play: schools are for children, not children for schools

Some years ago, the TES carried an article about a primary school that taught its pupils how to knit. I learned to knit at school. My mum dutifully used my first attempt – a cotton dishcloth – for months despite its resemblance to a fishing net with an annoying tendency to ensnare kitchen utensils. The reason I was taught knitting was primarily in order to be able to knit. But the thrust of the TES article wasn’t about the usefulness of knitting. It was that it improved the children’s maths. It seemed that at some point since the introduction of mass education in England the relationship between schools and the real world had changed. The point of schools was no longer to provide children with knowledge (like maths) that will help them tackle real-world problems (like knitting), but vice versa – the point of useful real-world skills was now to support performance in school.

school readiness

I was reminded of the knitting article when Sir Michael Wilshaw, then chief inspector of Ofsted, suggested to inspectors that not all early years settings are preparing children adequately for school. In a comment to the BBC he added;

More than two-thirds of our poorest children – and in some of our poorest communities that goes up to eight children out of 10 – go to school unprepared,” he said. “That means they can’t hold a pen, they have poor language and communication skills, they don’t recognise simple numbers, they can’t use the toilet independently and so on.”

His comments prompted an open letter to the Telegraph complaining that Sir Michael’s instruction to inspectors to assess nurseries mainly in terms of preparation for school “betrays an abject (and even wilful) misunderstanding of the nature of early childhood experience.” One of the signatories was Sue Cowley, who had blogged about the importance of play. Her post, like Sir Michael’s original comments, generated a good deal of discussion.

Old Andrew responded promptly. He comments “This leads me to my one opinion on early years teaching methods: OFSTED are right to judge them by outcomes rather than acting as the “play police” and seeking to enforce play-based learning“.

The two bloggers have homed in on different issues. Sue Cowley is concerned about the shift in focus from childhood experience to ‘school-readiness’; Old Andrew is relieved that Ofsted inspectors are longer expected to ‘enforce play-based learning’. The online debate has also shifted from the original question implicit in Sir Michael’s comments and in the response in the letter to the Telegraph i.e. what is the purpose of nurseries and pre-schools? to a question posed by Old Andrew; “Is there any actual empirical evidence on the importance of play? All the “evidence” seems to be theoretical.”

empirical evidence

Responses from early years teachers to questions about evidence for the benefits of play are often along the lines of “I have the evidence of my own eyes”, which hasn’t satisfied the sceptics. Whether you think it’s a satisfactory answer or not depends on the importance you attach to direct observation.

The problem with direct observation is that it’s dependent on perception, which is notoriously unreliable. David Didau has blogged about some perceptual flaws here. He also mentions some of the cognitive errors that occur when people draw conclusions from observations. The scientific method has been developed largely to counteract the flaws in our perception and reasoning. But it doesn’t follow that direct observation is completely unreliable. Indeed, direct observation is the cornerstone of empirical evidence.

Here’s an example. Let’s say I’ve noticed that every time I use a particular brand of soap, my hands sting and turn bright red. It wouldn’t be unreasonable to conclude that I have an allergic response to an ingredient in the soap – but I wouldn’t know that for sure. There could be many causes for my red, stinging hands; the soap might be purely coincidental. The conclusions about causes I could draw solely from my direct observations would be pretty speculative.

But the direct observations themselves – identifying the brand of soap and what happened to my hands – would be a lot more reliable. It’s possible that I could have got the brand of soap wrong and could have imagined what happened to my hands, but those errors are much less likely than the errors involved in drawing conclusions about causality. I could easily increase the reliability of my direct observations by involving an independent observer. If a hundred independent observers all agreed that a particular brand of soap was associated with my and/or other people’s hands turning bright red, those observations wouldn’t be 100% watertight but they would be considered to be fairly reliable and might prompt the soap manufacturer to investigate further. Increasing the reliability of my conclusion about the causal relationship – that the soap caused an allergic reaction – would be more challenging.

is play another Brain Gym?

What intrigued me about the early years’ teachers responses was their reliance on direct observation as empirical evidence for the importance of play. Most professionals, if called upon to do so, can come up with some peer-reviewed research that supports the methods they use, even if it means delving into dusty textbooks they haven’t used for years. I could see Old Andrew’s point; if play is so important, why isn’t there a vast research literature on it? There are three characteristics of play that would explain both the apparent paucity of research and the teachers’ emphasis on direct observation.

First, play is a characteristic typical of most young mammals, and young humans play a lot. At one level, asking what empirical evidence there is for its importance is a pointless question – a bit like asking for evidence for the importance of learning or growth. Play, like learning and growth, is simply a facet of development.

Second, play, like most other mammalian characteristics, is readily observable – although you might need to do a bit of dissection to spot some of the anatomical features of mammals. Traditionally, play has been seen as involving three types of skill; locomotor, object control and social interaction. But you don’t need a formal peer-reviewed study to tell you that. A few hours’ observation of a group of young children would be sufficient. A few hours’ observation would also reveal all the features of play Sue Cowley lists in her blog post.

Third, also readily apparent through direct observation is what children learn during play; the child who chooses to play with the shape-sorter every day until they can get all the shapes in the right holes first time, the one who can’t speak a word of English but is fluent after a few months despite little direct tuition, the one who initially won’t speak to anyone but blossoms into a mini-socialite through play. Early years teachers watch children learning through play every day, so it’s not surprising they don’t see the need to rely on research to tell them about its importance.

The features of play and what children can learn from it are not contentious; the observations of thousands of parents, teachers, psychologists, psychiatrists and anthropologists are largely in agreement over what play looks like and what children learn from it. This would explain why there appears to be little research on the importance of play; it’s self-evidently important to children themselves, as an integral part of human development and as a way of learning. In addition, much of the early research into play was carried out in the inter-war years. Try finding that online. Or even via your local library. Old Andrew’s reluctance to accept early years teachers’ direct observations as evidence might stem from his admission that he doesn’t “really have much insight into what small children are like.”

play-based education

The context of Old Andrew’s original question was Michael Wilshaw’s comments on school readiness and the response in the Telegraph letter. A recent guest post on his blog is critical of play-based learning, suggesting it causes problems for teachers higher up the food chain. Although Old Andrew says he’d like to see evidence for the importance of play in any context, what we’re actually talking about here is the importance of play in the education system.

Direct observation can tell us what play looks like and what children learn from it. What it can’t tell us about is the impact of play on development, GCSE results or adult life. For that, we’d need a more complex research design than just watching and/or recording before-and-after abilities. Some research has been carried out on the impact of play. Although there doesn’t appear to be a correlation between how much young mammals play and their abilities as adults, not playing does appear to impair responsiveness and effective social interaction. And we do know some things about the outcomes of the more complex play seen in children (e.g. Smith & Pellegrini, 2013).

Smith & Pellegrini agree that a prevailing “play ethos” has tended to exaggerate the evidence for the essential role of play (p.4) and that appears to be Old Andrew’s chief objection to the play advocates’ claims. Sue Cowley’s list describes play as ‘vital’, ‘crucial’ and ‘essential’. I can see how her choice of wording might give the impression to anyone looking for empirical evidence in the research literature that research findings relating to the importance of play in development, learning or education were more robust than they are. I can also see why someone observing the direct outcomes of play on a daily basis would see play as ‘vital’, ‘crucial’ and ‘essential’.

I agree with Old Andrew that Ofsted shouldn’t be enforcing play-based learning, or for that matter, telling teachers how to teach. There’s no point in training professionals and then telling them how to do their job. I also agree that if grand claims are being made for play-based learning or if it’s causing problems later on, we need some robust research or some expectation management, or both.

Having said that, it’s worth noting that for the best part of a century nursery and infant teachers have sung the praises of play-based learning. What’s easily overlooked by those who teach older children is the challenge facing early years teachers. They are expected to make ‘school-ready’ children who, in some cases and for whatever reason, have started nurseries, pre-schools and reception classes with little speech, who don’t understand a word of English, who can’t remember instructions, who have problems with dexterity, mobility and bowel and bladder control, or who find the school environment bewildering and frightening. Sometimes, the only way early years teachers can get children to engage or learn anything at all is through play. Early years teachers, as Sue Cowley points out, are usually advocates of highly structured, teacher-directed play. What’s more, they can see children learning from play in real time in front of them. The key question is not “what’s the empirical evidence for the importance of play?” but rather “if children play by default, are highly motivated to play and learn quickly from it, where’s the evidence for a better alternative?”

I’m all in favour of evidence-based practice, but I’m concerned that direct observation might be being prematurely ruled out. I’m also concerned that the debate appears to have shifted from the original one about preparation for school vs the erosion of childhood. This brings us back to the priorities of the school that taught knitting in order to improve children’s maths. Children obviously need to learn for their own benefit and for that of the community as a whole, but we need to remember that in a democracy school is for children, not children for school.

Originally posted in August 2014.


Pellegrini, A & Smith PK (2005). The Nature of Play: Great Apes and Humans. Guilford Press.
Smith, PK & Pellegrini, A (2013). Learning through play. In Tremblay RE, Boivin M, Peters (eds). Encyclopedia of Early Childhood Development [online]. Montreal, Quebec: Centre of Excellence for Early Childhood Developmentand Strategic Knowledge Cluster on Early Child Development 1-6. Available at Accessed 11.8.2014.