Research School Network: Tracking the Traces of Metacognition Mark Miller looks for the tangible signs that metacognition is happening

Metacognition and self-regulation sits at the top of the EEF Teaching and Learning Toolkit, with a potential impact of +7 months. For schools that implement strategies to improve metacognition and self-regulation, they want to see evidence of it taking place. Other than increased performance, it is difficult to catch metacognition in action. So how can we check that pupils are acting ​‘metacognitively’?

Two ways to do this, identified by Dent and Koenka (2015), are traces and think aloud protocols.

Traces are ​‘observable metacognitive strategies used by pupils while completing a task, such as underlining a passage or making notes’. (From the EEF’s Metacognition and Self-regulated Learning guidance report).

For example, there may be certain behaviours in answering an exam question that indicate metacognitive strategies being used. In a lesson you could model annotation of the task using a visualiser, making explicit the processes that are often implicit. Then pupils will practise independently. When they take the exam, this annotation will be replicated. Not only does it act as a process for them to follow, but the annotations are a very clear way of checking that they are following processes.

Some of these traces might not be explicitly taught. A pupil who crosses out a word and replaces it with another, or who goes back and changes an answer, is leaving a ​‘trace’. And this may even be a more interesting manifestation, as they are self-regulating, rather than simply executing a taught process.

Finding traces can be seen as an objective way of checking for metacognition as this does not rely on self-report. However, not all process can be made explicit. And there is a danger of adding a step merely to make the metacognition explicit, which may add a complication that makes things less effective. For example, if you asked pupils to underline any word that they had to work out from context in order to highlight where this is happening, then you may interfere with reading fluency.

Because metacognition is highly context-specific, we should be looking for traces that are context-specific too, rather than general strategies. We always recommend a grid like the one below to frame what you might be looking for in a particular context. Ask, e.g. if we were looking for a trace that a pupil was using a monitoring strategy to keep themselves on track, what would it look like? And the answer would be context specific.

Because metacognition is not always a tangible action, and does not always leave those traces, you need other ways to make the thinking process visible. One way is to have pupils articulate what they are thinking.

Much in the same way a teacher would talk through their thought processes when modelling, in a think-aloud protocol we ask pupils to articulate their thoughts: ​“I’m answering questions 5 first because I find that easier.”; ​“I see that it is a probability question.”; ​“There are only two tennis rackets, so we will need to take turns.”

Pupils may not be able to articulate all thoughts. Sentence stems might help them to get started, and this can also be supported with prompts from the teacher: ​“Why did you start with that question?”; ​“How can you try to work out what that word means?”

Asking pupils to articulate their thoughts may interfere with task performance. It might stop them from completing successfully, or the fact that we ask these may act as a metacognitive prompt and promote these behaviours rather than check they are happening. We might wait until afterwards to ask these pupils in a structured interview, accepting that this also has its challenges and limitations.

It’s not easy to see metacognition – the best academic minds find it hard too. But with these approaches we may be in a far better position that simply inferring from outcomes.

Dent, Amy L. and Alison C. Koenka.​“The Relation Between Self-Regulated Learning and Academic Achievement Across Childhood and Adolescence: A Meta-Analysis.” Educational Psychology Review 28 (2016): 425 – 474.