is systematic synthetic phonics generating neuromyths?

A recent Twitter discussion about systematic synthetic phonics (SSP) was sparked by a note to parents of children in a reception class, advising them what to do if their children got stuck on a word when reading. The first suggestion was “encourage them to sound out unfamiliar words in units of sound (e.g. ch/sh/ai/ea) and to try to blend them”. If that failed “can they use the pictures for any clues?” Two other strategies followed. The ensuing discussion began by questioning the wisdom of using pictures for clues and then went off at many tangents – not uncommon in conversations about SSP.

SSP proponents are, rightly, keen on evidence. The body of evidence supporting SSP is convincing but it’s not the easiest to locate; much of the research predates the internet by decades or is behind a paywall. References are often to books, magazine articles or anecdote; not to be discounted, but not what usually passes for research. As a consequence it’s quite a challenge to build up an overview of the evidence for SSP that’s free of speculation, misunderstandings and theory that’s been superseded. The tangents that came up in this particular discussion are, I suggest, the result of assuming that if something is true for SSP in particular it must also be true for reading, perception, development or biology in general. Here are some of the inferences that came up in the discussion.

You can’t guess a word from a picture
Children’s books are renowned for their illustrations. Good illustrations can support or extend the information in the text, showing readers what a chalet, a mountain stream or a pine tree looks like, for example. Author and artist usually have detailed discussions about illustrations to ensure that the book forms an integrated whole and is not just a text with embellishments.

If the child is learning to read, pictures can serve to focus attention (which could be wandering anywhere) on the content of the text and can have a weak priming effect, increasing the likelihood of the child accessing relevant words. If the picture shows someone climbing a mountain path in the snow, the text is unlikely to contain words about sun, sand and ice-creams.

I understand why SSP proponents object to the child being instructed to guess a particular word by looking at a picture; the guess is likely to be wrong and the child distracted from decoding the word. But some teachers don’t seem to be keen on illustrations per se. As one teacher put it “often superficial time consuming detract from learning”.

Cues are clues are guesswork
The note to parents referred to ‘clues’ in the pictures. One contributor cited a blogpost that claimed “with ‘mixed methods’ eyes jump around looking for cues to guess from”. Clues and cues are often used interchangeably in discussions about phonics on social media. That’s understandable; the words have similar meanings and a slip on the keyboard can transform one into the other. But in a discussion about reading methods, the distinction between guessing, clues and cues is an important one.

Guessing involves drawing conclusions in the absence of enough information to give you a good chance of being right; it’s haphazard, speculative. A clue is a piece of information that points you in a particular direction. A cue has a more specific meaning depending on context; e.g. theatrical cues, social cues, sensory cues. In reading research, a cue is a piece of information about something the observer is attending to, or a property of a thing to be attended to. It could be the beginning sound or end letter of a word, or an image representing the word. Cues are directly related to the matter in hand, clues are more indirectly related, guessing is a stab in the dark.

The distinction is important because if teachers are using the terms cue and clue interchangeably and assuming they both involve guessing there’s a risk they’ll mistakenly dismiss references to ‘cues’ in reading research as guessing or clues, which they are not.

Reading isn’t natural
Another distinction that came up in the discussion was the idea of natural vs. non-natural behaviours. One argument for children needing to be actively taught to read rather than picking it up as they go along is that reading, unlike walking and talking, isn’t a ‘natural’ skill. The argument goes that reading is a relatively recent technological development so we couldn’t possibly have evolved mechanisms for reading in the same way as we have evolved mechanisms for walking and talking. One proponent of this idea is Diane McGuinness, an influential figure in the world of synthetic phonics.

The argument rests on three assumptions. The first is that we have evolved specific mechanisms for walking and talking but not for reading. The ideas that evolution has an aim or purpose and that if everybody does something we must have evolved a dedicated mechanism to do it, are strongly contested by those who argue instead that we can do what our anatomy and physiology enable us to do (see arguments over Chomsky’s linguistic theory). But you wouldn’t know about that long-standing controversy from reading McGuinness’s books or comments from SSP proponents.

The second assumption is that children learn to walk and talk without much effort or input from others. One teacher called the natural/non-natural distinction “pretty damn obvious”. But sometimes the pretty damn obvious isn’t quite so obvious when you look at what’s actually going on. By the time they start school, the average child will have rehearsed walking and talking for thousands of hours. And most toddlers experience a considerable input from others when developing their walking and talking skills even if they don’t have what one contributor referred to as a “WEIRDo Western mother”. Children who’ve experienced extreme neglect (such as those raised in the notorious Romanian orphanages) tend to show significant developmental delays.

The third assumption is that learning to use technological developments requires direct instruction. Whether it does or not depends on the complexity of the task. Pointy sticks and heavy stones are technologies used in foraging and hunting, but most small children can figure out for themselves how to use them – as do chimps and crows. Is the use of sticks and stones by crows, chimps or hunter-gatherers natural or non-natural? A bicycle is a man-made technology more complex than sticks and stones, but most people are able to figure out how to ride a bike simply by watching others do it, even if a bit of practice is needed before they can do it themselves. Is learning to ride a bike with a bit of support from your mum or dad natural or non-natural?

Reading English is a more complex task than riding a bike because of the number of letter-sound correspondences. You’d need a fair amount of watching and listening to written language being read aloud to be able to read for yourself. And you’d need considerable instruction and practice before being able to fly a fighter jet because the technology is massively more complex than that involved in bicycles and alphabetic scripts.

One teacher asked “are you really going to go for the continuum fallacy here?” No idea why he considers a continuum a fallacy. In the natural/non-natural distinction used by SSP proponents there are three continua involved;

• the complexity of the task
• the length of rehearsal time required to master the task, and
• the extent of input from others that’s required.

Some children learn to read simply by being read to, reading for themselves and asking for help with words they don’t recognise. But because reading is a complex task, for most children learning to read by immersion like that would take thousands of hours of rehearsal. It makes far more sense to cut to the chase and use explicit instruction. In principle, learning to fly a fighter jet would be possible through trial-and-error, but it would be a stupidly costly approach to training pilots.

Technology is non-biological
I was told by several teachers that reading, riding a bike and flying an aircraft weren’t biological functions. I fail to see how they can’t be, since all involve human beings using their brain and body. It then occurred to me that the teachers are equating ‘biological’ with ‘natural’ or with the human body alone. In other words, if you acquire a skill that involves only body parts (e.g. walking or talking) it’s biological. If it involves anything other than a body part it’s not biological. Not sure where that leaves hunting with wooden spears, making baskets or weaving woolen fabric using a wooden loom and shuttle.

Teaching and learning are interchangeable
Another tangent was whether or not learning is involved in sleeping, eating and drinking. I contended that it is; newborns do not sleep, eat or drink in the same way as most of them will be sleeping, eating or drinking nine months later. One teacher kept telling me they don’t need to be taught to do those things. I can see why teachers often conflate teaching and learning, but they are not two sides of the same coin. You can teach children things but they might fail to learn them. And children can learn things that nobody has taught them. It’s debatable whether or not parents shaping a baby’s sleeping routine, spoon feeding them or giving them a sippy cup instead of a bottle count as teaching, but it’s pretty clear there’s a lot of learning going on.

What’s true for most is true for all
I was also told by one teacher that all babies crawl (an assertion he later modified) and by a school governor that they can all suckle (an assertion that wasn’t modified). Sweeping generalisations like this coming from people working in education is worrying. Children vary. They vary a lot. Even if only 0.1% of children do or don’t do something, that would involve 8 000 children in English schools. Some and most are not all or none and teachers of all people should be aware of that.

A core factor in children learning to read is the complexity of the task. If the task is a complex one, like reading, most children are likely to learn more quickly and effectively if you teach them explicitly. You can’t infer from that that all children are the same, they all learn in the same way or that teaching and learning are two sides of the same coin. Nor can you infer from a tenuous argument used to justify the use of SSP that distinctions between natural and non-natural or biological and technological are clear, obvious, valid or helpful. The evidence that supports SSP is the evidence that supports SSP. It doesn’t provide a general theory for language, education or human development.

synthetic phonics, dyslexia and natural learning

Too intense a focus on the virtues of synthetic phonics (SP) can, it seems, result in related issues getting a bit blurred. I discovered that some whole language supporters do appear to have been ideologically motivated but that the whole language approach didn’t originate in ideology. And as far as I can tell we don’t know if SP can reduce adult functional illiteracy rates. But I wouldn’t have known either of those things from the way SP is framed by its supporters. SP proponents also make claims about how the brain is involved in reading. In this post I’ll look at two of them; dyslexia and natural learning.

Dyslexia

Dyslexia started life as a descriptive label for the reading difficulties adults can develop due to brain damage caused by a stroke or head injury. Some children were observed to have similar reading difficulties despite otherwise normal development. The adults’ dyslexia was acquired (they’d previously been able to read) but the children’s dyslexia was developmental (they’d never learned to read). The most obvious conclusion was that the children also had brain damage – but in the early 20th century when the research started in earnest there was no easy way to determine that.

Medically, developmental dyslexia is still only a descriptive label meaning ‘reading difficulties’ (causes unknown, might/might not be biological, might vary from child to child). However, dyslexia is now also used to denote a supposed medical condition that causes reading difficulties. This new usage is something that Diane McGuinness complains about in Why Children Don’t Learn to Read.

I completely agree with McGuinness that this use isn’t justified and has led to confusion and unintended and unwanted outcomes. But I think she muddies the water further by peppering her discussion of dyslexia (pp. 132-140) with debatable assertions such as:

“We call complex human traits ‘talents’”.

“Normal variation is on a continuum but people working from a medical or clinical model tend to think in dichotomies…”.

“Reading is definitely not a property of the human brain”.

“If reading is a biological property of the brain, transmitted genetically, then this must have occurred by Lamarckian evolution.”

Why debatable? Because complex human traits are not necessarily ‘talents’; clinicians tend to be more aware of normal variation than most people; reading must be a ‘property of the brain’ if we need a brain to read; and the research McGuinness refers to didn’t claim that ‘reading’ was transmitted genetically.

I can understand why McGuinness might be trying to move away from the idea that reading difficulties are caused by a biological impairment that we can’t fix. After all, the research suggests SP can improve the poor phonological awareness that’s strongly associated with reading difficulties. I get the distinct impression, however, that she’s uneasy with the whole idea of reading difficulties having biological causes. She concedes that phonological processing might be inherited (p.140) but then denies that a weakness in discriminating phonemes could be due to organic brain damage. She’s right that brain scans had revealed no structural brain differences between dyslexics and good readers. And in scans that show functional variations, the ability to read might be a cause, rather than an effect.

But as McGuinness herself points out reading is a complex skill involving many brain areas, and biological mechanisms tend to vary between individuals. In a complex biological process there’s a lot of scope for variation. Poor phonological awareness might be a significant factor, but it might not be the only factor. A child with poor phonological awareness plus visual processing impairments plus limited working memory capacity plus slow processing speed – all factors known to be associated with reading difficulties – would be unlikely to find those difficulties eliminated by SP alone. The risk in conceding that reading difficulties might have biological origins is that using teaching methods to remediate them might then called into question – just what McGuinness doesn’t want to happen, and for good reason.

Natural and unnatural abilities

McGuinness’s view of the role of biology in reading seems to be derived from her ideas about the origin of skills. She says;

“It is the natural abilities of people that are transmitted genetically, not unnatural abilities that depend upon instruction and involve the integration of many subskills”. (p.140, emphasis McGuinness)

This is a distinction often made by SP proponents. I’ve been told that children don’t need to be taught to walk or talk because these abilities are natural and so develop instinctively and effortlessly. Written language, in contrast, is a recent man-made invention; there hasn’t been time to evolve a natural mechanism for reading, so we need to be taught how to do it and have to work hard to master it. Steven Pinker, who wrote the foreword to Why Children Can’t Read seems to agree. He says “More than a century ago, Charles Darwin got it right: language is a human instinct, but written language is not” (p.ix).

Although that’s a plausible model, what Pinker and McGuinness fail to mention is that it’s also a controversial one. The part played by nature and nurture in the development of language (and other abilities) has been the subject of heated debate for decades. The reason for the debate is that the relevant research findings can be interpreted in different ways. McGuinness is entitled to her interpretation but it’s disingenuous in a book aimed at a general readership not to tell readers that other researchers would disagree.

Research evidence suggests that the natural/unnatural skills model has got it wrong. The same natural/unnatural distinction was made recently in the case of part of the brain called the fusiform gyrus. In the fusiform gyrus, visual information about objects is categorised. Different types of objects, such as faces, places and small items like tools, have their own dedicated locations. Because those types of objects are naturally occurring, researchers initially thought their dedicated locations might be hard-wired.

But there’s also word recognition area. And in experts, the faces area is also used for cars, chess positions, and specially invented items called greebles. To become an expert in any of those things you require some instruction – you’d need to learn the rules of chess or the names of cars or greebles. But your visual system can still learn to accurately recognise, discriminate between and categorise many thousands of items like faces, places, tools, cars, chess positions and greebles simply through hours and hours of visual exposure.

Practice makes perfect

What claimants for ‘natural’ skills also tend to overlook is how much rehearsal goes into them. Most parents don’t actively teach children to talk, but babies hear and rehearse speech for many months before they can say recognisable words. Most parents don’t teach toddlers to walk, but it takes young children years to become fully stable on their feet despite hours of daily practice.

There’s no evidence that as far as the brain is concerned there’s any difference between ‘natural’ and ‘unnatural’ knowledge and skills. How much instruction and practice knowledge or skills require will depend on their transparency and complexity. Walking and bike-riding are pretty transparent; you can see what’s involved by watching other people. But they take a while to learn because of the complexity of the motor-co-ordination and balance involved. Speech and reading are less transparent and more complex than walking and bike-riding, so take much longer to master. But some children require intensive instruction in order to learn to speak, and many children learn to read with minimal input from adults. The natural/unnatural distinction is a false one and it’s as unhelpful as assuming that reading difficulties are caused by ‘dyslexia’.

Multiple causes

What underpins SP proponents’ reluctance to admit biological factors as causes for reading difficulties is, I suspect, an error often made when assessing cause and effect. It’s an easy one to make, but one that people advocating changes to public policy need to be aware of.

Let’s say for the sake of argument that we know, for sure, that reading difficulties have three major causes, A, B and C. The one that occurs most often is A. We can confidently predict that children showing A will have reading difficulties. What we can’t say, without further investigation, is whether a particular child’s reading difficulties are due to A. Or if A is involved, that it’s the only cause.

We know that poor phonological awareness is frequently associated with reading difficulties. Because SP trains children to be aware of phonological features in speech, and because that training improves word reading and spelling, it’s a safe bet that poor phonological awareness is also a cause of reading difficulties. But because reading is a complex skill, there are many possible causes for reading difficulties. We can’t assume that poor phonological awareness is the only cause, or that it’s a cause in all cases.

The evidence that SP improves children’s decoding ability is persuasive. However, the evidence also suggests that 12% – 15% of children will still struggle to learn to decode using SP. And that around 15% of children will struggle with reading comprehension. Having a method of reading instruction that works for most children is great, but education should benefit all children, and since the minority of children who struggle are the ones people keep complaining about, we need to pay attention to what causes reading difficulties for those children – as individuals. In education, one size might fit most, but it doesn’t fit all.

Reference

McGuinness, D. (1998). Why Children Can’t Read and What We Can Do About It. Penguin.

synthetic phonics and functional literacy: the missing link

According to Diane McGuinness in Why Can’t Children Read, first published in 1997, California’s low 4th grade reading scores prompted it in 1996 to revert to using phonics rather than ‘real books’ for teaching reading. McGuinness, like the legislators in California, clearly expected phonics to make a difference to reading levels. It appears to have had little impact (NCES, 2013). McGuinness would doubtless point out that ‘phonics’ isn’t systematic synthetic phonics, and that might have made a big difference. Indeed it might. We don’t know.

Synthetic phonics and functional literacy

Synthetic phonics is important because it can break a link in a casual chain that leads to functional illiteracy:

• poor phonological awareness ->
• poor decoding ->
• poor reading comprehension ->
• functional illiteracy and low educational attainment

The association between poor phonological awareness and reading difficulties is well established. And obviously if you can’t decode text you won’t understand it and if you can’t understand text your educational attainment won’t be very high.

SP involves training children to detect, recognise and discriminate between phonemes, so we’d expect it to improve phonological awareness and decoding skills, and that’s exactly what studies have shown. But as far as I can tell, we don’t know what impact SP has on the rest of the causal chain; on functional literacy rates in school leavers or on overall educational attainment.

This is puzzling. The whole point of teaching children to read is so they can be functionally literate. The SP programmes McGuinness advocates have been available for at least a couple of decades, so there’s been plenty of time to assess their impact on functional literacy. One of them, Phono-graphix (developed by a former student of McGuinness’s, now her daughter-in-law), has been the focus of several peer-reviewed studies all of which report improvements, but none of which appears to have assessed the impact on functional literacy by school leaving age. SP proponents have pointed out that might be because they’ve had enough difficulty getting policy-makers to take SP seriously, let alone fund long-term pilot studies.

The Clackmannanshire study

One study that did involve SP and followed the development of literacy skills over time was carried out in Clackmannanshire in Scotland by Rhona Johnston and Joyce Watson, then based at the University of Hull and the University of St Andrews respectively.

They compared three reading instruction approaches implemented in Primary 1 and tracked children’s performance in word reading, spelling and reading comprehension up to Primary 7. The study found very large gains in word reading (3y 6m; fig 1) and spelling (1y 9m; fig 2) for the group of children who’d had the SP intervention. The report describes reading comprehension as “significantly above chronological age throughout”. What it’s referring to is a 7-month advantage in P1 that had reduced to a 3.5-month advantage by P7.

A noticeable feature of the Clackmannanshire study is that scores were presented as group means, although boys’ and girls’ scores and those of advantaged and disadvantaged children were differentiated. One drawback of aggregating scores this way is that it can mask effects within the groups. So an intervention might be followed by a statistically significant average improvement that’s caused by some children performing much better than others.

This is exactly what we see in the data on ‘underachievers’ (fig 9). Despite large improvements at the group level, by P7 5% of children were more than two years behind their chronological age norm for word reading, 10% for spelling and 15% for reading comprehension. The improvements in group scores on word reading and spelling increased with age – but so did the proportion of children who were more than two years behind. This is an example of the ‘Matthew effect’ that Keith Stanovich refers to; children who can decode read more so their reading improves, whereas children who can’t decode don’t read so don’t improve. For the children in the Clackmannanshire study as a group, SP significantly improved word reading and spelling and slightly improved their comprehension, but it didn’t eliminate the Matthew effect.

The phonics check

There’s a similar within-group variation in the English KS1 phonics check, introduced in 2012. Ignoring the strange shape of the graph in 2012 and 2013 (though Dorothy Bishop’s observations are worth reading), the percentage of Year 2 children who scored below the expected standard was 15% in 2013 and 12% in 2014. The sharp increase at the cut-off point suggests that there are two populations of children – those who grasp phonics and those who don’t. Or that most children have been taught phonics properly but some haven’t. There’s also a spike at the end of the long tail of children who don’t ‘get’ phonics at all for whatever reason, representing the 5783 children who scored 0.

It’s clear that SP significantly improves children’s ability to decode and spell – at the group level. But we don’t appear to know whether that improvement is due to children who can already decode a bit getting much better at it, or to children who previously couldn’t decode learning to do it, or both, or if there are some children for whom SP has no impact.

And I have yet to find evidence showing that SP reduces the rates of functional illiteracy that McGuinness, politicians and the press complain about. The proportion of school leavers who have difficulty with reading comprehension has hovered around 17% for decades in the US (NCES, 2013) and in the UK (Rashid & Brooks, 2010). A similar proportion of children in the US and the UK populations have some kind of learning difficulty. And according to the Warnock report that figure appears to have been stable in the UK since mass education was introduced.

The magical number 17 plus or minus 2

There’s a likely explanation for that 17% (or thereabouts). In a large population, some features (such as height, weight, IQ or reading ability) are the outcome of what are essentially random variables. If you measure one of those features across the population and plot a graph of your measurements, they will form what’s commonly referred to as a normal distribution – with the familiar bell curve shape. The curve will be symmetrical around the mean (average) score. Not only does that tell you that 50% of your population will score above the mean and 50% below it, it also enables you to predict what proportion of the population will be significantly taller/shorter, lighter/heavier, more/less intelligent or better/worse at reading than average. Statistically, around 16% of the population will score more than one standard deviation below the mean. Those people will be significantly shorter/lighter/less intelligent or have more difficulties with reading than the rest of the population.

Bell curves tend to ring alarm bells so I need to make it clear what I am not saying. I’m not saying that problems with reading are due to a ‘reading gene’ or biology or IQ, and so we can’t do anything about them. What I am saying is that if reading ability in a large population is the outcome of not just one factor, but many factors that are to all intents and purposes random, then it’s a pretty safe bet that around 16% of children will have a significant problem with it. What’s important for that 16% is figuring out what factors are causing reading problems for individual children within that group. There are likely to be several different causes, as the NCES (1993) study found. So a child might have reading difficulties due to persistent glue ear as an infant, an undiagnosed developmental disorder, having a mother with mental health problems who hardly speaks to them, having no books at home or because their family dismisses reading as pointless. Or all of the above. SP might help, but is unlikely to address all of the obstacles to word reading, spelling and comprehension that child faces.

The data show that SP enables 11 year-olds as a group to make huge gains in their word reading and spelling skills. That’s brilliant. Let’s use synthetic phonics.

The data also show that SP doesn’t eliminate reading comprehension problems for at least 15% of 11 year-olds – or the word reading problems of around 15% of 6-7 year-olds. That could be due to some SP programmes not being taught systematically enough, intensively enough or for long enough. But it could be due to other causes. If so, those causes need to be identified and addressed or the child’s functional literacy will remain at risk.

I can see why the Clackmannanshire study convinced the UK government to recommend then mandate the use of SP for reading instruction in English schools (things are different in Scotland), but I haven’t yet found a follow-up study that measured literacy levels at 16, or the later impact on educational attainment; and the children involved in the study would now be in their early 20s.

What concerns me is that if more is being implicitly claimed for SP than it can actually deliver or if it fails to deliver a substantial improvement in the functional literacy of school leavers in a decade’s time, then it’s likely to be seen as yet another educational ‘fad’ and abandoned, regardless of the gains it brings in decoding and spelling. Meanwhile, the many other factors involved in reading comprehension are at risk of being marginalised if policy-makers pin their hopes on SP alone. Which just goes to show why nationally mandated educational policies should be thoroughly piloted and evaluated before they are foisted on schools.

References
McGuinness, D. (1998). Why Children Can’t Read and What We Can Do About It. Penguin.
NCES (1993). Adult Literacy in America. National Center for Educational Statistics.
NCES (2013). Trends in Academic Progress. National Center for Educational Statistics.
Rashid, S & Brooks, G (2010). The levels of attainment in literacy and numeracy of 13- to 19-year-olds in England, 1948–2009. National Research and Development Centre for adult literacy and numeracy.

the nation’s report card: functional literacy

Synthetic phonics (SP) proponents make some bold claims about the impact SP has on children’s ability to decode text. Sceptics often point out that decoding isn’t reading – comprehension is essential as well. SP proponents retort that of course decoding isn’t all there is to reading, but if a child can’t decode, comprehension will be impossible. You can’t argue with that, and there’s good evidence for the efficacy of SP in facilitating decoding. But what impact has it had on reading? I feel as if I’ve missed something obvious here (maybe I have) but as far as I’ve been able to ascertain, the answer is that we don’t know.

Despite complaints about literacy from politicians, employers and the public focussing on the reading ability of school leavers, the focus of the English education system has been on early literacy and on decoding. I can understand why; not being able to decode can have major repercussions for individual children and for schools. But decoding and adult functional literacy seem to be linked only by an assumption that the primary cause of functional illiteracy is the inability to decode. This assumption doesn’t appear to be supported by the data. I should emphasise that I’ve never come across anyone who has claimed explicitly that SP will make a significant dent in functional illiteracy. But SP proponents often tut-tut about functional literacy levels and when Diane McGuinness discusses it in Why Children Can’t Read and What We Can Do About It, she makes the implication quite clear.

Armed with a first degree from Birkbeck College, a PhD from University College London and now Emeritus Professor of Psychology at the University of South Florida, McGuinness’ work has focussed on reading instruction. She’s a tireless advocate for SP and is widely cited by SP supporters. Her books are informative and readable, if rather idiosyncratic, and Why Children Can’t Read is no exception. In it, she explains how writing systems developed, takes us on a tour of reading research, points us to effective remedial programmes and tops it all off with detailed instructions for teachers and parents who want to use her approach to teaching decoding. But before moving on to what she says about functional literacy, it’s worth considering what she has to say about science.

“This is doing science.”

Her chapter ‘Science to the rescue’ consists largely of a summary of research into reading difficulties. However, McGuinness opens with a section called ‘What science is and isn’t’ in which she has a go at Ken Goodman. It’s not her criticism of Goodman’s work that bothers me, but the criteria she uses to do so. After listing various kinds of research carried out by journalists, academics doing literature reviews or observing children in classrooms, she says; “None of these activities qualify as scientific research. Science can only work when things can be measured and recorded in numbers” (p.127). This is an extraordinary claim. In one sentence, McGuinness dismisses operationalizing constructs, developing hypotheses, and qualitative research methods (that don’t measure things or put numbers on them) as not being scientific.

She uses this sweeping claim to discredit Goodman, who, as she points out elsewhere, wasn’t a ‘psycholinguist’ (p.55). (As I mentioned previously, McGuinness also ridicules quotes from Frank Smith – who was a ‘psycholinguist’ – but doesn’t mention him by name in the text; that’s tucked away in her Notes section.) She rightly points out that using the words ‘research’ and ‘scientific’ doesn’t make what Goodman is saying, science. And she rightly wonders about his references to his beliefs. But she then goes on to question the phonetics and linguistics on which Goodman bases his model;

“There is no ‘science’ of how sounds and letters work together in an alphabet. This is strictly an issue of categorisation and mapping relationships… Goodman proceeds to discuss rudimentary phonetics and linguistics, leading the reader to believe that they are sciences. They are not. They are descriptive disciplines and depend upon other phoneticians and linguists agreeing with you. …Classifying things is not science. It is the first step to begin to do science.” (p.128)

McGuinness has a very narrow view of science. She reduces it to quantitative research methods and misunderstands the role of classification in scientific inquiry. Biology took an enormous leap forward when Linnaeus developed a classification system that worked for all living organisms. Similarly, Mendeleev’s periodic table enabled chemists to predict the properties of as yet undiscovered elements. Linguists’ categorisation of speech sounds is, ironically, what McGuinness used to develop her approach to reading instruction. What all these classification systems have in common is not just their reliability (level of agreement between the people doing the classification) but their validity (based on the physical structure of organisms, atoms and speech sounds).

McGuinness’s view of science explains why she seems most at home with data that are amenable to measurement, so it was instructive to see how she extracts information from data in her opening chapter ‘Reading report card’. She discusses the results of four large-scale surveys in the 1990s of ‘functional literacy’ (p.10). Two, published by the National Center for Education Statistics (NCES) compared adult and child literacy in the US, and two by the Organisation for European Economic Co-operation (OECD) included the US, Canada and five non-English-speaking countries.

Functional literacy data

Functional literacy was assessed using a 5–level scale. Level 1 ranged from not being able to read at all to a reading task that “required only the minimum level of competence” – for example extracting information from a short newspaper article. Level 5 involved a fact sheet for potential jurors (NCES, 1993, pp.73-84).

In the NCES study, 21% of the US adult population performed at level 1 “indicating that they were functionally illiterate” (McGuinness, p.10) and 47% scored at levels 1 or 2. Despite the fact that level 2 was above the minimum level of competence, McGuinness describes the level 1+2 group as “barely literate”. Something she omits to tell us is what the NCES report has to say about the considerable heterogeneity of the level 1 group. 25% were born abroad. 35% had had fewer than 8 years of schooling. 33% were 65 or older. 26% reported a ‘physical, mental or health condition’ that affected their day-to-day functioning, and 19% a visual impairment that made it difficult for them to read print (NCES, 1993, pp.16-18).

The OECD study showed that functional illiteracy (level 1) varied slightly across English-speaking countries – between 17% and 22%. McGuinness doesn’t tell us what the figures were for the five non-English speaking countries, apart from Sweden with a score of 7.5% at level 1 – half that of the English-speaking countries. The most likely explanation is the relative transparency of the orthographies – Swedish spelling was standardised as recently as 1906. But McGuinness doesn’t mention orthography as a factor in literacy results; instead “Sweden has set the benchmark for what school systems can achieve” (p.11). McGuinness then goes on to compare reading proficiency in different US States.

The Nation’s Report Card

McGuinness describes functional illiteracy levels in English-speaking countries as ‘dismal’, ‘sobering’, ‘shocking’ and ‘a literacy crisis’. She draws attention to the fact that after California mandated the use of the ‘real books’ (whole language) approach to reading instruction in 1987, it came low down the US national league tables for 4th grade reading in 1992, and then tied ‘for a dead last’ with Louisiana in 1994 (p.11). Although California’s score had decreased by only 5 points (from 202 to 197 – the entire range being 182-228) (NCES, 1996 p.47), there was perhaps a stigma attached to being tied ‘dead last with Louisiana’, as phonics was reintroduced into Californian classrooms together with more than a billion dollars for teacher training in 1996, the year before Why Children Can’t Read was first published.

What difference did it make? Not much, it seems. Although California’s 4th grade reading scores had recovered by 1998 (NCES,1999, p.113), and improved further by 2011 (NCES, 2013b), the increase wasn’t statistically significant.

Indeed, whatever method of reading instruction has been used in the US, it doesn’t appear to have had much overall impact on reading standards. At age 17, the proportion of ‘functionally illiterate’ US readers has fluctuated between 14% and 21% – an average of 17% – since 1971 (NCES, 2013b). And in the UK the figure has remained ‘stubbornly’ around 17% since WW2 (Rashid & Brooks, 2010).

Functional illiteracy levels in the English-speaking world are higher than in many non-English-speaking countries, and have remained stable for decades. Functional illiteracy is a long-standing problem and McGuinness, at least, implies that SP can crack it. In the next post I want to look at the evidence for that claim.

References

McGuinness, D. (1998). Why Children Can’t Read and What We Can Do About It. Penguin.
NCES (1993). Adult Literacy in America. National Center for Educational Statistics.
NCES (1996). NAEP 1994 Reading Report Card for the Nation and the States. National Center for Educational Statistics.
NCES (1999). NAEP 1998 Reading Report Card for the Nation and the States. National Center for Educational Statistics.
NCES (2013a). Mega-States: An Analysis of Student Performance in the Five Most Heavily Populated States in the Nation. National Center for Educational Statistics.
NCES (2013b). Trends in Academic Progress. National Center for Educational Statistics.
Rashid, S & Brooks, G (2010). The levels of attainment in literacy and numeracy of 13- to 19-year-olds in England, 1948–2009. National Research and Development Centre for adult literacy and numeracy.

whole language and ideology

It took my son two years to learn to read. Despite his love of books and a lot of hard work, he just couldn’t manage it. Eventually he cracked it. Overnight. All by himself. Using whole word recognition. He’s the only member of our family who didn’t learn to read effortlessly – and he’s the only one who was taught to read using synthetic phonics (SP). SP was the bee’s knees at the time – his reading breakthrough happened a few months before the interim Rose Report was published. Baffled, I turned to the TES forum for insights and met the synthetic phonics teachers. They explained systematic synthetic phonics. They questioned whether my son had been taught SP systematically or intensively enough. (He hadn’t.) And they told me that SP was based on scientific evidence, whereas the whole language approach, which they opposed, was ideologically driven.

SP supporters are among the most vocal advocates for evidence-based education policies, so I checked out the evidence. What I could find, that is. Much of it predated the internet or was behind a paywall. What I did find convinced me that SP was the most effective way of teaching children to decode text. I’m still convinced. But the more I read, the more sceptical I became about some of the other claims made by SP proponents. In the next few posts, I want to look at three claims; about the whole language approach to learning to read, the impact of SP and reading and the brain.

whole language: evidence and ideology

The once popular whole language approach to learning to read was challenged by research findings that emerged during the 1980s and 90s. The heated debate that ensued is often referred to as the Reading Wars. The villains of the piece for SP proponents seemed to be a couple of guys called Goodman and Smith. I was surprised to find that they are both academics. Frank Smith has a background in psycholinguistics, a PhD from Harvard and a co-authored book with his supervisor, George “the magical number seven” Miller. Ken Goodman had accumulated an array of educational awards. Given their credentials, ideology clearly wasn’t the whole story.

In 1971 Frank Smith published Understanding Reading: A Psycholinguistic Analysis of Reading and Learning to Read, which explains the whole language approach. It’s a solid but still readable and still relevant summary of how research from cognitive science and linguistics relates to reading. So how did Smith end up fostering the much maligned – and many would say discredited – whole language approach?

bottom-up vs top-down

By 1971 it was well established that brains process sensory information in a ‘bottom-up’ fashion. Cognitive research showed that complex visual and auditory input from the environment is broken down into simple fragments by the sense organs. The fragments are then reconstituted in the brain, step-by-step, into the whole visual images or patterns of sound that we perceive. This process is automatic and pre-conscious and gets faster and more efficient the more familiar we are with a particular item.

But this step-by-step sequential model of cognitive processing didn’t explain what readers did. Research showed that people read words faster than non-words, can identify words from only a few key features, and that the meaning of the beginning of a sentence influences the way they pronounce words at the end of it (as in ‘her eyes were full of tears’).

According to the sequential model of cognition, this is impossible; you can’t determine the meaning of a word before you’ve decoded it. The only explanation that made sense was that a ‘top-down’ processing system was also in operation. What wasn’t clear at the time was how the two systems interacted. A common view was that the top-down process controlled the bottom-up one.

For Smith, the top-down model had some important implications such as:

• Young children wouldn’t be able to detect the components of language (syllables, phonemes, nouns, verbs etc) so teaching reading using components wouldn’t be effective.
• If children had enough experience of language, spoken and written, they would learn to read as easily as they learned to speak.
• Skilled readers readers would use contextual cues to identify words; poorer readers would rely more heavily on visual features.

Inspired by Smith’s model of reading, Keith Stanovich and Richard West, then graduate students at the University of Michigan, decided to test the third hypothesis. To their surprise, they found exactly the opposite of Smith’s prediction. The better readers were, the more they relied on visual recognition. The poorer readers relied more on context. It wasn’t that the skilled readers weren’t using contextual cues, but their visual recognition process was simply faster – they defaulted to using context if visual recognition failed.

As Stanovich explains (Stanovich, 2000, pp.21-23) the flaw in most top-down models of reading was that they assumed top-down controlled bottom-up processing. What Stanovich and West’s finding implied (and later research supported) was that the two systems interacted at several levels. Although some aspects of Smith’s model were wrong it was based on robust evidence. So why did SP proponents think it was ideologically driven? One clue is in Ken Goodman’s work.

a psycholinguistic guessing game

Smith completed his PhD in 1967, the year that Goodman, then an Associate Professor at Wayne State University, Detroit, published his (in)famous article in the Journal of the Reading Specialist “Reading: A psycholinguistic guessing game”. The title is derived from a key concept in contemporary reading models – that skilled readers used rapid, pre-conscious hypothesis testing to identify words. It’s an eye-catching title, but open to misunderstanding; the skilled ‘guessing game’ that Goodman was referring to is very different from getting a beginner reader to have a wild stab at an unfamiliar word. Which was why Goodman and Smith recommended extensive experience of language.

Goodman’s background was in education rather than psycholinguistics. According to Diane McGuinness, (McGuinness, 1998, p.129) Goodman does have some peer-reviewed publications, but the most academic text I could find was his introduction to The Psycholinguistic Nature of the Reading Process published in 1968. In contrast to the technical content of the rest of the book, Goodman’s chapter provides only a brief overview of reading from a psycholinguistic perspective and in the four-sentence chapter summary he refers to his ‘beliefs’ twice – a tendency McGuinness uses as evidence against him. (Interestingly, although she also ridicules some quotes from Smith, his name is tucked away in her Notes section.)

Although Goodman doesn’t come across as a heavyweight academic, the whole language model he enthusiastically supports is nonetheless derived from the same body of evidence used by Smith and Stanovich. And the miscue analysis technique Goodman developed is now widely used to identify the strategies adopted by individual readers. So where does ideology come in?

Keith Stanovich sheds some light on this question in Progress in Understanding Reading. Published in 2000, it’s a collection of Stanovich’s key papers spanning a 25-year career. In the final section he reflects on his work and the part it played in the whole language debate. Interestingly, Stanovich emphasises what the two sides had in common. Here’s his take on best practice in the classroom;

“Fortunately the best teachers have often been wise enough to incorporate the most effective practices from the two different approaches into their instructional programs.” (p.361)

and on the way research findings have been used in the debate;

“Whole language proponents link [a model of the reading process at variance with the scientific data] with the aspects of whole language philosophy that are legitimately good and upon which virtually no researchers disagree.” (p.362)

correspondence and coherence

For Stanovich the heat in the debate didn’t come from disagreements between reading researchers, but from the clash between two conflicting theories about the nature of truth; correspondence vs coherence. Correspondence theory assumes that there is a real world out there, independent of our perceptions of it. In contrast the coherence theory assumes that our “knowledge is internally constructed – that our evolving knowledge is not tracking an independently existing world, but that internally constructed knowledge literally is the world” (p.371, emphasis Stanovich’s). The whole language model fits nicely into the coherence theory of truth, so research findings that challenged whole language also challenged what Stanovich describes as the “extreme constructivism” of some whole language proponents.

Stanovich also complains that whole language proponents often fail to provide evidence for their claims, cherry-pick supporting evidence only, ignore contradictory evidence and are prone to the use of straw men and ad hominem attacks. He doesn’t mention that synthetic phonics proponents are capable of doing exactly the same. I don’t think this is due to bias on his part; what’s more likely is that when his book was published in 2000 the whole language model had had plenty of time to filter through to classrooms, policy makers’ offices and university education and philosophy departments. The consensus on synthetic phonics was relatively new and hadn’t gathered so much popular support. Fifteen years on, that situation has changed. In my experience, some SP proponents are equally capable of making sweeping claims, citing any supporting evidence regardless of its quality, and being dismissive towards anyone who disagrees with anything they believe. Which brings me to the subject of my next post; claims about what SP can achieve.

references

Goodman, K (Ed). 1998. The Psycholinguistic Nature of the Reading Process. Wayne State University Press.
McGuinness, D. (1998). Why Children Can’t Read and What We Can Do About It. Penguin.
Smith, F. (1971). Understanding Reading: A Psycholinguistic Analysis of Reading and Learning to Read. Lawrence Erlbaum. (My copy is 4th edition, published 1988).
Stanovich, K (2000). Progress in Understanding Reading. Guilford Press.