Direct instruction or discovery learning – is there a middle way?

I recently attended Paul Kirschner’s presentation at ResearchEd London where he laid out the evidence for the power of direct instruction, which Kirschner describes as a process whereby teachers 1) emphasise academic goals, 2) ensure that learners are involved in learning, 3) select the learning objectives and monitor learner progress 4) structure the learning activities and give immediate academically focused feedback, 5) create a task-oriented yet ‘relaxed’ learning environment.. And the evidence for direct instruction is very strong.

Alongside this strong argument for teachers understanding and using direct instruction, we were all very clearly told that discovery learning does not work. And here is where the presentation fell down for me. Because it’s not the case that discovery learning has no value. There’s a reason why early years curriculums around the world are designed around the principles of discovery learning. And as a result of this, early years practitioners say the children in their care are curious, have a growth mindset, take control of their own learning and ask a lot of questions, all surely characteristics we would want for all learners in our care. Much care is taken to ensure that university learners experience opportunities for discovery learning: following their own lines of enquiry and drawing their own conclusions.

Can we say the same about learners in the primary years, or the secondary years? Many teachers would say that this curiosity, propensity to ask questions, and growth mindset reduces as learners go through the school system and many university lecturers bemoan the absence of these skills in their students. Reasons for this may include naturally increasing maturity and the social pressures that come with the solidification of peer groups throughout the primary and secondary years, but could we also see this as attributable to the move away from discovery learning as pupils progress through their school career?

When working with teachers to develop mathematics teaching, a subject in which the vast majority of lessons are taught through direct instruction, we begin by asking them to consider the advantages and disadvantages of teaching mathematics through discovery, and teaching through direct instruction, and the below table is broadly representative of the answers they give:


And when this discussion is over, we ask them whether they think it is possible to teach mathematics solely through one of the two methods – clearly it isn’t. Children will be very unlikely to discover decimal notation or the written numerals or the protocols for drawing a graph. This can only come through a lesson which tells or instructs the learners. And yet, if we shift towards teaching entirely through direct instruction, where are the opportunities for learners to apply learning to complex unseen problems, which are the kinds of problems for which adults routinely and often unconsciously use mathematics in their daily personal and professional lives? And how are we promoting self-regulation and the value of mistake making?

In fact, Kirschner himself admits there is value in discovery learning in which the teacher is not a passive agent of pupils’ interests, but an active shaper of pupils’ thinking, within the parameters of the discovery activity – much like we see in the best early years good practice.

And in fact, there is a third possibility beyond the polarised concepts of discovery and direct intruction: the concept of ‘guided discovery’. This is one where the teacher has carefully constructed a ‘low threshold, high ceiling’ problem with which all the learners can engage, and in which, crucially, there is a ‘cognitive conflict’ they need to resolve. This problem is designed to align with a very specific mathematical schema, within which the teacher intends to challenge learners’ thinking and enable progression in reasoning. Learners work together in mixed achieving groups to find a way through the cognitive conflict, and the teacher enables a metacognitive discussion amongst learners in which they review the different approaches to the problem and decide which is the most elegant, the most efficient and the one they would use if they tackled a problem like this in the future.

This ‘guided discovery’ approach is Cognitive Acceleration, now known as Let’s Think, a rigorously evaluated approach to teaching and learning from King’s College London which, since the 1980s, has demonstrated it can make a significant difference to all learners, and not just in the subject in which it was taught. Recent studies also demonstrate high impact in new subject areas like English, and the EEF describe it as ‘particularly promising’Programmes for teachers new to Let’s Think run every year, especially in London.

If this is the case, why is Let’s Think not being routinely used in classrooms across the UK?  For two very good reasons:

  1. It’s difficult for teachers. As the table above shows, direct instruction involves less workload, is more predictable, and means the teacher feels more in control of the learning. Contingent talk, which is integral to discovery learning, is difficult: it requires very secure subject knowledge and a familiarity with the thought patterns of young learners, which usually only accrues to teachers with several years of teaching experience, something the UK is struggling to achieve.
  2. It only works in mixed achieving classes. The last 30 years have seen setting and streaming introduced to almost all secondary schools and the majority of primary schools. This is despite a huge body of evidence showing the damage it can do to learning.

So, in summary, what can senior leaders and teachers do to get the right balance between discovery and direct instruction?

  1. Support teachers to develop the confidence and capacity to teach contingently.  Scripted lessons won’t do this, as they suggest that learners are predictable – and they aren’t!
  2. Ensure your schemes of work and teaching and learning policies enable the correct balance to be struck between discovery and direct instruction – for learners of all ages. Don’t be fooled by people who promote one approach and denigrate the other.
  3. Find out about ‘guided discovery’ approaches like Let’s Think that do work and give teachers opportunities to spend time with experts in the approaches, through properly structured professional development programmes – and this means giving time over an extended period of time!

And finally, don’t be fooled by the ‘what works’ agenda and its proponents on Twitter and in the blogging world. The body of evidence is more complicated than they would have you believe! To get a more balanced view of the research evidence join our research-focused programmes, including our first Journal Club in October.