Last week I attended an in-house training session led by a number of my colleagues in the MFL Department including Louise Miller, Dr. Gianfranco Conti and Dylan Viñales.The focus was on developing students’ metacognitive skills. I was particularly taken by an activity which they suggested that they called ‘Think Aloud’.

On paper this was identical to a Kagan ‘Rally Coach’ activity that I have seen elsewhere numerous times before. Students work in pairs and talk through a particular question while their partner quizzes them on what they are doing. This always felt like something of a gimmick in the past- a case of getting students to talk for its own sake or perhaps because it checks a box for an observation. At best I have heard some generic comment about how somehow, *as if by magic*, student learning is enhanced by working cooperatively with no explanation of the mechanism behind this .

It was refreshing to hear some research and rationale justifying this strategy (that it improves metacognition) and some advice as to how best to implement it. Modelling a **constant stream** of discussion worked well, but what was most useful was asking students to model the sorts of ‘coaching questions’ they could ask as I made deliberate mistakes on the board. Initial comments such as ‘you’ve got it wrong there sir’ or ‘you need to square root at the end’ were corrected and improvements such as ‘have you considered what is happening between lines three and four?’ or ‘what have you found here, does it make sense?’, ‘could a different diagram help you here’ quickly came to the fore.

Students then worked through a fairly difficulty Pythagoras’ Theorem problem set that I had put together for them. I circulated listening for good coaching questions and periodically asking students to share with the class good questions they had been asked.

Questions courtesy of http://corbettmaths.com/

Problems with the questions included students ‘getting lazy’ and neglecting to talk their way through the problem (vigilance is key here) and students not speaking when they become unsure of how to proceed (convince them that just saying exactly what they are thinking is a good strategy and often leads to a ‘lightbulb moment’ of how to proceed.)

Did the students find completing the problem set any easier using these strategies? Probably not. Were the students stretched in other ways and becoming more aware of the ‘voice in their head’- probably. Were the students being forced to think like a mathematician to frame questions that supported students but didn’t involve just telling them the answers? Definitely. Indeed it was refreshing, if a little spooky, to notice that some of the questions students asked each other sounded remarkably like something I would ask.

The key with any metacognitive improvement strategy is progressing slowly over time. Rome wasn’t built in a day and effective metacognition is not developed in one lesson (or even one year- it’s a lifetime skill). Crucially a proportion of the training session focused on a neglected aspect of metacognitive strategies- that of developing automation. In Maths terms, this should include students reviewing their own solutions, drawing diagrams and ‘sense checking’ answers **without prompting **(though I think I could soon end up out of a job were this to become the norm!).

To recap:

- Model, model and model again. This was the most important part of the process.
- Encourage students to share good examples periodically.
- Be aware of pitfalls, particularly student reluctance to speak when struggling- nip these in the bud by modelling what to do in these situations.
- Don’t expect improvements to happen straight away, progress in any aspect of learning is not predictable or linear (or even monotonic) and indeed
**should not be rapid**despite what many observers and school leaders might have you believe. Metacognition is no different

In some cases students may find completing the problems *more* difficult using these strategies. Good. Difficulty in learning is beneficial for students and is an indicator that they are being stretched.

One further point- a good proportion of the work I do at my school is with students to students applying to Oxbridge. Oxbridge interviews require students to talk *a lot*. Indeed, any interview for any mathematics or science interview will almost certainly involve students having to talk their way through a number of problems. Many impressive mathematics students lack the ability to do this. Like anything, practice is crucial and activities aimed at developing metacognition like this one are an excellent way of developing students for these intellectual challenges.