How much can you remember from your high school classes? If I gave you one of your final exams today, could you still pass it?
In their now-famous study [1], Harry Bahrick and Lynda Hall asked this question of nearly two thousand people.
Unsurprisingly, they found what most of us expect: a few years out from our classes, we forget much of what we’ve learned. Wait long enough, and our performance barely exceeds someone who has never taken the class.
But there was one group Bahrick and Hall identified that didn’t forget. For this group learning wasn’t just durable—it was permanent. Even fifty years later, without any additional rehearsal or practice, they managed to maintain almost exactly the same test performance they had from the start.
Forgetting is the Default
How did some people manage to retain knowledge for decades? And why do so many of us forget?
Before we look at the factor driving near-perfect memory performance in one of the groups Bahrick and Hall studied, we need to start by recapping how normal forgetting works.
Because, it should be stressed, forgetting is the default.
Forgetting was one of the first subjects of experimental psychology. In the early 1880s, German psychologist Hermann Ebbinghaus conducted extensive tests [2] on his own ability to remember.
Being a meticulous scientist, he was worried about his results being contaminated by information he encountered in his day-to-day life. To test this likelihood, he made vast lists of thousands of nonsense syllables and set himself the task of memorizing them and seeing how much time it took to relearn them after varying intervals.
It’s from Ebbinghaus’s original work that we have the forgetting curve. This curve plots the decay of recall over time, showing that forgetting is steepest right after we learn something. After that, the strength of the memory flattens out; we continue to forget, but at a less rapid rate than before.
Now a problem with Ebbinghaus’s study, and with most psychological research into memory that’s been conducted since, is that the intervals he tested for forgetting are pretty short. Ebbinghaus’s longest interval for relearning was just one month.
Yet for subjects we learn in school, we’d like to be able to retain them much longer than that. Ideally, at least, an education would last a lifetime.
How Do We Remember?
So, one hundred years after Ebbinghaus, Bahrick and Hall decided to investigate exactly that. They recruited nearly two thousand people of varying ages. They gathered some basic biographical information, then they gave all the participants a test in high-school algebra or geometry. Finally, the researchers compared the participants’ scores with archival records of their performance when they were in school.
From these results, Bahrick and Hall calculated a forgetting curve: how much did people remember from classes they took, whether they took them last year or half a century ago?
At first glance, the data from this paper seem to show a curve with forgetting accelerating over time, rather than Ebbinghaus’s classic decay curve with most of the forgetting happening at the beginning.1 [3]
But we need to note that Bahrick and Hall graphed the x-axis logarithmically. If we were to redraw Bahrick and Hall’s graph without that transformation, we would get the same shape for the forgetting curve Ebbinghaus originally proposed.
One possibility might be that those who learned the math better initially would likely remember more than those who had lower math grades to start with.
As we can see in the figure above, this holds true. People who did better in their courses also did better recalling that information years later. Indeed, the relative rankings of the A students, B students and C students are remarkably stable even over very long intervals.
But even though better students remembered more, they still forgot at the same rate. Being a better student didn’t spare those from the forgetting curve.
Now let’s look at another group. This group, even fifty years after taking algebra, barely declined in their performance on the test. Who were they? People who had taken calculus.
And this relationship held when the researchers controlled for the amount of exposure to algebra through work and hobbies. In other words, it’s not simply that people who study calculus all become scientists or engineers, thus keeping their math fresh.
What is it about learning calculus that allowed those participants to defy the usual laws of forgetting?
Make Learning Stick
For hints as to the cause, we must go all the way back to Ebbinghaus’s original studies. In addition to his contribution to our understanding of the forgetting curve, Ebbinghaus also tried various experimental manipulations to see how they would affect his relearning speed.
One of these interventions was overlearning. Normally, he would practice his nonsense syllables until he could recall them perfectly. When he instead continued to practice beyond perfect—continuing to practice even after he could successfully recall them all—the extra effort wasn’t wasted. Overlearning doesn’t boost performance—at that point you’re already performing the task perfectly—but it does boost retention.
Psychology has since amply demonstrated [4] the principle of overlearning. Extra practice beyond what’s needed for mastery makes learning more durable.
This is a useful technique whenever you need to learn something deeply but are unlikely to use it regularly. For instance, overlearning is a good strategy for a skill like CPR. You probably won’t ever need it, but you don’t want your skills to be rusty in an emergency. Thus, a training session that has you practicing much more than necessary to pass the test isn’t wasteful—it might be the difference between life and death!
Of course, overlearning is exactly what occurs when you study calculus. The basic operations of algebra are embedded in calculus. To do calculus, you have to do a lot of algebra. This results in additional practice of algebra techniques, far beyond what’s needed to perform it adequately, thus making that knowledge last longer.
The second effect Ebbinghaus observed was spacing. If the practice sessions for a skill were spread out over time, the effort needed to recall the information was less than if the same amount of practice was done back-to-back. Ten sessions spread out over time led to more durable knowledge than ten sessions crammed together.
Bahrick and Hall note that their experiment couldn’t detect whether the calculus group’s lasting knowledge of algebra was due more to overlearning or spaced practice. Taking advanced mathematics courses does force overlearning, but because the calculus classes were typically taken anywhere from six months to a few years after the initial algebra class, spacing was leveraged as well.
All of this suggests a simple method for making learning permanent: if you want skills and knowledge to last, study beyond the level you think you’ll need long-term. Even if you don’t need advanced skills, overlearning and spaced practice will ensure your knowledge of the basics lasts much longer.
Overlearning and spacing are just two keys to more durable memory, but there’s far more. I’ve put together a guide [5] that explains the science of how memory works as well as what are the steps you can take to improve yours.