |
1.296 |
1.084 |
1.404 |
1.607 |
|
Alpha |
0.783 |
0.828 |
0.591 |
0.753 |
0.644 |
0.700 |
0.612 |
0.691 |
|
SEM |
0.658 |
0.448 |
0.810 |
0.758 |
0.774 |
0.594 |
0.874 |
0.893 |
|
Mean P |
0.815 |
0.914 |
0.695 |
0.747 |
0.782 |
0.879 |
0.712 |
0.576 |
|
Mean Item-Tot. |
0.736 |
0.764 |
0.610 |
0.707 |
0.648 |
0.673 |
0.629 |
0.668 |
|
Mean
Point-Biserial |
0.983 |
1.000 |
0.870 |
0.959 |
0.922 |
0.966 |
0.842 |
0.856 |
*Please note that this item
analysis included the 25-item phoneme blending scale which is shown in Table
3.5.
As might be expected, the Alpha scores are
considerably lower on the five-item scale that they were on either the 65-item
or the 40-items scales discussed earlier. However, in four of the subtests, the
Alpha score is above .70, which indicates a moderately high level of
reliability. It is interesting to note that the mean point-biserial scores on
the subscales are very high, suggesting a high level of internal consistency in
the test.
3.10.6 Six subscale analysis
The final stage of the item analysis
involved placing all of the phonemic discrimination items on a single scale and
the 25 items in the blending section of the test on 5 separate subscales. A
summary of the item analysis using this approach is presented in Table 3.8. Compared to the mean scores obtained by
children on the phonemic discrimination items, the mean scores on these subtest
are quite low. Also, the anticipated order of difficulty of the items
anticipated when the test was developed did not materialise. The subtest which
was found to be most difficult was the long vowel subtest and the easiest was
the short vowel subtest. The two-syllable nonsense words were found to be less
difficult than either the consonant digraphs or the long vowel nonsense words. Alpha scores for all of the subtests were
consistently above .70, indicating a moderately high level of reliability for
the individual subtests. The mean item-tot. represents the average
point-biserial correlation across all items of the scale. All mean item-tot.
scores on this scale are in the region of .70, with the exception of the
phoneme discrimination mean item-tot. score.
Table 3.8: Item analysis of the phonemic discrimination scale and five phonemic blending subscales
|
Scale: |
Phoneme
discrim. |
Short
vowel |
Cons blends |
Cons digraphs |
Long vowels |
Two syll |
N of Items |
40 |
5 |
5 |
5 |
5 |
5 |
|
N of
Examinees |
109 |
109 |
109 |
109 |
109 |
109 |
|
Mean |
30.596 |
3.101 |
2.495 |
1.651 |
1.541 |
2.128 |
|
Std. Dev. |
6.310 |
1.597 |
1.759 |
1.605 |
1.611 |
1.751 |
|
Minimum |
9.000 |
0.000 |
0.000 |
0.000 |
0.000 |
0.000 |
|
Maximum |
40.000 |
5.000 |
5.000 |
5.000 |
5.000 |
5.000 |
|
Median |
32.000 |
3.000 |
2.000 |
1.000 |
1.000 |
2.000 |
|
Alpha |
0.861 |
0.703 |
0.752 |
0.724 |
0.749 |
0.780 |
|
SEM |
2.355 |
0.871 |
0.876 |
0.843 |
0.808 |
0.821 |
|
Mean P |
0.765 |
0.620 |
0.499 |
0.330 |
0.308 |
0.426 |
|
Mean Item-Tot. |
0.392 |
0.676 |
0.709 |
0.692 |
0.707 |
0.728 |
|
Mean Point Biserial |
0.569 |
0.877 |
0.892 |
0.901 |
0.932 |
0.927 |
3.10.7 Correlation of subscales
Table 3.9 presents the correlation of the
phoneme discrimination subscales with each other and with the phoneme blending
scale. The correlations are positive in
all but one case, where the correlation between the syllable subscale and the
onset subscale produced a very low negative correlation of -.083. The correlation between the onset subscale
and other subscales are generally low and indicates that the onset subtest may
be measuring a qualitatively different trait, compared to the other subtests.
Table 3.9: Correlation matrix of phoneme discrimination subscales and the phoneme blending scale
|
|
Pattern test |
Onset test
|
Missing sound test |
Onset deletion |
Odd-one-out rhyme
identification test |
Provide a rhyming word for
target word |
Syllable test |
Phoneme identification
test |
Phoneme
Blending |
|
Pattern test |
1 |
0.069 |
0.242 |
0.435 |
0.239 |
0.233 |
0.340 |
0.307 |
0.449 |
|
Onset test |
|
1 |
0.446 |
0.020 |
0.011 |
0.138 |
-0.083 |
0.230 |
0.118 |
|
Missing sound
test |
|
|
1 |
0.308 |
0.189 |
0.150 |
0.200 |
0.276 |
0.320 |
|
Onset deletion |
|
|
|
1 |
0.331 |
0.325 |
0.400 |
0.305 |
0.568 |
|
Odd-one-out rhyme
identification test |
|
|
|
|
1 |
0.470 |
0.442 |
0.122 |
0.341 |
|
Provide a
rhyming word for target word |
|
|
|
|
|
1 |
0.373 |
0.153 |
0.246 |
|
Syllable test |
|
|
|
|
|
|
1 |
0.245 |
0.351 |
|
Phoneme identification
test |
|
|
|
|
|
|
|
1 |
0.339 |
|
Phoneme
Blending |
|
|
|
|
|
|
|
|
1 |
3.11 Summary of the result of the analysis of the data for first class pupils
The phonemic awareness development test
proved to be highly reliable overall, with good psychometric properties of
reliability and internal consistency. The phonemic discrimination section of
the test was somewhat easy for the pupils in the sample, with a ceiling effect
for some items. The blending section of the test was highly reliable and
discriminated well between high and low achievers. It can therefore be
concluded that the phonemic awareness development test could be used to
accurately measure the phonemic awareness development of children with a
specific learning difficulty in the main phase of the study.
3.12 Assessment of children with specific learning disabilities at enrolment
The children selected for enrolment in special SLD provision in September, 1999 were assessed by staff from the special schools and support units early in June 1999. The June assessment was already standard procedure in a number of the schools and all but one of the schools undertook the assessment in advance of the enrolment of the children. The remaining support unit assessed the pupils immediately on enrolment in September.
The schools were asked to administer the phonemic awareness development test and the Neale Analysis of Reading Ability. Schools also collected a sample of the children’s free writing, although this is not included in the analysis. The phonemic awareness test that was administered to the 109 First Class children was also administered to the SLD pupils at the initial assessment. Pupils’ responses to the phonemic awareness items were recorded on the Phonemic Awareness Test Record Sheet (Appendix C) The reading percentile score achieved on the Neale Analysis of Reading Ability was also reported on the Phonemic Awareness Test Record Sheet, as was the date of birth of the pupil, their sex and the class in which they would be enrolled if they were in a mainstream school. Pupils with SLD were tested by teachers in each of the special schools and support units and the data were recorded on Phonemic Awareness Test Record Sheet. (A copy of the record sheet is available in Appendix C.) These record sheets were returned to the researcher where they were input into an Excel Spreadsheet in order to facilitate subsequent analysis.
3.13 Assignment of pupils to treatments groups, OM and US
The hypothesis for the present study stated
that children using computer software would develop phonemic awareness more
effectively using an approach that involved the use of computer software than
if they were taught phonemic awareness through other methods alone. In order to
test this hypothesis, children with SLD enrolling in special provision in
September, 1999 were assigned to one of two treatment groups. In the ‘Units of
Sound’ treatment group, children used the ‘Units of Sound’ programme for
approximately 15 minutes per day, while in the ‘other methods’ treatment group,
children were taught phonemic awareness directly by the teacher. Reinforcement
activities such as games and workbook exercises were used to supplement direct
instruction in the case of children in the OM group.
3.14 ‘Units of Sound’ treatment
‘Units of Sound’ includes an initial screening test that
determines where children should start the instructional section of the
programme. Children were placed in the
programme on the basis of their initial assessment and were assigned to work on
the programme for approximately 15 minutes daily from September 1999 to the end
of February, 2000. There was slight variation in the commencement date, but all
children in the ‘Units of Sound’ treatment group (US) had commenced the
programme before the end of September.
Each school had a daily period of
individualised instruction for the SLD children. During this period, children
were assigned written work, or read aloud to the teacher, or worked in pairs or
small groups on assignments. It was usually during this period that the US
group worked on the ‘Units of Sound’ program. In the vast majority of cases,
the US pupils remained in the classroom under the supervision of the teacher.
However, in one school, a classroom assistant supervised the pupils in the US group while they worked
on the programme in a computer lab adjacent to the classroom.
The number of pupils in any given classroom
assigned to the US group ranged between 2 and 11 pupils. Where the number of
pupils in the classroom was toward the upper end of this range, it was not
always possible to allow all of the pupils in the US treatment to work on the
programme every day. However, almost all of the US group had access to the
programme on at least four days per week.
3.15 ‘Other Methods’ treatment
Pupils who were assigned to the ‘other
methods’ treatment (OM) were taught phonemic awareness using approaches and
material that were already in use in the school. In all cases, schools follow a
structured programme in phonemic awareness, so that phonemic awareness is
explicitly taught to all pupils. Some components of the programme in phonemic
awareness are supported by resources developed by the teachers, but in the
main, teachers use commercially-produced materials.
Teaching of phonemic awareness in the OM
treatment involved using a variety of methods and approaches. The structured
programme in phonemic awareness follows a sequence where children are taught
consonant sounds, followed by consonant blends and short vowel sounds in the
initial and medial position. Usually, the phoneme sounds are taught directly by
the teacher for about 5 minutes daily. Direct instruction is frequently
complemented by a related reinforcement activity using material devised by the teachers
or using commercially produced materials. A variety of material is used for
reinforcement activities, some of which was produced to accompany reading
schemes, such as The Oxford Reading Tree (Oxford University Press) and
other material was designed as part of an independent phonemic awareness
programme, such as Alpha to Omega (Hornsby & Pool, 1987).
It should be noted that phonemic awareness
activities are not confined to a single period during the day in the SLD
special schools and support units. In accordance with the principles of the
primary schools curriculum, the schools
adopt an integrated approach to teaching and learning and thus avail of
incidental opportunities for developing phonemic awareness whenever the
opportunity is presented. Therefore, children in the US group also engaged to
some extent in the other activities to develop phonemic awareness, as well as
the ‘Units of Sound’ treatment. In the context of the experiment, some
contamination of the US group by the OM treatment was therefore unavoidable.
3.16 Data collection and analysis
Teachers in the special
schools and support units for pupils with specific learning disabilities
collected data on the pupils who enrolled in September, 1999. In all but one
case, data from the pre-treatment tests were collected in June, 1999. In the
remaining school, pre-treatment data were collected in early September, 1999.
These data were returned to the researcher where they were recorded in a
spreadsheet. Treatments for both the OM and US groups commenced in all cases in
early September 1999. Following a six-month period, the principals were asked
to administer the post-treatment tests as soon as possible after March 1st,
2000. Data were again returned to the researcher where they were input in Excel.
The data were analysed using a variety of
statistical software packages, depending on the output required. Simple
correlations and mean scores were calculated using Excel. Datadesk and Minitab
were use mainly when descriptive statistics, t-tests and regression analyses
were required. All three software packages were used to produce graphs.
Chapter 4: Results
4.1 Pre-treatment SLD group
Pre-treatment and post-treatment data were obtained on the 140
pupils selected for enrolment in special schools and support units for children
with specific learning disabilities. Descriptive statistics for the subjects in
the study are presented in Table 4.1. Of the 140 pupils, 99 were boys and 41
were girls. The proportion of girls enrolled was higher than usual in 1999, as
normally there are more than three boys enrolled for every girl.
Table 4.1: Descriptive statistics for SLD children at the pre-treatment stage
Variable |
Sex |
N |
Mean |
Median |
StDev |
Range |
|
CA |
Boys |
99 |
10.128 |
10.200 |
1.010 |
8.200 - 12.200 |
|
|
Girls |
41 |
9.741 |
9.500 |
1.133 |
7.900 - 12.200 |
|
Read. Percentile |
Boys |
99 |
7.444 |
4.000 |
7.732 |
1.000 - 30.000 |
|
|
Girls |
41 |
8.44 |
5.00 |
8.81 |
1.00 - 34.00 |
The mean age of the girls
(9.7yrs.) was slightly lower than the
mean age of the boys (10.1yrs.). Boys ranged in age from 8.2yrs. to 12.2yrs., a
range of exactly 4 years, but the standard deviation of just over one year in
the case of both boys and girls suggests that approximately two-thirds of
children were aged between 9.0yrs. and 11.0yrs. The mean reading percentile of
the girls (8.4) was slightly higher than that of the boys (7.4). The range in
the reading percentile scores was considerably higher than might be expected
based on the enrolment criteria, but it is important to note that more than
half of the children achieved a reading percentile of 5 or less, because the
median reading score for boys is 4.0, and for girls the median score is 5.0.
4.2 Comparison between the phonemic awareness development of the First Class children and the pre-treatment group
The phonemic awareness development test was administered to the pre-treatment group prior to, or immediately upon their enrolment in the special schools and support units. A two-sample t-test was used to determine whether there was a significant difference between the average scores achieved by the First Class group and the pre-treatment SLD group. The results of the t-test (Table 4.2) revealed that there was a significant difference in favour of the First Class group (t = 3.20, p=.0016, df = 208).
Table 4.2: T-test of
phonemic awareness development of the First Class and SLD pre-treatment
groups
|
|
N |
Mean |
St. Dev |
|
SLD
Pre-treatment |
140 |
37.01 |
9.86 |
|
First Class |
109 |
41.5 |
11.8 |
In order to pinpoint the subtests which contributed to the differences in the mean scores of the two groups, a comparison was made between the mean scores of the two groups on each of the subtests
Table 4.3 (a): Mean scores of First Class pupils (N=109) and SLD
pre-treatment group (N=140) on the 8
phonemic discrimination sub-tests.
|
|
Pattern test |
Onset test |
Missing sound
test |
Onset deletion |
Odd-one-out
rhyme identification test |
Provide a
rhyming word for target word |
Syllable test |
Phoneme
identification test |
|
First Class |
4.07 |
4.57 |
3.48 |
3.73 |
3.91 |
4.39 |
3.56 |
2.88 |
|
SLD pre-treat. |
3.91 |
4.82 |
3.63 |
3.14 |
3.70 |
4.20 |
3.86 |
2.97 |
The results of this comparison are presented in Table 4.