Teaching students Skills is possible and necessary

Teaching students skills is possible and necessary

I read on blogs and on twitter, that teachers should only teach knowledge and should not teach generic skills. I believe this is wrong and dangerous.  This blog looks at how to teach skills, and examines the arguments for and against skills teaching.

Procedural knowledge and declarative knowledge.

It is well established that long term memory is divided into two parts:

Declarative memory: ‘knowing what’ commonly called ‘knowledge’ For example
the battle of Hastings was in 1066.
The square of 5 is 25.

Procedural memory “knowing how” commonly called ‘skills’. How to ride a bike, play the piano, do a long jump, for teachers I would add how to read a difficult text.

Procedural and declarative memories are stored in different parts of the brain.

Procedures we need to teach our students might include:

  • how to plan an essay,
  • how to read a text to gain maximum understanding,
  • how to approach a maths problem unlike one you have seen before
  • how to justify an argument in a history essay for example by using historical facts
  • How to critique a psychological experiment, and its findings

Teaching how to comprehend a text (an example of skills teaching)













The flow diagram shows a procedure you could explain to students, ask them to use, and to self-assess against.  You might adapt it: for example make a more detailed version for more able learners. In response to a comment below I would add that you might like to adapt this process to make it subject specific.

This procedure is not obvious to students. Some students are surprised to be told that you should not give up half way through a text when you begin to lose understanding, or that reading something more than once improves understanding.

Teaching the skill

How could you use this flow diagram with your students?  You explain the process, stressing what would happen if one of the steps was missed out. Your students use the process on a piece of text, this text explains new learning your students need anyway.  Then, following Reuven Feuerstein, I suggest you teach for transfer using class discussion on metacognition and bridging:

The teacher says:

‘We’ve just read a difficult text:
How did we do that? Why did we do it that way? What would have happened if we had missed out steps in the flow diagram? (Class discussion).

Where else could we use that process? Does it only work for this topic? (No!) Could you use the process for material on the internet? (Print it off, then you can). Could you use it for material in a library book? (photocopy the book, then you can). Could you use it for History? Maths?

I call the questions in bold Feurerstein’s ‘killer questions’.

Note you have taught some content that you had to teach anyway.  But by getting your students to read material using the given procedure you have simultaneously helped teach a skill. Students discuss their old reading habits and why this method is better. Maybe they improve the procedure. Maybe students use slightly different versions of the process.  They self assess against the process after each trial and set targets for next time. After sufficient repeated practice, preferably with feedback, the process becomes a habit, they don’t need the flow diagram any more.  You have taught the skill.  This is not a new teaching technique, its sometimes called scaffolded learning.

There is more detail on this here, including downloads for flow diagrams etc on other skills:

Teaching students skills

The best account is in chapter 21 in Evidence Based Teaching.

Does it work if you teach academic skills?

The best source of evidence for what works, and what doesn’t, is systematic reviews, or meta studies. See here https://geoffpetty.com/the-uses-and-abuses-of-evidence-in-education/

On my website you can download a summary of a review of high quality research on the teaching of study skills, thinking skills, and the like.  The main findings were:

  • It is best to integrate or embed the teaching of academic skills with content learning, rather than teach a bolt-on unit on study skills, though both methods work. There is an average effect size of above 0.7 if skills are embedded, which shows that teaching skills is very effective. (A bolt-on study skills unit averages at about 0.45)
  • It’s best to use tasks that are real, embedded, and subject specific.  E.g. Teach essay planning while they are writing a real essay for their course; or teach note taking by looking at the notes they have made in a real lesson.
  • Students need repeated practice on real material that they have to learn anyway
  • Students need to use the skills with a clear sense of purpose and to orchestrate their use.
  • Students need to self monitor, self assess and self regulate their use of the skills.
  • It’s greatly helps to teach even able students skills such as extracting key points; concept mapping; note making; summary writing. Strategies that stress the meaning and structure of information seem to have a very large effect on understanding.
  • Meta-cognition is a notable feature of all the successful (high effect size) studies that the review finds.  Meta-cognition is students thinking about their learning, and self-regulating their own learning. For example, students reflecting about the way they read a piece of text, and so setting themselves goals for improvement, then evaluating how this went.

These findings directly contradict the idea that skills can’t be taught. Whether skills can be taught or not is of course an empirical question, and can only be answered by doing it, and seeing what happens in carefully arranged experiments in real classrooms. Over and over again, teachers have taught skills effectively, and although it took a little time to do this, their students got much better results.

(Yes, I know effect sizes are far from perfect measures, but this is a high average effect size for hundreds of studies, and effect sizes are the only way we have of comparing teaching methods and gauging their relative effectiveness. Any other way of gauging the effectiveness of teaching study skills, including your own guesses, will be less reliable.)

The controversy

Some cognitive scientists seem to be saying that skills cannot be taught, despite the strong empirical evidence to the contrary.  The argument goes something like this I think, but I have this wrong I expect. Its hard to express an argument you don’t agree with.

Procedures or skills such as riding a bike playing the piano and doing a long jump are stored in the procedural memory which is a subset of implicit memory.  This memory is unconscious, its a set of habits brought about by repeated use of the skill.  Someone with a skill finds it hard to verbalise or explain the skill, because it is automatic and unconscious.  We develop procedural skills though experience without really trying, we just get better at things the more we do them.  Therefore these skills don’t need teaching.

The trouble with this argument is that empirically it is clear that if you teach skills well, they can be learned well.


An example of this ‘don’t teach skills’ argument was drawn to my attention on twitter by @LeoToAquarius to a paper by Prof John Sweller Emeritus Professor of Education, in a ResearchEd talk he said that “our knowledge of human cognition,” tells us:

“We should be teaching domain subject specific knowledge, not generic skills”

“Generic skills are far more basic and far more important than domain-specific knowledge, but they do not need to be taught because we have evolved over countless generations to acquire them effortlessly and unconsciously simply by membership of a society’

“*It is a waste of students’ time placing these skills in a curriculum because we have evolved to acquire such skills easily, automatically, unconsciously and without tuition’”

See http://www.learnandteachanything.com/blog for a useful summary of his presentation

This begs the question what Prof Sweller means exactly by ‘generic skills’. But I worry that the impression given is that even study skills cannot be taught, which we know from classroom trials is not the case.
In his defence I think the term ‘generic skills’ was used in the 1960s and 1970s by those who believed that just getting students to use a skill will improve it. This doesn’t work well.

Dr Y Weinstein-Jones (@doctorwhy) is also a cognitive psychologist and she believes that skills can be taught as you can see from this very useful link that she tweeted:
She like me made use of the Hattie, Biggs and Purdie reference below.

(1) Donker, A. S., de Boer, H., Kostons, D., Dignath-van Ewijk, C. C., & van der Werf, M. P. C. (2014). Effectiveness of learning strategy instruction on academic performance: A meta-analysis. Educational Research Review, 11, 1–26.

(2) Hattie, J., Biggs, J., & Purdie, N. (1996). Effects of learning skills interventions on student learning: A meta-analysis. Review of Educational Research, 66, 99–136.

2 thoughts on “Teaching students Skills is possible and necessary

  1. Ref :Teaching students skills is possible and necessary @LeoToAquarius @doctorwhy @LearningSpy @olivercavigliol https://geoffpetty.com/?p=1162

    I’d be obliged if you could add the word ‘generic’ before skills in the first sentence of your blog as that’s what was being criticised & not subject specific skills 🙂 which you do mention at the end of your post.

    The notion the procedural skills cannot be taught is false see

    Teaching Procedural Skills

    Thomas E Norris, MD,1 Sam W Cullison, MD,1 and Stephan D Fihn, MD, MPH2


    They must be taught explicitly using different modalities – i.e the five senses and then repeatedly practiced in order to appear ‘unconsciousness’
    The structure of knowledge is itself a domain i.e epistemology and therefore I tried applying your flowchart to a chemistry text, computer programming and maths textbooks – I can give examples if you wish and it rapidly fails because it is generic

    There are better alternatives but they are all domain specific eg

    Learning Domain Knowledge for Teaching Procedural Skills


    That paper has a model like you flowchart for machine learning but it does not transfer

    For science at any level & maths at degree level (for maths replace ‘hypothesis’ with ‘proposition’ & ‘experiment’ with ‘proof’) your flowchart does not work

    To understand a science textbook at any level a simple flowchart of the scientific methods such as


    is needed – I can tweet the links to you separately if you wish to check the pics
    eg Observation–>Darwins finches Beaks–>Hypothesis ‘the type of food determines the design of the body and behaviour of the animal’ & ‘hypothesis of natural selection & some more hypotheses’ –>verifed using experimental observations leads to the theory of Evolution

    [Your flow chart can’t work for Science and Maths as it is not domain specific because it does not emphasise the need to build a network of experimentally verified hypotheses of which the student must know the experiments & evidence that proved them to build the relevant theory]

    It also does not work for Geography or History as those subjects try and use the scientific method by replacing ‘experiments’ with ‘case studies’ in Geography and using paired ‘historical examples’ in History
    So in summary we can & must teach explicitly procedures for learning that apply to different subjects as in the 1st part of your blog post but these do not transfer to other domains eg science to music

    So learning the scientific method does not make you better at playing or composing music but it may help you design better musical instruments

    There is no experimental evidence that generic thinking skills eg Blooms / SOLO (or any other pseudoscientific taxonomy) can be taught or that doing lots of them makes a person better at them in general – doing lots of scientific analysis does not make a person better at analysing poetry as it’s a different procedure nor does does it make them quicker at learning how to analyse poetry


    • I’ve changed the first sentence as you suggest, however see my other blog on skills.

      If skills are entirely subject specific as you seem to be arguing, why is it that someone with a good higher education is often better at reading a text unrelated to their specialism, than someone only educated up to the age of 16? Take for example reading a complex newspaper article.

      I think some skills are more transferrable than others. Reading comprehension has a good degree of transferability, but of course no where near complete transferability. You will always be able to understand something better when it is within a subject you have specialised in as you point out. However I can read and understand articles on US politics, but have never studied it formally.

      My main point is, that for a given level of background knowledge, you will nearly always understand something better if you read it twice than if you read it once. You will almost always understand it better if you isolate the difficult passages and give them more attention. This is why it is fair to call reading comprehension a generic skill, even if you are better at reading in your specialism than you are outside it.

      I don’t approve of all attempts to teach such generic skills, for example the vague notion that just doing a creative tasks will make you necessarily more creative. The generic skill needs to be taught explicitly within a domain, on a particular task, and then the teacher needs to teach for transfer. This is hard to achieve I grant you. but that doesn’t mean its impossible to achieve.

Leave a Reply

Your email address will not be published. Required fields are marked *