Deprecated: mysql_pconnect(): The mysql extension is deprecated and will be removed in the future: use mysqli or PDO instead in /home/rrforg/public_html/Connections/rrf.php on line 9
Reading Reform Foundation (RRF) - Promoting Synthetic Phonics : Page Title

RRF Newsletter 47 back to contents
Do we have a literacy crisis?Byron Harrison

In Australia this question has been almost impossible to answer because we have no independent national inspectorate. Whole Language, ‘Top-Down’ philosophy thereby controls both the training of teachers and the testing of literacy outcomes that result from that training.  

Tests are designed to fit within Top-Down philosophic limitations and are largely limited to higher order, context-driven processing such as cloze procedures and comprehension.

We present evidence showing that such tests do not reveal sub-skill deficits such as:

·        knowledge of sounds,

·        confusions between names and sounds,

·        capacity to blend sounds into syllables,

·        capacity to read 1,2 and 3 syllabic words,

·        capacity to proof read,

·        confusions between the shapes and sounds of the letters b, d, p etc.

Survey population

The raw data in this paper was therefore drawn from an independent source, an optometric practice conducting literacy research . This paper is part of a larger study of 3000 consecutive children. The children were mainly resident in or around Hobart, the beautiful capital of Tasmania in Australia.

Ten years ago, when Whole Language reigned supreme, Tasmania claimed to be at the leading edge of Australian education. Since that time there has been a gradual retreat from an overt anti-phonics sentiments to the current ‘but-we-teach-everything’ policy position.

However since California’s 1998 dismissal of Whole Language as ‘unscientific’, Tasmania has seen some introduction of Synthetic Phonics but its introduction is voluntary and its application piecemeal. We still employ teachers who believe phonics is a four letter word, many infant teachers have not received additional phonic training and the bureaucrats and teacher training institutions, responsible for more than 20 years of Top-Down teaching, are still drawing salary and still training teachers.

The study

The full 3000 child study (in preparation) will report on 25 different aspects of literacy ranging from the child’s capacity to simply point to matching letters, to examining a child’s responses when reading at speed, through to a child’s capacity to detect when a word’s spelling looks correct.

This interim paper only reports on five of those twenty five aspects, those which involve the children’s capacity to read phonetically-regular 1,2 and 3 syllabic words.


The practice sees a disproportionate number of struggling children. This paper therefore confines itself to the 487 children who had been lead to believe that they were ‘average’ readers.

The children were first asked to categorise themselves as:

1.      ‘average’ readers or

2.      ‘below average’ readers or

3.      ‘above average’ readers.

That self-assessment was then independently confirmed or denied by their parents acting on the basis of teacher interviews, school reports and often by parental observation in the classroom. In the few cases where there was disagreement the data was excluded from the study.

·        Selected words appeared on the computer screen.

·        The child was given unlimited time to read the words.

·        The words appeared without pictures or context.

·        If the child misread a word, a second and similarly structured word was immediately displayed to confirm the error pattern.

·        The response was only recorded as an error if both words were misread.








UNLIMITED TIME ALLOWED: (only multiple errors recorded)





1 syllable


2 syllables


3 syllables













































Column 1: Confusions between letter names and sounds

The three letters c, g and y have multiple sounds. This table therefore only reports on the percentage of children who made four or more errors. If a child confused the name and sound of a vowel, they were told ‘Yes that is the name but what sound does it make?’ This data only records children, who despite that prompting, clearly had no knowledge of short vowel sounds.

The fact that 34% of 6 year olds confuse letter names and sounds is understandable and perhaps inevitable in a situation where infant schools fail to advise preschool parents and teachers to teach sounds not names. What is of more concern is the finding that at the age of 9, after three years of compulsory education, 30% of children still hadn’t mastered letter sounds, arguably the most basic phonic skill.

Given that children’s reading habits, both good and bad, have usually been habituated by the age of 9, it is alarming but not surprising to find evidence of  confusions between names and sounds in 31% of children entering high school.

Column 2: Misreads 1 syllabic words

The four most common causes of misreading 3 letter words were:

1.      guessing errors (paying attention to some letters but misguessing others. e.g. ‘pet’ misread as ‘pot’). At the age of nine, 66% of all errors were due to misguessing[1].

2.      confusions between the names and sounds of vowels (e.g. ‘net’ misread as ‘neat’)

3.      b,d,p confusions (e.g. ‘big’ misread as ‘dig’).

4.      Word reversals (e.g. ‘was’ misread as ‘saw’).

Our data indicates that whilst ‘was/saw’ word reversals usually disappear by the age of 10, the ‘big/dig’ letter reversals and the ‘net/neat’ name/sound confusions often persist throughout primary school undermining both self confidence and phonic skills. The ‘pet/pot’ guessing errors persist longer still.

Column 3: Misreads 2 syllabic words. (e.g. ‘picnic’ misread as ‘picture’)

The fact that readers make fewer errors on two syllabic words than when reading either one or three syllabic words may seem to be an odd finding. In fact it is precisely what we predicted[2] should happen whenever whole word-guessing becomes the dominant mode of word attack.

Column 3: misreads 3 syllabic words. (e.g. ‘cormorant’ misread as ‘computer’)

Our team explained, in papers and lectures as far back as 1988, why whole-word guessing has an upper limit of about 7 letter words. We actually predicted that about 75% of children entering high school from a primary school with a whole word emphasis should in theory have difficulty with long particularly unfamiliar three and four syllabic words. Our finding reported here that 72% of ‘average’ readers experienced these difficulties merely confirms those predictions. The figures for poor readers are of course even worse. The word length limitation is even observed in some readers who believed they were ‘above average’ readers.

Contrast with official figures

Last year the Australian public was reassured that ‘only’ about 19% of grade 3 (age 9) children failed to meet the national standards. Such a finding is patently incompatible with the above data which shows that 39% of average nine year olds misread 3 letter words, 6% misread 2 syllabic words and 87% were unable to read longer words.

We do not believe that these Australian results are unusual; indeed these are the same error patterns that we observed during our lecture tours in the UK and New Zealand. By testing within context, these underlying subskill problems go undetected and official figures thereby overestimate the skill levels of children.

We conclude that:

a)      basic phonic skills are still not being effectively taught in infant grades;

b)      if good phonic skills are not established in the infant schools, the resultant guess-induced error patterns often persist right into high school;

c)      this population is unlikely to receive scarce remedial resources because they are ‘average’ readers.

All of these outcomes and more were predicted by Visual Attention Span (VAS) Theory back in 1988. VAS Theory is discussed further on the RRF website. Teachers wishing to access the diagnostic methods are referred to our team’s website

VAS Theory

A child learning to read using Synthetic Phonics is learning to pay attention to every letter, to process letters in a left-to-right direction and to connect each letter to its sound.

The whole word processor however is learning to select those letters that attract most attention. The infant then tries to make a match with a word that they know. There is no necessary left-to-right processing going on; they are learning to process words like a pictures. It is fast, inherently inaccurate and weak in sequencing.

The letters that attract most attention are called High Visibility Letters and they are:

·        the two end letters

·        any letters with limbs intruding into the space above and below the word.

Consider the word ‘magnet’ as an example:

The word ‘magnet’ therefore has three High Visibility Letters, m_g__t , which attract visual attention  If the word-guessing child can simultaneously hold all three of these high visibility letters letters in memory, then m_g__t  becomes the basis of their guess.

Unfortunately that visual pattern fits more than one word (e.g. maggot, midget, magnet etc) which is why whole word processing is inherently inaccurate.

But some infants have not developed the capacity to hold all three High Visibility Letters in working memory. Their guess may therefore be based on only two of the three letters m____t and there are now over forty words that fit..

There is thus an intimate relationship between the number of letters that a child can hold in working memory and their capacity to accurately word-guess. The number of letters that a child can hold in memory is called their Visual Attention Span level (hereafter abbreviated to VAS). Teachers can determine a child’s VAS level (and thereby their capacity to word-guess) in a few seconds simply by flashing symbols[3] on screen and asking the infant what they saw.

VAS level 3 appears to be the minimum level of VAS consistent with some success in whole word guessing. Teachers should be horrified to learn that our current study of over 3000 children demonstrates that, at the age of 7 years 7 months, 50% of infants had still not developed this level of VAS and were therefore deemed inappropriate candidates for the whole word guessing and predictive cueing strategies taught in their schools.

The problem is not limited to children with low VAS memory:

The high VAS child

The high VAS child develops visual memory very early. They may have developed the level 3 VAS memory, the level necessary for whole word guessing, even before they can read. Guessing therefore comes easy and early. Such children they are often held up as proof that whole word processing works. A very different picture emerges if the progress of these children is followed in subsequent years.

Their high VAS level leads to initial success both in recognising words and in predicting from context. The child therefore acquires a fast word-guessing strategy. They therefore have little need to learn a slower phonic approach and thereby become guess-dependent.

In our lectures we have called these high VAS children/low phonic children, the ‘Brick Wall Kids’ because their initial success due to guessing with a high VAS hits a brick wall and begins to fail about the age of nine or ten. The ‘brick wall’ is caused by the fact the each year the words are getting longer and guesssing has an upper limit of about 7 letter words.

As a rule of thumb, if you add 2 to a child’s VAS level, you can determine the word length that may be guessable.[4]. A child with a VAS level of 3 therefore tends to struggle with words longer than five letters (VAS 3 + 2).

The problem is that even adults seldom exceed level 5 VAS. VAS Theory therefore predicts that guessing -dependence should lead to difficulties with words longer than VAS 5 + 2= 7 letter words.

In 1988 we put our reputations on the line and predicted that if children were encouraged to word-guess in infant grades, about 75% of them, on entering high school, should display difficulties in reading unfamiliar, phonetically regular words such as ‘Eromanga’ or ‘Consonant’. Eight years later we were saddened to find our predictions confirmed.

Phonics and whole word guessing strategies should therefore be viewed as competitive rather than cooperative strategies at the ‘learning to read’ stage of development. If teachers attempt to teach both strategies, they should expect to find that many children will choose the faster, inaccurate, unsequenced, context-dependent, whole word guessing alternative. This is one reason why massive phonic deficits are showing up in schools which protest that they do teach phonics. The other reason for such phonic failure of course is that these schools are usually teaching Analytic Phonics rather than Synthetic Phonics.

Consider the graph below:

The left hand scale denotes the VAS level, the bottom scale the age.

The top curved line shows the typical development of VAS in an average reader.

The heavier horizontal line is level 3 VAS, the minimum level necessary for whole word processing.

Case histories

Consider the case of John, an 8 year old with a low VAS level of 2.6

There are 3 steps to develop John’s VAS graph:

1.      draw a vertical line upwards from John’s age (8) on the bottom line and

2.      draw a horizontal line starting from John’s VAS level (2.6) on the left hand side. The point where those two lines intersect is John’s current VAS level. It is marked ‘A’.

3.      draw a line passing through ‘A’ and parallel with the ‘Average VAS’ development line’. (This is the A-B-C line)

What does the chart tell you?

The first thing that you look for is the Guessing Age, the age when VAS level 3 is reached (marked ‘B’ on the graph). In John’s case we can see that he won’t reach this age until he is 9 years 7 months.

Immediately the alarm bells should sound because that data indicates that if John has been encouraged to rely on whole word guessing and predictive cueing between the age of 6 and his current age of 8, then he will almost certainly have failed.

You should pray that John knows how to blend sounds and syllables together. If John does NOT have control over his phonic skills then this graph tells you that:

1.      John has to be a non-reader, is ‘at-risk’ and needs urgent remedial attention.

2.      He probably has developed a poor attitude towards learning as a result of having experiencing two years of failure.

3.      His stock of sight words should be low.

4.      Proof  reading and spelling[5] should be poor.

You know all this without asking the child to even read a page!

Case history No. 2

Let us consider another case. (This is in fact John again but in this example you are meeting him for the first time at the age of twelve).

John is just starting high school and you are his new teacher. It has taken you less than 2 minutes to measure his VAS level and his capacity to blend sounds into syllables. You find that he has a VAS level of 3.2. You also find that he is not confident in blending sounds and syllables together mainly because he has confusions between letter names and letter sounds.

You carry out the same 3 steps of

1.      draw a vertical line upwards from John’s age (12) on the bottom line and

2.      draw a horizontal line starting from John’s VAS level (3.2) on the left hand side. The point where those two lines intersect is John’s current VAS level. It is marked ‘C’.

3.      draw a line passing through ‘C’ and parallel with the ‘Average VAS’ development line’.

What does the graph now tell you?

It tells you that this child did not reach level 3 VAS until the age of 9 years 7 months.

The fact that he still has confusions between names and sounds at the age of 12, tells you that his phonic skills have never been reliable.

Combining those two bits of information tells you that he must have been a non-reader until the age of 9 years 7 months:

a)      because his phonics were undermined by the name/sound confusions and

b)      because his low VAS level prevented him from using whole word processing.

The predictions would not stop there:

you would expect attitudinal changes to have developed during those 6 years of failure; that damage may never be undone;

he may have developed strategies of overusing context, of relying on pictures, of making random guesses and none of these will have made him a good reader;

those bad habits may have made him resistant to remediation and his low self-esteem may have lead him to form associations with other struggling children, adopting their poor attitudes towards authority, slovenly speech and dress code (we all need friends).

You therefore shouldn’t be surprised if this child exhibited behavioral problems at home and/or in school.

The two lessons arising from this are:

1.      that the low VAS would have identified this child as being at risk in less that a minute as early as age 6 and

2.      had Synthetic Phonics been introduced from the start, the low VAS would have been relatively unimportant because a low VAS handicaps guessing skills more than phonic skills (although sight words and proof reading may have needed some extra work).


·        The time has come for teachers to reclaim their profession from those who for almost four decades substituted dogma for data, rhetoric for reason.

·        The new tools are now available on the Internet which means that for the first time parents will have direct access to detailed and independent information about their children’s reading deficits.

·        Parents and teachers will be able to generate on-line reports and may then begin to openly question both the time spent in preparing our current end-of-term reports and the reliability of the information being reported to parents.

The USA has already identified the need for reform. Can the rest of the English-speaking world afford to do less?


Byron Harrison

VAS Research

ibm building

5th floor, 147 Macquarie St.,

Hobart, TAS. 7005, Australia

[1]               Whole Language philosophy asserts that such habits should disappear but  VAS Theory (see later) predicts that once guessing habits are established, they should tend to persist. The data supports VAS Theory for we find that at the age of twelve , 44% of errors were still due to misguessing. (see column 5).

[2]               The Hole In Whole Language: Harrison, Zollner and Magill. Aust. J. Rem. Ed. Vol.27. No. 5 1996

[3]               We use carefully selected and sequenced numbers to ensure that the child is processing every number individually. We record how many of the flashed numbers the child can recall. Numbers enable us to start testing VAS at a very early age, as soon as the child can match symbol to word.

[4]               It varies a little because some words have a unique shape and are therefore easier to guess. In other cases the context cues may be so strong that guessing becomes easy The low VAS children often cannot use context cues because there are simply too many unreadable words in sentences to develop any meaning.

[5]               If John misspells ‘laid’ as ‘laed’, the misspelled word ‘laed’ will look perfectly OK (because when John learned to pay attention to the end letters l__d, he will have developed inattention to mid-word letters and their sequence).




Copyright Notice
All rights, including copyright, in the content of these RRF web pages are owned or controlled for these purposes by the RRF. In accessing the RRF's web pages, you agree that you may only download the content for your own personal non-commercial use. You are not permitted to copy, broadcast, download, store (in any medium), transmit, show or play in public, adapt or change in any way the content of these RRF web pages for any other purpose whatsoever without the prior written permission of the RRF.
© Reading Reform Foundation 2010
Home  |   RRF Conferences  |   Resources / Articles  |   Newsletter Archive  |   About Us  |   Contact  |   Donate

Sites for Teachers