Are all skills composed of knowledge?

And if not, what does that mean for how we teach them?

Last week, I wrote about the relationship between skills and knowledge - and how Keir Bloomer, one of the architects of the skills-based Scottish Curriculum for Excellence, had said “The problem is we did not make sufficiently clear that skills are the accumulation of knowledge. Without knowledge there can be no skills.”

One of the big responses to my article and Bloomer’s quotation was something along the following lines.

‘Maybe there is some truth to this, but are all skills composed of knowledge? Maybe this is true of history and literature, but is it true of practical skills like drama and football? Is it even true of something like maths?’

Here's my answer.

• All complex skills¹ are composed of smaller units, and have to be taught by building up those smaller units.

• Sometimes (typically, but not always, in "academic" subjects) we call those smaller units knowledge. Sometimes (typically, but not always, in "practical" subjects) we call those smaller units sub-skills.

Here are some examples of what I mean

• History breaks down into knowledge: a typical end goal of a history curriculum unit might be to write an analytical essay about the causes of the First World War. The sub-units needed to achieve these skills include a lot of what we'd typically call knowledge - memorising dates, understanding sequences of events, knowing the roles played by key characters.

• Football breaks down into sub-skills: a typical end goal of training to play football might be to play - and win! - an 11-a-side match. The sub-units needed to achieve these skills are what we'd typically call sub-skills: being able to control the ball in tight spaces, pass accurately, tackle and head the ball, etc.

There are different labels for the small steps, but the crucial point is that both skills can be broken down into small steps and taught that way.

So is it sometimes misleading to talk about breaking down skills into knowledge?

I agree that it can be misleading to talk about 'knowledge' in the context of teaching practical skills like football, because then people assume you are talking about learning to play football using textbooks. In my 2013 book Seven Myths about Education, I specifically addressed this issue and explained how “breaking it down” might work for something like football.

“In football, you eventually want a football team to play an 11-a-side game and win that game. But the best way to achieve that is not to get children playing regular competitive 11-a-side games from a young age. Just as with projects, you should break down the complex problem of winning an 11-a-side game into smaller, simpler problems, and practise those. The fundamental basic team skill required in football is to be able to keep the ball. The fundamental basic individual skill follows from this – it is to be able to control the ball, which means having a good first touch, and being able to give and take a pass.”

However, that didn’t stop lots of people accusing me of wanting to teach football with textbooks. In my 2017 book, Making Good Progress, I spent a lot more time on this issue - in fact, in some ways, the entire book is about “breaking it down”. I also reframed the language I used. Instead of talking about skills vs knowledge, I talked about the contrast between a “generic skills” model and a “deliberate practice” model. Generic skills models assume you can acquire a skill directly, by practising the final skill over and over (eg writing an essay, playing an 11-a-side match). Deliberate practice models think you need to break the skill down into smaller steps (eg memorising chronologies, passing drills).

I also came back again and again to one of my favourite analogies, which is from a practical skill - marathon running. You don’t train for a marathon by running endless marathons.

Another term I used a lot in Making Good Progress was the ‘model of progression’, taken from Dylan Wiliam’s Embedded Formative Assessment. The model of progression is essentially the chain of knowledge and sub-skills that builds to the final end goal.

In some school subjects like early reading, and in many sports, the model of progression is well-researched and understood - perhaps with some points of controversy, and with some room for innovation. In most school subjects, however, there is relatively little research on this - and that’s one of the most frustrating things about constantly having to have the skills-knowledge and the generic skills-deliberate practice debate. It stops you from moving on to the far more important and productive next question, which is how do you know what the right “small steps” are?

OK then, how do you know what the right small steps are?!

This is the million-dollar question! Ultimately, it is an assessment question, and it is one that we at No More Marking are really interested in gathering empirical data on! We have two assessment techniques that can help. One is Comparative Judgement, which makes it easy to gather information on the end goal, particularly in subjects where the end goal involves writing. The second technique is multiple-choice questions, which can be used formatively to gather information on the steps on the way. We can then compare results from these two different assessments to see which steps on the way result in better performance at the end goal.

For any complex skill, you could hypothesise a number of different ways of breaking it down. Not all will be equally effective, and some will result in dead ends instead of the end goal.

How can you tell if your “small step” assessments are causative or just correlations?

This is also a million-dollar question. For example, we have already identified that there are some multiple-choice writing questions that are strongly correlated with scores on extended pieces of writing, as measured by Comparative Judgement. It’s tempting to say, let’s teach every student to answer this MCQ right and then they will all get better at writing. They might - but it might also turn out that performance on that MCQ and the extended writing are actually being caused by some other factor that we aren’t measuring.

There are analogies in sport. For example, in marathon running, one popular workout is “Yasso 800s”, where you train for a marathon by running 10 * 800 metre repeats. These are supposed to be predictive of your final marathon time, in the following way: the pace you can run ten 800m intervals in minutes corresponds to your marathon time in hours. So if you run your 800-meter repeats in 3:30, for instance, this predicts a marathon time of 3:30.

Yasso 800s are very popular, but there is some debate about whether a) they are truly predictive and b) even if they are predictive to begin with, are they causative - eg, if you do them again and again, do they build up the physical adaptations necessary to make you better at running 26.2 miles?

My own take is that for complex skills like reading, writing, marathon running, etc., there is never going to be one component we can focus on exclusively. The whole point is that complex skills are made up of many sub-units. Data analysis can help us work out the best set of activities and the right amount of time to spend on each.

Our future posts will set out more of the research we’re doing to answer this question for writing instruction.

1. Strictly speaking, all biologically secondary skills. We don’t need to be explicitly taught biologically primary skills.