Metacognition in practice: what the updated EEF guidance means for teaching

A pupil completes a task correctly after several prompts. That is useful information, but it leaves an important question: what will the pupil choose to do when the prompts are absent? Metacognition becomes practical when teachers can identify that decision, teach an approach and look for its use on another task.

The EEF's second-edition guidance was published on 13 November 2025. It is the edition verified for this briefing on 7 September 2026, not a new 2026 research finding. It develops recommendations for primary and secondary schools and is aimed particularly at teachers and leaders supporting professional development.

What the updated guidance asks teachers to teach

The report connects pupils' knowledge of themselves, available strategies and the task with how they direct their learning. It recommends explicit teaching, teacher modelling, purposeful discussion, appropriate challenge and guided practice leading towards independence. These approaches sit within subject teaching: pupils need useful subject knowledge and methods before they can decide which method fits.

Our interpretation is that a reflection sheet cannot carry this work on its own. A pupil needs something specific to notice or change while working. The full report provides the detail behind the recommendations and should be read before adopting a school-wide routine.

Read the impact figure with its limitations

The Toolkit strand, whose evidence review is dated May 2025, reports an average equivalent to eight additional months of progress and identifies 355 studies. It rates the evidence as high, while explaining that many studies were not independently evaluated. Evaluations involving organisations connected to an approach tend to show larger effects, which can influence the overall estimate.

This is evidence about a broad family of teaching approaches, not a forecast for a purchased programme, an individual pupil or a particular school. The Toolkit also notes that practice is difficult to implement well. Its low-cost rating does not remove the need for staff time and professional development. Treat the headline as a reason to examine the approach carefully, not as a target to promise governors.

A classroom example to discuss together

This is an original Schoolgle planning example, not a reported trial. Imagine a class comparing two accounts of the same event. Before reading, ask pupils to decide which question they are trying to answer. Give them a small evidence space with room for a claim, a supporting detail and something the account leaves uncertain.

After the first account, pause. A pupil may have a plausible claim but no supporting detail. Instead of rewarding confidence, ask them to point to the passage their claim depends on. After the second account, invite them to keep, qualify or change the claim and say what prompted that decision.

For the follow-up, use two short accounts of a different event and change the order of presentation. Note whether pupils use evidence without being told which sentence matters. Do not treat a longer answer as automatically better reasoning; a brief, well-supported correction may reveal more.

Give feedback a job to do

The EEF's feedback guidance, published on 11 June 2021, focuses on teacher-delivered feedback. Its recommendations include establishing learning intentions, identifying gaps and planning how pupils will use the feedback. It does not establish that a particular automated hint or game reward has the same effect.

For our classroom example, the useful feedback concerns the missing support for a claim. Leave time for a revision and inspect what changed. Recording that feedback was delivered is different from recording that the pupil acted on it.

Plan a manageable trial of the teaching routine

The EEF's implementation guidance, third edition dated 24 April 2024, treats implementation as work involving people, context and a structured process. Its process guidance covers exploration, preparation, delivery and sustaining an approach. A launch meeting is therefore only one part of the job.

Our suggested starting record has four decisions: the subject task, the pupil difficulty being investigated, the teaching change and the evidence the team will revisit. Choose a review date that fits the teaching sequence. Include what staff need to rehearse and what existing activity will make room for it.

• Keep an initial response and a later response to a different but comparable task.
• Record the help actually given, including reading support, modelled examples and hints.
• Look for pupils who still struggle, rather than allowing a class average to end the discussion.
• Separate whether the routine was used from whether learning improved.

A digital record can preserve those observations, but it cannot infer a pupil's thinking from clicks alone. Look for an explanation, a purposeful change or successful use on a new task. Review such evidence with the teacher's knowledge of the pupil. This is a proposed evaluation routine, not a claim that Schoolgle has demonstrated improved attainment.