I’ve been reading American psychologist and geneticist Robert Plomin’s latest book Blueprint:How DNA Makes Us Who We Are. It was on my ‘to read’ list anyway, but I was surprised to see on Twitter that in it, Plomin had made what might be construed as negative comments about the UK Phonics Screening Check, so my interest was piqued.
For context, while I know that many educators find Plomin’s work uncomfortable and controversial, I have no particular problem with it, although I don’t necessarily agree with all of his conclusions. Indeed, as his work was already a subject of debate back in 2014 I was prompted to attend Andrew Sabisky’s talk on genetics at that year’s ResearchEd and I enjoyed it so much that I wrote a blog about it: Nature and Nurture – the Genetics of Education. Why do children differ?
In Blueprint, The Phonics Screening Check discussed in two chapters, 6 and 8.
Firstly, Chapter 6. The point, I believe, Plomin is making in this chapter is that, contrary to the established belief of most experts involved in the treatment/definition/management of human behaviour etc., the genes that affect cognition and behaviour have very general rather than specific effects. In other words they are ‘generalist genes’. Consequently, many conditions such as anxiety, schizophrenia, bipolar disorder etc. that experts may previously have seen as very separate, even mutually exclusive, entities, are in fact expressions of the same genetic effects and they share high genetic correlations. Plomin notes that many experts are very surprised at such findings.
He goes on to say that the Phonics Screening Check is a favourite example of his where he believes that the ‘experts’ have failed to appreciate the general nature of the genetic effects involved (the highlights in the following quote are mine):
“Education-related skills such as reading,mathematics and science show even higher genetic correlations: about 0.7. One of my favourite examples of generalist genes involves reading. A test called the Phonics Screening Check was developed to distinguish two components of reading that were thought to be fundamentally different processes. One is the ability to read familiar words quickly and accurately (fluency). The other is the ability to sound out non-words(phonetics). A test like this is administered to all 600,000 five- and six-year-olds in the UK because there is an assumption that it separates out these two components of reading, fluency and phonetics. The test involves reading aloud as quickly as possible a list of age-appropriate familiar words and‘non-words’. For example, familiar words might be ‘dog’ and ‘exercise’.Non-words are word-like combinations of letters never seen before that are matched in difficulty level to the real words, such as ‘pog’ and ‘tegwop’. The reasonable idea underlying this interesting test is that reading familiar words should be automatic, but non-words that children have never seen before require sounding them out, which is phonetics. Reasonable ideas are often wrong, as in this case. The genetic correlation between reading familiar words and non-words is 0.9, making this one of the most powerful examples of generalist genes. That is, the same DNA differences are responsible for individual differences in fluency and phonetics, even though fluency and phonetics are thought to be completely different neurocognitive processes. “
Plomin, Robert. Blueprint (pp.68-69). Penguin Books Ltd. Kindle Edition
The problem is that the assertions made by Robert Plomin here simply aren’t true. The single purpose of the Phonics Screening Check from its inception and as consistently stated by the DfE is “to confirm that all children have learned phonic decoding to an age-appropriate standard.” Furthermore, the aim states, “children who have not reached this level should receive extra support from their school to ensure they can improve their decoding skills and will then have the opportunity to retake the phonics screening check.”
Neither do I recognise his description of reading as comprising “fluency and phonetics”. In fact, even the use of the word ‘phonetics’ rather that ‘phonics’ raises questions as to where Plomin has drawn his information from. As for his description of the check itself. There is no ‘speed’ element whatsoever. It is accuracy only that is measured. The child who can quickly read the word, as if by sight, will score no better than the child who still needs to sound out and then blend no matter how long it takes them. Children are given as long as they need to answer, and the guidance to teachers is clear:
“Children should be given as long as necessary to respond to a word, although in most cases, 10 seconds should be sufficient. The teacher should decide when it is appropriate to tell the child to move onto the next word, taking care not to try to move the child on if they are still trying to decode the word.”
It may sound pedantic, but again, language is important. The PSC isn’t a ‘test’ such as one that might seek to measure a child’s relative ability compared to other children, it is a check to see if the individual child has developed sufficient decoding skill. The bar is set at a level that experts in the science and teaching of reading deem to be sufficient, in combination with their wider language skills, for the child to become a competent reader. Should a child fail to achieve the required standard, the onus is on the school to provide the support necessary to address this. The non-word section of the check is there, as it is in almost all similar diagnostic assessments of phonic ability, to provide a finer level of analysis to inform the teacher administering the check and to avoid the children who are over reliant on a memorised bank of sight words or ‘lucky guesses’ from slipping through the diagnostic net.
The high correlation between reading familiar words and non-words at 0.9 comes as no surprise at all to the experts on reading who designed the check, or who advocate that systematic synthetic (or linguistic) phonics is taught well in all schools. They, and the science, would tell you that the evidence unequivocally shows that successful, highly competent readers use decoding to identify words, although this eventually becomes such a fluent process that they don’t realise it is what they are doing. To reading/phonics experts “fluency” means fast, automatic decoding. Therefore no reading/phonics experts regard “fluency and phonetics” as “completely different neurocognitive processes”. If it were true that the PSC sought to discriminate between performance on two “fundamentally different process” then you would expect that it would report the results of the two separately but this data is neither collected nor reported.
Then in Chapter 8, Plomin revisits the PSC:
“In 2010 Michael Gove, the recently appointed UK Secretary of State for Education, decided that UK schools should go back to teaching reading using phonics to sound out letters and words. At that time the national curriculum used the ‘look and say’ technique, in which children learn whole words on sight and are expected gradually to pick up the ability to recognize letter sounds. To make sure that teachers are following through on this curriculum change, all Year 1 pupils are tested on the Phonics Screening Check. The Phonics Screening Check, mentioned earlier, involves reading aloud as quickly as possible a list of forty age-appropriate familiar words and non-words. For example, some easy words are ‘dog’, ‘big’ and ‘hot’ and more difficult words are ‘project’, ‘frequent’, ‘exercise’. Non-words are word-like combinations of letters the child has never seen before that are matched to the real words in difficulty level. They also vary from easy (‘pog’, ‘dat’, ‘bice’)to difficult (‘supken’, ‘tegwop’, ‘slinperk’). The reasonable idea behind this interesting test is that reading familiar words should be automatic, but non-words that the child has never seen before require sounding out, which is phonetics. How well children performed on the Phonics Screening Check was assumed to be due to how well their teachers taught phonics. Schools are named and shamed if their students do not meet national standards. As usual in education, genetics was not even mentioned in the debate surrounding the phonics test. However, when we administered the test in the Twins Early Development Study, we found that it was among the most highly heritable traits ever reported at this age, with heritabilities of about 70 per cent. This means that the test is not measuring how well children are taught reading. Instead,it is a sensitive measure of genetically driven aptitudes for learning to read.Environmental factors shared by children going to the same school as well as growing up in the same family account for less than 20 per cent of the variance in children’s performance on the phonics test.”
Plomin, Robert. Blueprint (p.82). Penguin Books Ltd. Kindle Edition.
Again the erroneous assertion regarding the ‘idea’ behind the ‘test’ and the suggestion of a ‘speed’ element. I, for example, did not have to sound out ‘supken’, ‘tegwop’, ‘slinperk’ to read them fluently, and neither will the most able Year 1s, but neither I, nor they, would score any better on the PSC than a child who slowly sounded out and then (correctly) blended each and every real and pseudo-word.
Where he is absolutely right though, is that the PSC was at least in part introduced to “make sure that teachers are following through on this curriculum change” although the requirement to teach phonics was in fact first set out not by Gove in 2010, but in 2006 by the then Labour Government’s Education Secretary Ruth Kelly, in response to the clear findings of former Ofsted inspector Jim Rose, as set out in his report the Independent Review of the Teaching Early Reading published in March that year. Sadly, this requirement was largely ignored by many, not least because most teachers had received very little instruction in phonics during their Initial Teacher Training. To that end I consider the PSC, and the focus it brought to bear on this aspect of the teaching of reading, to have been the single greatest milestone to ensuring that all children receive a greater equality of educational opportunity.
Writing and reading are a human-invented technology. Acquiring the knowledge and skills required is not a natural process and while some might pick enough of it up on their own to do sufficiently well, for many, if not most pupils, it needs to be explicitly taught in order for them to do as well as possible. In the same way, schools must teach the curriculum in order that individual pupils might achieve their genetically determined potential on curriculum-based tests. While Plomin explains, and I am happy to accept, that different schools and school types make little difference to the attainment of pupils, I am assuming that this presupposes that all the schools are at least teaching the basic requirements of the curriculum.
Hitherto, the teaching of reading was always bedevilled by the so called ‘long tail of under-achievement’ in other words, a distorted pattern of results with more than the expected (if reading is like other cognitive abilities) number of pupils languishing in the lower range. What Robert Plomin’s work leads me to expect is, that if phonics were being universally well taught, then I should expect to see a normal, bell-shaped distribution of reading/phonics results. What I would like to help to achieve is for that curve to be shifted as far to the right as possible and for the vast majority of children to meet, or exceed, the standard set by the Phonics Screening Check.
Addendum
Interestingly, while Professor Plomin is clearly mistaken in his understanding of the underpinnings and conduct of the Phonics Screening Check, what he describes in Blueprint does, however, bear a striking resemblance to the Test of Word Reading Efficiency (TOWRE) reading test that he and his colleagues used in their “Telephone Testing and Teacher Assessment of Reading Skills in 7-year-olds” paper published in 2005. It is that test, and not the PSC, that is built on the false premise that Professor Plomin ascribes to the PSC i.e. that fluent (supposed sight word) reading is a fundamentally different process to decoding.
