Phonics: 13 things every teacher should know

Research is clear that using phonics to teach reading works, argues Clare Sealy, but why don’t more teachers know about the science behind it?

Clare Sealy


Nearly every primary school “does” phonics to some degree and it is often seen as a necessary but dull chore, akin to trying to get children to eat some greens when they’d more naturally want to be eating ice cream. Doubts over whether it is the “right” thing for children to be doing still lurk in some quarters, too. 

This negative perception exists despite reading probably being the most widely researched subject in education and the evidence for the superiority of systematic phonics being overwhelming. 

Why is this evidence so poorly known among educators?

Read more:

The ignorance about the evidence matters because there are still instances of phonics being taught badly. For example, using levelled books rather than cumulative, decodable readers, in which all the words encountered matched the phonemes already taught. Doing the former fundamentally undermines systematic phonics. 

What is phonics?

The whole point of systematic phonics is that we explicitly teach children the link between the sounds we already use when we speak and those squiggles on a page. There are about 44 different sounds (the exact amount will depend on your accent) and 175+ different ways of writing those sounds using letters. Contrast this with having to learn every single word you ever want to read – what a laborious way of doing things. 

Of course, critics of phonics will argue that no one is expecting children to learn every single word as a separate entity. They will argue that readers use a variety of cues and cross-reference between them. These cues will be led by the quest for meaning, with what makes sense in context and in terms of grammar, aided and abetted by reference to word shape, picture cues and, if all else fails, phonics.

They cite frustrating experiences of teaching children who do not understand what they read, who read without meaning, and say surely phonics is to blame for that? And learning to read via lovely stories as opposed to learning letter-sound correspondences in isolation seems much more fun.

But this is where we have to do what is actually true, and not what we want to be true. It’s our professional duty to attend to what the science tells us and, where necessary, change our practice.

Phonics research

So, what does the science tell us?

  1. The first thing to realise is that there is really overwhelming evidence that systematic phonics is far more effective than other approaches. This has been repeatedly acknowledged by various national groups tasked with reporting back to governments about the evidence. For example, the National Reading Panel in the US (2000), the Australian National Inquiry into the Teaching of Literacy (2005) and, in the UK, the Rose Report (2006) all surveyed the literature and concluded that systematic phonics was by far the most effective teaching strategy for early literacy acquisition.

  2. The hardest part of learning to decode is learning to hear speech sounds individually, as this is not something we have evolved to do – we hear the sounds blended together in words, not as separate entities. However, it is a mistake to think that the ability to hear individual speech sounds is a prerequisite for learning to read, and therefore to avoid linking sounds with print until speech sounds can be identified orally. This is because the thing that helps most in learning to distinguish individual speech sounds is being taught sound-symbol correspondences. So, Reception teachers need to crack on with teaching phonics whether or not children already have well-developed phonological awareness.

  3. Fluent adult readers might not think that they still use phonics, but we know from many experiments that they still do, alongside simultaneously reading whole words. Fluent readers have two routes used simultaneously, a sounds route and a spelling route that uses whole- or part-word recognition (See Daniel Willingham, The Reading Mind).

  4. The existence of a spelling route might seem, at first glance, to undermine the necessity of starting to teach reading through phonics. However, the research is very clear here: starting with phonics is much more effective. What is more, whole-word recognition – what researchers call orthographic representations – will just happen automatically over time with sufficient practice. It is an important second step, but it is just that, a second step. Once children can decode, it is very important they read really regularly, both at school and at home, in order to build up this orthographic store.

  5. It is often said that English is just too complicated to be taught by phonics alone and that children need to be taught high-frequency sight words alongside phonetically regular words. In actual fact, it is perfectly possible to teach children to read – and spell – very well without resorting to teaching sight words as whole global shapes. Instead, children are taught that the different sounds we make can be written in a variety of different ways. Note how this is different from the more usual approach of teaching letters “saying” or “making” sounds. If children are explicitly taught the range of different ways a particular sound can be written, then so-called irregular words can be decoded (read) and encoded (written) using sounds. For example, if we teach children that  the grapheme <ai> can spell the sound /e/, they can read not just “said”, but also “again” and “against”. It’s a “buy one, get several free” situation.

  6. Teachers often worry that too strong a focus on decoding interferes with children making sense of what they read. The phrase “barking at print” is sometimes invoked to describe children who can decode but don’t understand what they read. Some may appear not to attend at all to the meaning behind the words they are decoding. This is an understandable worry, but knowing more about the various stages in learning to read can help reassure us that phonics is not the enemy here. Indeed, it is (in the long run) the enabler of meaning. But initially, learning to decode requires intensive use of working memory, so much so that there is very little left over to attend to the meaning behind what you have just decoded. This is rather similar to when you first learn to drive and are expending so much attention on clutch control that attending to the traffic is difficult – which is why we have dual control cars for beginning drivers and we tend to have our first lessons in quiet back roads. When children are at the early stages of decoding, teachers can “control” the meaning aspect by reading the successfully decoded sentence back to the child.

  7. Over time, with regular decoding practice, children will become fluent at decoding and then will have space in their working memory to attend to what the words actually mean themselves, without needing the teacher to do this for them. Not all children will realise that they now need to do this though, which is why – once children can decode fluently – teachers need to explicitly model how they think about texts as they read them, and how they check for sense and make links to what they know as they do. This is particularly important when children are making the jump from mainly reading out loud to an adult to reading silently to themselves. Unless teachers make it clear that readers should expect what they read to make sense, and have strategies such as re-reading a sentence to fall back on when it doesn’t, some children think that reading is limited to just decoding.

  8. Fluent decoders also need teachers to model reading with prosody – that is to say, reading with the “music” of the text in mind – the changes in pace, in tone and the way we stress different words in sentences, all of which add meaning to the words on the page. Some children don’t realise that this is how their reading in their head is meant to sound to them!

  9. Some teachers seem to believe that people who argue for phonics think that phonics is the only important element of learning to read. This is not true at all. As well as excellent, evidence-informed phonics teaching, children also need to be immersed in a stimulating language environment and exposed to the wealth of rich children’s literature. These are in no way mutually exclusive, you just don’t do them at the same time of day.

  10. The most important factor determining whether or not children understand what they read is the strength of their understanding of spoken language. Providing children with opportunities to develop their spoken listening comprehension is crucial for their future ability to read with comprehension (as shown in the Simple View of Reading (Gough and Tunmer, 1986) – recommended by Sir Jim Rose in his review).

  11. The second most important factor determining whether or not children understand what they read is their background knowledge. Providing children with opportunities to learn about the world around and beyond them right from the start of their education provides the necessary bedrock for later reading for meaning.

  12. The searchlight, or “three-cue” model of reading grew out of Goodman’s model of reading as a “psycholinguistic guessing game”, where readers cross-reference their guesses about what a word might be against what makes sense in context, what makes sense grammatically, word shape etc.  Goodman’s work was explored in the 1970s by psychologist Keith Stanovich, who started off assuming Goodman was probably correct that, as people become better readers, they rely more on their knowledge of vocabulary and language structure to read words and don’t need to pay as much attention to the letters. But much to his surprise, he found that the opposite was true. Skilled readers recognise words almost instantly; poorer readers can’t do this so are reliant on contextual cues to guess what the word is (Keith E Stanovitch, Progress in Understanding Reading, 2000). This finding has been replicated many times by other researchers. Goodman’s three-cue model has also been shown by MRI brain scans to be false. It takes less than 0.2 seconds for us to recognise words – far too short a time for the putative cross-checking process to take place. However, despite having been categorically disproved, Goodman’s model lingers on zombie-like – the theory that will not die. It is the model that lies behind using banded or levelled readers that use repetitive text. While some children will learn to read using this model (presumably working out the code for themselves), far too many do not. In effect, using this model is teaching children to read in the same way that poor readers do! 

  13. Some teachers argue that phonics is tedious and the decodable early readers are boring, which is to completely miss the point – they aren’t intended to be great works of literature, they are intended to teach children to read (so that they can go on to read great works of literature). Teaching phonics is as tedious as you want it to be. Young children love learning to do “grown-up” things like reading, and if you show how excited and impressed you are that children can now blend p-i-n, then they will be excited and impressed with themselves, too. That’s where the engagement comes in, with the success. The lesson isn’t meant to entertain the teacher after all, it’s meant to develop the child and some of the things that help a child develop, especially early on, are pretty prosaic. The sentence “Clap, clap, clap on the big, red bus” is not, of itself, desperately interesting. However, being able to turn all those squiggles into actual words that make up a sentence is amazing! Criticising a decodable reader because of its limited storyline is like telling a babbling baby their conversation is boring. To the “phonics is boring” brigade, I say, “stop raining on the children’s parade!” Rather you should be delighted and enthused by helping young children take their first steps on the reading journey.

Clare Sealy is head of curriculum and standards for Guernsey

This article was amended on 9 August 2020 to replace "systematic synthetic phonics" with "systemic phonics".

Register to continue reading for free

It only takes a moment and you'll get access to more news, plus courses, jobs and teaching resources tailored to you

Clare Sealy

Latest stories