‘Mechanistic’ metacognition: the key to student agency?

Nearly a decade ago, I was asked to develop a study skills programme for a group of secondary schools in Australia. The aim of this programme (as you can guess) was to ensure students were well-versed in the most effective and efficient study strategies supported by research.

Following the implementation of this short course, we gave students a survey exploring whether or not they were now able to differentiate between empirically strong study strategies (eg, note summarisation) and empirically weak ones (eg, rereading notes).

We found that 90 per cent of the students were able to successfully differentiate between strong and weak revision techniques.

That was the good news.

The bad news? When we tracked individual study habits throughout the year, only 30 per cent of students employed those techniques previously identified as effective.

Furthermore, and perhaps unsurprisingly, those students who did implement said techniques were invariably among the top academic performers by the end of the year.

So we had a clear problem: why was it that, when nine out of 10 of these secondary students knew what strategies they should be using, only three out of 10 actually used those strategies?

When I finally solved this problem, it significantly changed how I understood the concept of “agency” and what was required for students to meaningfully own their own learning.

Student agency: what does it actually mean?

It’s been well established that students in most English-speaking countries are less engaged with school today than the students of previous generations. And in a bid to reinvigorate engagement, many schools are turning toward the concept of agency - most commonly conceptualised as allowing students “choice” over their learning practices.

The idea is fairly simple: if individuals lose motivation when they feel as though they do not have control over their actions, then increasing their sense of control via agentic choice should enhance motivation and increase engagement levels.

The academic literature exploring learner choice typically revolves around three levels: content, material and presentation.

• “Content” suggests students should be given a choice over what information to study (eg, they can choose to study mammals or classical music).
• “Material” says students should be given a choice over the medium by which said info is accessed (eg, they can read a book or interview a practising professional).
• “Presentation” says students should be given a choice over how to demonstrate their learning (eg, they can make a video documentary or write an essay).

Research into these levels of choice reveals two very important principles.

First, when students are given too much freedom without relevant guidance or support, they are more likely to become overwhelmed and actually disengage more from learning.

For this reason, it may be worthwhile allowing choice across only one of the aforementioned levels (eg, students can select a topic to study, but they all must do their research using primary-source library books and demonstrate their learning via a structured poster).

Limiting choice to a single level should help students to avoid becoming overburdened, while simultaneously allowing teachers to maintain an effective academic trajectory.

Second, when students are allowed absolute choice within a particular level, many opt to pursue only those things they are already well versed in (eg, the football player will always choose to study football while the filmmaker will always choose to make films).

Accordingly, rather than giving free choice, it may be worthwhile embracing structured choice by allowing students to choose from among a set of relevant options.

Furthermore, recent data suggests that the optimal number of options in any domain is four (two options tend to stifle a sense of agency whereas six tend to drive a sense of being overwhelmed).

There is no question that the idea of student choice is incredibly important and should certainly be embraced across education; however, the astute reader will recognise that this brings us little closer to meaningful learner agency.

Allowing a student to study aeroplanes instead of dinosaurs may certainly improve motivation, but this says nothing about what that student needs to do in order to effectively develop her or his understanding of aviation.

Interestingly, the more practical aspects of decision-making with regard to learning practice are rarely discussed by researchers who specialise in agency. Rather, they are discussed by researchers who specialise in a related but qualitatively different field: metacognition.

What is metacognition?

Simply defined, metacognition is thinking about one’s own thinking; it is self-awareness of one’s own cognitive processes while engaged in the pursuit of a particular aim or goal.

Discussions of metacognition within education typically revolve around the concept of active contemplation: can teachers employ timely and relevant questions that encourage students to reflect upon their own thoughts and behaviours during the learning process?

This “metacognitive questioning” is commonly aligned along the dimensions of planning (“What do I already know that can be applied to this assignment?”), monitoring (“Are my current strategies working or is it time to change my approach?”) and evaluation (“If I were the teacher, what would I identify as strengths or weaknesses of this work?”).

Again, although ensuring that students have ample opportunity to interrogate their own thinking is incredibly important and should certainly be embraced across education, these reflections will necessarily be constrained by a student’s prior experiences.

When we ask “What strategies might you use to complete this task?”, the answer will always be confined to those strategies that students have previously been exposed to. We wouldn’t expect anyone to answer “Text-to-question alignment” if they’d never before heard of or employed that technique.

In other words, if we conceive of student agency not merely as choice over what to learn but also choice over learning behaviours, then each student’s choices will always be restricted to his or her prior knowledge.

It is for this reason that many researchers in the field of metacognition argue that educators must explicitly teach study techniques.

If we ensure that students are well versed in varied effective learning strategies, we can expand their decision space and, in turn, drive meaningful engagement.

Unfortunately, this brings us back to the problem outlined at the start of this article: despite explicitly teaching secondary students the method and impact of key study skills - and despite 90 per cent of them being able to differentiate between strong and weak skills - only 30 per cent of students chose to align their agentic choices with those techniques most likely to drive their learning in a positive direction.

After several years of beating my head against a wall trying to figure out the reason behind this 60 per cent discrepancy, my breakthrough came during a dinner conversation with my lawyer father-in-law.

Mechanistic metacognition

Several years back, a group I was working with decided to take a company to court for failure to deliver on a contract.

During our initial legal consultation, the lawyer outlined his primary strategy. After hearing him out, I chimed in to ask what would happen if his strategy failed? Without skipping a beat, the lawyer was able to outline the next two strategies he’d employ if things fell to ruin.

Ever the teacher, I couldn’t leave it there.

I asked the lawyer to expand: why was he opting for those particular strategies in that particular order?

After a moment’s thought, his response was: “Because that’s the way we do it.”

This reminded me of the myriad times I’d asked students why they opted to employ certain learning strategies, only to be confronted with: “Because that’s the way I was taught.”

More from Jared Cooney Horvath:

• Why you’re probably wrong about the science of learning
• The real problem with the learning styles ‘myth’
• Are you teaching your students to be smarter?

That evening, I asked my father-in-law the same question: why would a lawyer opt for these particular strategies in that particular order?

His response was quite different: over 30 minutes, he outlined the inner workings of the court system and explained why certain cases needed to be tried according to a particular pattern. He then went one step further and described several legal loopholes that could be used to skip over certain steps and expedite the entire process.

And herein lies the secret to meaningful student agency.

My father-in-law didn’t merely know what things to do; he understood why he should be doing each. It was this deeper understanding that allowed him to take greater agency over his behaviours (beyond “That’s how I was taught”) and imagine novel strategies that might lead to greater impact.

Ensuring that students have the opportunity for choice and are well-versed in effective learning strategies may be necessary for agency, but it is insufficient.

Without a deeper understanding of why each strategy has the impact that it does, students will either blindly follow explicit instructions or make decisions based on personal intuition (which is rarely well-aligned with effective learning).

I call this concept mechanistic metacognition.

In order for lawyers to make meaningful and effective judicial choices, they must have a strong understanding of why courts function in the manner they do. In order for engineers to make meaningful and effective choices about structure, they must have a strong understanding of why materials interact in the manner they do. In order for students to make meaningful and effective learning choices, they must have a strong understanding of why learning functions in the manner that it does.

How does this work in schools?

With this realisation, I retooled my secondary school programme to focus on “mechanistic metacognition”.

Rather than exploring study skills, we explored the learning process: what differentiates surface learning from deep learning? What is required to transfer skills between contexts? How does memory work? What are the different levels of attention?

With this new focus, varied study strategies became tools for reflection rather than recipes for behaviour.

As an example, in the study skills programme, I taught students to “print it out”. Put simply, any time a text reaches three pages or longer, then human beings will almost always comprehend better and remember more when reading from hard-copy paper as opposed to digital screens.

To be fair, digital texts have several advantage over print (eg, the ability to change font size, the ability to easily search for keywords), but if the goal of a particular reading is to understand and learn relevant material, then print it out.

Remember: 90 per cent of students could repeat this strategy, but only 30 per cent consistently employed it.

Conversely, in the new “mechanistic metacognition” programme, I taught students about spatial memory.

If you’ve ever played the computer game Grand Theft Auto or something similar, then you know that a common element of sandbox-type video games is a real-time map; sitting in the corner of the screen is a continuously updating diagram that reminds the player where he or she is located within the larger game world.

“If we want students to take meaningful control over their own learning, they must first deeply understand how learning works”

Believe it or not, the human brain has a nearly identical system.

Riddled throughout the hippocampus (the brain’s gateway to memory) are millions of tiny structures called place cells. Not only do these cells automatically build a real-time map of the world around us and mark our position within it, but they also subconsciously imprint every new memory we form with a relevant three-dimensional spatial location.

This is why even though most of us never explicitly set out to memorise the location of the stapler, mug or notebooks on our desk, it is immediately apparent when someone has unexpectedly (and, likely, uninvitedly) cleaned up or rearranged our desktop.

Here’s the important bit: books have both a clear and static spatial layout - information presented in print will forever exist in an unambiguous and unchanging three-dimensional location in space.

Digital files, on the other hand, have neither a clear nor static spatial layout; information presented digitally will start at the bottom of the screen, move through the middle, and then disappear out the top.

Only after exploring this specific mechanism will I then ask students, “Can you explain why many learning researchers believe you should read from hard-copy paper?”

With the concept of spatial memory in tow, many students are now able to explain why printing it out would prove such a powerful technique (if spatial layout is an integral and essential part of human memory, then ensuring that learned information has a clear three-dimensional location assists memory formation and later access).

Furthermore, many are also able to explain why techniques like digital cramming and scrolling would be so ineffective (if media does not have a clear three-dimensional location, then an important aspect of memory formation is removed, weakening the ultimately formed memory).

So, did teaching students mechanisms prior to teaching strategies have any impact on agentic choice?

After exploring mechanistic metacognition, the number of students who consistently employed strong study techniques jumped from 30 per cent to 65 per cent; still not the 90 per cent I was hoping for but a significant move in the right direction.

Furthermore, and perhaps more importantly, learner agency (as determined via student surveys) and student engagement (as determined via teacher surveys) both improved significantly.

Finally, academic outcomes in these schools improved significantly as well.

So, what next?

We have covered a ton of ground in this article. If I were to try and sum everything up in a single sentence, what would it be?

Simply put: if we want students to take meaningful control over their own learning, then they must first deeply understand how learning works.

When it’s said that “schools should ensure students have agency over their learning”, this treats learning as a universal, innate skill like breathing or digestion: something everyone can (and does) do without conscious consideration.

Whereas it’s true that very young children are able to engage in a purely experiential form of development, the neurological mechanisms that drive this relatively “automatic” neuronal refinement are largely dismantled and expired by the age of 5.

This means that learning among traditional school-aged children is neither intuitive nor spontaneous; rather, it is an explicit process that must be learned and practised to be mastered, just like any other significant skill.

And this is why mechanistic metacognition is the secret to meaningful student agency.

Only when students are able to verbalise the mechanisms behind why varied strategies and techniques work can we expect them to consistently make meaningful choices that serve to drive their learning in a positive direction.

Jared Cooney Horvath is a neuroscientist, educator and author