How accurate are exam results as a measure of learning?

In the past, Ofqual has conceded that results in the UK are only reliable “to one grade either way”. That might sound pretty good, but the implications for a student of being awarded a grade B at A level instead of the A they needed to get that place at Cambridge is potentially life-changing.

While many agree that exams are the fairest way to assess learning, we know that the system is not perfect.

But what if there was another method that we could use to predict students’ learning in a more accurate way than just looking at test scores? 

Research led by neuroscientists at Georgetown University, in Washington DC, suggests that there is - and it involves scanning pupils’ brains in order to observe the neurological changes that happen as they learn. 

In the new study, which adds to a growing body of evidence in this area, academics found that observing brain changes in students could “predict learning better than traditional tests”, providing strong evidence that “the inside view afforded by neuroscience can give educators insights about far-transfer learning that they have long sought but that traditional learning assessments often miss”. 

So, just what can a brain scan show us about students’ learning?

Adam Green, a professor in the Department of Psychology and Neuroscience at Georgetown and senior author of the research, explains that we have long been able to “see” learning in the brain and that there are several measures we can use here.

“You can see this in very simple organisms, you can see learning in terms of the strengthening of connections between neurons - and then scaling that up to human brains, you’re more or less looking at the same mechanism,” he says. In fact, he adds, there are “lots of ways that we can look at learning in the brain.” 

However, what has been less well-researched up to this point is the extent to which science can identify changes in the brain that are associated with learning a particular kind of concept, specifically in school. 

Instead, Green says, neuroscientists have often focussed on work in the lab that might involve only scanning the brain on one occasion, based on an activity conceived for the research in question - which, he adds, is “not the way most people learn”.

Green and his colleagues wanted to help to close that gap between what happens in the lab and what happens in the classroom.

“We wanted to look at the brain ahead of somebody taking a real class, and then look to predict - based on various neural characteristics - whether they’re going to struggle with reading, or if they’re going to need more remedial help with maths,” he explains.

In order to get a measurement of learning, the researchers knew that they needed to observe the brain at multiple time points. “Because learning is a change, right? So a longitudinal approach that looks at the brain both before and after, is really critical to actually measuring that change,” says Green. 

Read more:

• Why explicit instruction matters, according to neuroscience

• How neuroscience can help struggling pupils

• How understanding the brain can help your teaching

The research took place in five public high schools in the state of Virginia. Green and his co-researchers looked at the extent to which a curriculum that emphasised spatial thinking - the process of visualising information in the mind’s eye, or laying information out in an image - improved verbal reasoning skills. The research team also used brain imaging to detect the changes that came with learning a specific curriculum in the classroom and used these brain changes to compare the effects of different curricula.

When the researchers tested verbal reasoning, they found marked improvements in the students who had followed the curriculum that emphasised spatial thinking - and the better students got at spatial thinking, the more their verbal reasoning improved. 

Meanwhile, MRI scans showed that changes in students’ brains throughout the course predicted just how well a student who learned to think spatially would transfer that learning to improved verbal reasoning. Indeed, they found that the brain changes more accurately predicted how well learning would carry over than existing measures.

“They were better predictors of learning transfer than grades and tests and all the traditional things we can look at,” Green explains. “Also, the better kids got at thinking spatially, the more they improved at verbal reasoning - and that effect of learning to think spatially on improvement in verbal reasoning was mediated, was statistically explained, by these changes in spatial regions of the brain that focus on spatial cognition. 

“So it really does seem to be that these changes in the brain are explaining why learning to think spatially leads to better verbal reasoning.”

But what, exactly, can teachers take from these findings? While they obviously won’t have access to the regular brain scans, the research does demonstrate clear benefits from this approach - so is there an easy way for them to implement this?

“Yes, and no,” says Green. “Neuroscience is a useful tool, but it’s not a scalable tool. So, the neuroscience itself doesn’t scale up to testing every kid everywhere - that would be crazy, and it’s not possible.” For a start, an MRI scan in a UK hospital can cost hundreds of pounds. 

“But what can scale up are ways of teaching the curriculum, right? So if you can use the neuroscience to identify the ways of teaching that are going to lead to the most transferable learning, then you can scale up those ways of teaching, you can scale up those curricula,” Green adds.

For instance, he continues, this study has shown the transferable benefits of focusing on teaching spatial abilities, and therefore suggests it could be worthwhile encouraging students to think in spatial terms, in any subject. He gives the example of getting students to visualise the different characters in a book in terms of their emotional distance to each other in English, or producing an event timeline in history.

“The more that we can build in these spatial resources…I would posit that we would see at least some of the gains that we’ve been observing in the curriculum we focussed on,” he says. 

And while this study focused specifically on measuring the benefits of developing spatial awareness, future studies could look at the effects of other teaching approaches or types of curricula.

In fact, wider research is already underway. In 2021, researchers from the University of Princeton published a study that used brain scans to determine how far “neural alignment” - that is the degree to which developing neural representations formed during learning match up to the representations in the brains of experts and other learners in the class - predicts exam performance.

One of the findings was that “alignment among students successfully predicts overall performance in a final exam” - in other words, the more closely students were aligned with their peers on the concepts they were learning, the better.

In a University of Princeton blog, Meir Meshulam, a postdoctoral research associate at the Princeton Neuroscience Institute, and lead author of the paper, suggested that activities that encourage students to communicate with one another about learning, allowing them to become more “aligned”, could therefore be helpful.

“I would definitely encourage students to work in groups, to talk to each other, to communicate as much as possible, because the class as a whole is a good thing to be aligned with,” he said. 

It’s important to note, though, that these are just two studies, and more research is needed to draw firm conclusions about what works.

While we can see learning in the brain, and use those observations to predict student performance - potentially in a more accurate way than looking at previous exam results - that technology is unlikely to be rolled out across the education system any time soon.  

However, as the studies described in this article show, we are only just beginning to scratch the surface of the possibilities that further research in this space can bring.

Using brain scans to determine which teaching approaches work best, and in which contexts, is just one avenue worth exploring.