When learning feels right but goes wrong

I just contacted Henry (Roddy) Roediger to write an endorsement for my new book I wrote with Carl Hendrick, Jim Heal about Instructional Illusions (coming at the end of the summer). The original idea that we had was to write a short book on teaching (and learning) paradoxes. Boy, am I glad we changed course because he sent me a chapter that he had written in 2011 with Andrew Butler on the Paradoxes of Learning and Memory.

As teachers, we often rely on intuition to guide how we help students learn. We know what feels effective, what seems to work in the moment, and what looks like engagement. But what if our instincts are misleading us? What if the strategies that feel good don’t actually support long-term learning, and what if our students are just as misled?

In that chapter, they ask ‘Why do our memories and learning habits so often lead us astray—even when we’re sure we’re right? To answer that question, they explore twelve surprising ways our minds fool us, revealing that what feels true or effective is often the opposite of what works. And these paradoxes aren’t just academic curiosities—they cut to the core of how students make sense of their learning experiences and how teachers structure effective instruction.

Rather than just giving a list of the twelve with a sentence or two on each, I’ve attempted to bundle them into themes.

Memory isn’t a recorder, it’s a storyteller

We tend to think of memory as a kind of mental recording device—a reliable trace of what we’ve seen, heard, and done like a photo or a video- or an audio recording. But psychological research paints a very different picture. Memory isn’t reproductive, but reconstructive. It doesn’t just store facts; it actively reshapes them every time we retrieve them. Carl and I spoke about this in How Learning Happens when we handled episodic and semantic memories.

This explains the unsettling reality that people can form vivid, confident memories of events that actually never happened. One striking example comes from Jean Piaget, who remembered an attempted kidnapping from his childhood in great detail—until he discovered that the entire story had been invented by his nanny. False memories like this aren’t rare exceptions. They’re baked into how memory works.

In the classroom, this means students can be genuinely convinced they “remember” hearing something in a lesson that was never said. And teachers, too, can fall prey to the illusion that they’ve taught something clearly when, in reality, they’ve only implied it or covered it too quickly.

I experienced this when I was doing research for my PhD thesis. I was the instructional designer of a lab-practical course aimed at helping students acquire the complex cognitive skill of carrying out a scientific experiment in the natural sciences. In one of the studies, I asked the content experts who helped design the course to tell me whether certain skills and subskills were included in the course and to what extent. I did the same with students taking the course. As expected, there were many skills that both populations said were either present or absent. Also. there were skills that the experts said were included but which the students didn’t ‘see’. But most interesting was that there were skills that the experts swore were not included in the course which the students were sure were in there!

Learning that feels easy can be deceptive

One of the most important takeaways from their chapter is: what feels like effective learning often isn’t. When students reread notes or highlight text, it feels familiar and fluent. But that fluency is deceptive. We confuse ease of processing with depth of understanding.

Roediger and Butler discuss the concept of desirable difficulties -the counterintuitive idea that making learning harder in the short term often leads to better retention in the long term. Techniques like spacing, interleaving, retrieval practice, and varied problem formats can all slow down performance during learning, but lead to stronger memory traces later on.

This goes against the grain of much classroom experience, where both students and teachers tend to favour smooth, flowing lessons and immediate success (i.e., performance on a quiz or test). But real learning requires struggle; cognitive struggle. A well-timed pause to retrieve information, a challenge that requires thought, a bit of forgetting before review—all of these can strengthen memory more than a perfect explanation.

Our knowledge can blind us

One of the most insidious paradoxes described is what’s called the curse of knowledge. The more you know about a subject, the harder it becomes to imagine what it’s like not to know it. I have likened this, at times, to a butterfly who has forgotten what it was like to be a caterpillar. This is a trap that expert teachers fall into all the time. Concepts that seem obvious to us – because we’ve internalised the background knowledge – may be opaque to our students. The explicit small steps that the learner often needs to solve a problem are bunched into one large step for the expert that at the same time are implicitly carried out.

It’s not just a matter of slowing down or giving more examples. It requires a deliberate effort to see the material through the eyes of the novice, and to scaffold learning in ways that gradually build understanding from the ground up. If we assume too much, we risk losing students without realising it. This is where instructional designers often carry out cognitive task analyses (CTAs) of expert behaviour. CTA is a method used to break down and examine the mental processes experts use to perform complex tasks. It identifies the knowledge, thought processes, and decision-making steps that are often invisible or automatic, making them easier to teach and learn.

Interference: When learning collides with itself

Another category of paradoxes has to do with interference – the way memories can overwrite or block each other. Students may struggle to recall new learning because older, similar information gets in the way This is known as proactive interference. This happens when old knowledge interferes with learning new information. For example, a student who learned how to spell in Dutch may keep spelling English words phonetically (e.g., writing “fone” instead of “phone”) because Dutch spelling rules are interfering with their English learning. Another example is someone who is learning to program in an object oriented computer language may be hindered by earlier learning of a declarative language. Or they may forget something they previously knew after encountering conflicting material, a phenomenon known as retroactive interference. This happens when new learning interferes with the recall of old information. For example, after learning French for several months, a student starts Spanish lessons. Now, when trying to speak French, the student accidentally uses Spanish words – like saying ‘pero’ instead of ‘mais’ for ‘but’, The newly learned Spanish is interfering with their older French knowledge.

In the classroom, this can happen when subjects overlap in confusing ways, or when misconceptions persist despite new teaching. Even peer conversations can contaminate memory. Research has shown that people adopt each other’s memory errors just by discussing what they saw or learned. Social contagion is real, and in classrooms it can undermine clarity and precision.

Imagination, fluency, and the illusion of knowing

Sometimes, students confuse what they imagined with what they actually learned. Simply imagining performing an action can, over time, lead people to believe they actually did it. This blurring between imagination and memory is called imagination inflation. For example, if students are asked a couple of times to imagine what it would have been like to go on a school trip to Rome – even though they never went – they may later ‘remember’ specific details (e.g., eating gelato at the Trevi Fountain) as if they truly experienced them. It reminds us that memory isn’t just fragile, it’s creative, and not always in ways we intend.

Similarly, fluency – how easily something comes to mind – can mislead us into thinking it’s true or important. If a student has just reviewed something, they’ll feel confident about it, but that confidence may evaporate a week later. What we actually do when reviewing something is saying to ourself: “Oh yeah, that looks familiar. I remember reading it.” But that’s quite different from being able to recall the information at a later time. On the other hand, if something is hard to recall, we tend to assume it wasn’t learned well, yet, paradoxically, this struggle often signals a deeper learning process.

Even we as teachers aren’t immune. In hindsight, we often convince ourselves we “knew it all along” when we reflect on outcomes. But hindsight bias obscures the real learning process—and makes us less able to guide students through it effectively.

So, what should teachers do?

These paradoxes might seem discouraging at first. They show us that our minds are fallible, and that our students’ learning is full of traps. But in truth, they offer a roadmap for more intentional teaching. Rather than relying on what feels intuitive, we can embrace what research tells us really works—even if it’s messier, harder, or initially slower. We can:

• design our lessons to include retrieval practice and spacing, even if it means revisiting old material instead of pressing forward;
• help our students reflect on what they know and how they know it, nurturing metacognitive awareness rather than overconfidence;
• build challenge and variability into our lessons, trusting that effortful learning pays off; and
• recognise the limits of our own knowledge and perspective, seek feedback and watch closely for signs of misunderstanding.

Above all, we can teach our students not just what to learn, but how learning itself works and how they can optimally do it. The more we surface these paradoxes in our teaching, the better equipped our learners will be to navigate them.

Roediger, H. L., & Butler, A. C. (2011). Paradoxes of learning and memory. The Paradoxical Brain. Cambridge University Press, (pp. 151-176).