Most parents of good readers have no idea how they contribute to their children becoming good readers. A conversation I had with a New York Times editor is typical: he told me his daughter was devoted to her books, and when I asked what he had done to encourage this proficiency and love of reading, his response was “not a damn thing”.

Of course, he obviously had done things - reading to his daughter when she was young, filling the house with books, being a newspaper editor - but what he meant was he didn’t plan it. There was no thought-out intervention that facilitated the skill being instilled.

Teachers must be more strategic. There is no room in the schools system to leave reading to chance: the skill needs to be taught and students need to be highly accomplished in it, as soon as possible.

But quite how teachers should teach reading is contentious. All educators will be aware of the reading wars: a vociferous and nasty set of arguments about the best way to teach children to read.

On one side you have advocates of the system stipulated and tested by the UK government: systematic synthetic phonics. This is reading words by learning individual (or combination) letter sounds. The sounds are then blended together to make words.

On the other side you have the whole-word approach: whole words are learned by sight, often through repetition (look and say), or “worked out” by focusing on meaning, context, pictures and other clues.

The rival armies are entrenched and the battles vicious. Yet in this age of research-informed teaching, shouldn’t this debate have been put to bed already? Well, science can indeed give us some answers, but whether it crowns an absolute victor is debatable. And whether anyone is prepared to accept the arguments is another matter altogether.

Symbols and sounds

Before we look at how we teach reading, we have to look at what science has taught us about reading.

Because writing uses visual symbols that signify sounds, children who are learning to read must master three things. First, they must be able to distinguish letters. Second, they must learn the mapping between these visual symbols and their auditory counterparts - that the letter “o” sometimes goes with one sound (as in the word tone) and sometimes another (as in ton). Third, they must be able to hear individual speech sounds.

The latter turns out to be especially hard for kids. Let me give you an example: what sound do you associate with the letter “p”? You might think of it as “puh” but that’s actually two sounds: the sound of the letter “p” and the vowel sound “uh”. The sound associated with the letter “p” is actually just a plosion of air - your vocal chords don’t vibrate at all.

And these individual speech sounds vary depending on the surrounding context. Try this: put your hand in front of your mouth and say “pot”. You feel the puff of air when you say the “p”. Now do the same thing saying “spot”. The puff is stronger for pot than spot. So we talk about “the sound the letter `p’ makes” as if there were one sound associated with “p”, but that’s an abstraction, an ideal.

Children who have trouble learning to read often have difficulty hearing individual speech sounds; children who more or less teach themselves to read hear them easily.

Understanding where one word ends and another begins is another must-have skill for mastering reading. Kids don’t hear individual words as well as adults do. In a standard test of this ability, you give the child a short sentence to keep in mind - say, “I like yellow bananas”. Then you give him a small basket of blocks and ask him to arrange a line of them, one block for each word in the sentence. More often than not, the child will use more or fewer blocks than the number of words. If they are to become a good reader, this needs to be fixed.

And lastly, readers need comprehension, too - a way of creating meaning across sentences, something akin to the way syntax connects meaning across words. Sometimes that’s straightforward. For example, if you read “He dropped the tissues. They fell to the floor”, it’s simple to connect the meaning of the two sentences: the action in the first caused the action in the second.

But writers often omit information needed to connect sentences, as in this example: “The stranger tapped the window. The dog barked.” To connect these sentences, you must know something about dogs. The author could provide that information, but providing all the information to make connections across sentences would make writing tedious. So authors gamble on when they can safely omit information and when it must be included. Where the connecting information is not in the text, comprehension demands it to be in your head or clear from the context.

Figuring things out in this way takes time and mental effort. It interrupts the flow of the text and you may lose the thread of the argument or story. Although a bit of this sort of problem-solving is satisfying, even fun, too much of it makes reading slow and difficult. How much unknown stuff can a text have in it before a reader will declare “Mental overload!” and call it quits? The answer varies depending on the reader’s attitude towards reading and motivation to understand that particular text. Still, studies have measured readers’ tolerance of unfamiliar vocabulary and have estimated that they need to know about 98 per cent of the words for comfortable comprehension.

Hard to comprehend

What does all this mean for the reading wars? Well, first of all, the notion of being a “good reader” is clearly problematic. If being a “good reader” actually means “knowing a little bit about a lot of stuff” so that you’re assured of knowing the information the author felt could be omitted, then reading tests don’t work the way most people think they do. Reading tests purport to measure a student’s ability to read: once I know your ability to read, I ought to be able (roughly) to predict your comprehension of any text I hand you. But reading comprehension depends heavily on how much you happen to know about the topic of the text, not just whether you can decode the words.

So when we talk about the reading wars we need to be careful, as the whole-word versus phonics debate is actually a decoding war. It’s just one part of learning to read.

With that out of the way, we can move on to looking at these decoding strategies through the lens of what we know about how kids read.

With phonics, you plan instruction to introduce the letter-sound relationships and do so in a particular order, teaching the most common letter-sound pairs first. There are about 44 phonemes (unique sounds) in the English language, and 120 graphemes (written representations of unique sounds). The idea with phonics is that you give children a code for reading that can be applied to any text, because it encapsulates the skills mentioned earlier that kids need to master (letter recognition and mapping to sounds).

You’d think this would be a pretty sensible approach but there is a counterargument with a long history. The position of the whole-words troops is at least 200 years old, appearing in a French monograph from 1787: The Real Way to Learn Any Language, Living or Dead, Through the French Language by Nicholas Adam.

Adam points out that when you teach a child the name of an object - a shirt, say - you don’t list the parts, telling the child, “These are the buttons, here are the cuffs,” and so on. No, you tell the child, “It’s a shirt.” Likewise, Adam says, “Hide from them all the ABCs. Entertain them with whole words which they can understand and which they will retain with far more ease and pleasure than all the printed letters.”

Some 50 years later, American education pioneer Horace Mann agreed that “the advantages of teaching children by beginning with whole words are many”. He referred to isolated letters as “skeleton-shaped, bloodless, ghostly apparitions” and remarked that it was little wonder children felt deathlike when confronted by them.

Both Adam and Mann suggested that learning to read was natural. Indeed, they thought that it was as natural as learning to speak - a suggestion still put forward today by some theorists, although they are very much in the minority.

The argument suggests that what makes reading unnatural and difficult is the process of drilling children in letters. Instead, they say, we should immerse children in reading and writing tasks that are pleasurable and authentic. Give the child experience with reading that makes sense to him, that has a clear point to it, and he’ll be motivated to engage with it and will learn to read more or less effortlessly.

But this strategy brings a substantial disadvantage. Learning to read in this way represents a huge task to human memory because kids must remember what each word looks like, and the average high schooler needs to know something like 50,000 words.

A whole-word advocate would reply that not all 50,000 words need to be learned immediately, and you can often make a good guess as to the identity of a word based on the meaning of other words in the sentence. They argue that readers should be encouraged not to rely solely on print to read. For new readers, we should use reading materials that have other supports to figure out meaning - pictures, for example, that tell the story. If we do these things (provide rich, authentic literary experiences; start with words, not letters; and teach children to use all the sources of information available), whole-word advocates say that children will figure out the letter-sound correspondences as they go.

Acting natural

So who is right? The problem with whole-word theory is that it makes a fundamental assumption that is almost certainly wrong. Reading is not natural. When a skill is “natural”, we expect to observe three things. First, everyone will learn the skill without great difficulty, and typically they will learn it by observation without the need for overt instruction. Second, given that it’s part of our inheritance as human beings, we expect that the skill can be observed in all cultures all over the world. Third, the proposal that our nervous system is primed to learn the skill implies that it will probably be evolutionarily old.

These three features are true for some human skills: walking, talking, reaching and appreciating social interactions, for example. But none of them is true for reading.

Next, if you look at the studies conducted on reading then there seems to be a clear winner. Numerous experiments have compared how well kids learn to read when instructed using phonics or when instructed using whole-word methods, going back nearly a hundred years. The governments of three English-speaking countries (the US, Britain and Australia), as well as the European Union countries, all came up with the same strategy to reach a conclusion from these studies: blue-ribbon panels of scientists were appointed to sort through the data and write a report.

All four panels came to the same conclusion: it’s important to teach phonics, and to teach it in a planned, systematic way, not on an as-needed basis. Similar conclusions have been reached by panels assembled by US scientific organisations.

So is that the debate sorted? Not exactly. Although all these reports were in agreement that phonics instruction is important, it’s not as though kids taught using whole-word methods don’t learn to read. In fact, if you compare their reading achievement to that of kids taught using phonics, there’s a lot of overlap. The advantage conferred by using phonics instead of the whole-word method is moderate, not huge.

A sense of balance

It’s also not the case that every child does better with phonics than he or she would have done with whole-word instruction. The importance of phonics instruction varies depending on what the child knows when reading instruction begins. It is less important for kids who, when they start school, can already hear individual speech sounds and understand that letters stand for sounds. They are likely to figure out the code with just a little support. But for kids who lack that knowledge, phonics instruction is likely to be very important.

Finally, the emphasis that whole-word strategies place on children’s literature as a key tool is pretty much uncontested as a positive, beneficial part of learning to read.

Hence, what you actually get in many classrooms is “balanced literacy”. This suggests that phonics be taught, but in the context of an array of activities, especially ones that offer authentic literacy experiences (as opposed to, for example, completing a worksheet), and that differentiation between students as to the amount of each strategy needed is preferable.

It’s certainly an appealing compromise, a truce that all should be able to sign up to. But does balanced literacy work? The quick-and-dirty answer is that it should. We know that the pieces - phonics instruction and children’s literature - work. Unfortunately, giving a firmer research-based answer will have to wait, for two reasons. First, we know that reading success is influenced by many factors, so it’s hard to draw firm conclusions from individual experiments. Second, there is a lot of variation in what actually happens in a balanced literacy programme.

But what we can say is that increasing evidence confirms what is likely to be your intuition: different kids learn better from different activities, depending on the strengths and interests they bring to the classroom. Until we get evidence suggesting otherwise, then, it would be safe to say that balanced literacy is the best solution we have and that with reading, as with any other area of teaching, the best course of action is to react to the child with different strategies, not to make the child react to just one.

Daniel Willingham is a professor of psychology at the University of Virginia in the US and the author of Why Don’t Students Like School? This is an edited version of two chapters from his new book Raising Kids Who Read: what parents and teachers can do, published by Jossey-Bass, which will be released in the UK on 3 April and is available to pre-order from Amazon at bit.lyKidsWhoRead

Classroom reading activities: lessons from cognitive psychology

We’re not completely in the dark on which activities are more likely to help kids read. Here are four principles from cognitive psychology that - until we have firmer research about what works - would be prudent to bear in mind when thinking about classroom activities.

1. Don’t forget phonics. The point of balanced literacy is “phonics plus rich literacy experiences”. If phonics instruction is one of 16 possible activities, you do want to guard against the possibility that they are viewed as all being equally important. Phonics should be the largest chunk: if the English language block of teaching is between 90 and 120 minutes, I’d hope to see 20 to 30 minutes devoted to phonics.

2. Students can focus on only one new thing at a time. Some literacy activities seem to demand that kids do two things at once. For example, when the teacher is composing text and thinking aloud as she does so, she’s both modelling the writing process and giving the students an implicit phonics lesson. But we know from other research that kids (and adults) can’t focus on two things at a time - especially two ideas that are pretty challenging. Lessons that focus on one thing at a time are more likely to be successful.

3. You learn more from doing than watching. When you’re told to watch someone, it’s easy to let your mind wander and think something else. Hence, I’m wary of shared reading, in which the students follow along while someone else reads.

4. Feedback matters. Corrected errors contribute to learning. Uncorrected errors do not and may contribute to an error-laden habit. Hence, when children are learning to decode, silent reading is not going to be nearly as helpful as reading aloud.