How We Learn (10 page)

Shereshevsky’s recall of words, numbers, and voices was so complete, in fact, that often one performance encroached on another performance, especially when they occurred in the same place, with no difference in context. He had to work to block related material. “Writing something down means I’ll know I won’t have to remember it,” he told Luria. “So I started doing this with small matters like phone numbers, last names, errands of one sort or another. But I got nowhere, for in my mind I continued to
see what I’ve written.” He lacked a normal forgetting filter, and it often frustrated him.

Luria had Shereshevsky study one of his number-letter matrices on May 10, 1939. S. examined it for three minutes. After a short break, he could recite it without error, row by row, column by column, or along the diagonals. Several months later, Luria tested him again—without warning—on the same table. “The only difference in the two performances was that for the latter one he needed more
time to ‘revive’ the entire situation in which the experiment had originally been carried out,” Luria wrote. “To ‘see’ the room in which we had been sitting; to ‘hear’ my voice; to ‘reproduce’
an image of himself looking at the board.” S. reinhabited the May 10 study session to bring back the matrix.

Shereshevsky was a prodigy, and his methods are largely off-limits to the rest of us. We can’t revive our study surroundings in nearly so much detail, and even if we could, there’s no chance that the entire table would scroll back up in pristine clarity. Our minds don’t work in the same way. Yet S.’s use of multiple perceptions—audio, visual, sensual—hints at how we can capitalize on context. We can easily multiply the number of perceptions connected to a given memory—most simply, by varying
where
we study.

How much could a simple change in venue aid recall?

In the mid-1970s, a trio of psychologists performed an experiment
to answer that question. Steven Smith, Robert Bjork, and another psychologist, Arthur Glenberg, all then at the University of Michigan, wondered what would happen if people studied the same material twice, only in two different places. They presented a group of students with a list of forty four-letter words, like “ball” and “fork.” Half the students studied the words in two ten-minute sessions, a few hours apart, either in the same small, cluttered, basement room or in a neat windowed room looking out on a courtyard. The other half studied the words in two settings: once in that small, windowless room and again in the neat windowed one overlooking the courtyard. Two groups. The same words. In the same order. The same amount of time. One group in the same environment both times, the other in two distinct ones.

“I considered myself, the experimenter, part of the environment, too,” Smith told me. “In the windowless, basement room I looked like I usually did, long wild hair, flannel shirt, construction boots. In the modern conference room, I had my hair slicked back, I wore a tie, I had on the suit my dad wore to my bar mitzvah.
Some of the students who studied in both places thought I was a different guy.”

After the second session the students rated each word on whether it evoked positive or negative associations. This was a ruse, to give them the impression that they were done with those words, that there was no reason to think about or practice them. In fact, they weren’t done. In the third phase of the experiment, three hours later, researchers had the students write down as many of the words as they could in ten minutes. This test occurred in a third, “neutral” room, a regular classroom. There was no reinstatement, as in previous context studies. The third room was one that the participants
hadn’t been in before
and was nothing like the other two where they had studied.

The difference in scores was striking. The one-room group recalled an average of sixteen of the forty studied words. The two-rooms group recalled twenty-four. A simple change in venue improved retrieval strength (memory) by 40 percent. Or, as the authors put it, the experiment “showed strong recall improvements with variation of environmental context.”

No one knows for sure why changing rooms could be better for recall than staying put. One possibility is that the brain encodes one subset of the words in one room, and a slightly different set in the other. Those two subsets overlap, and two subsets are better than one. Or it may be that rehearsing in two rooms doubles the number of contextual cues linked to any single word, fact, or idea being studied. In one room, the beige walls, fluorescent lighting, and clutter of stacked books color the memory of the word “fork”; in the other, “fork” is intertwined with the natural light pouring through the window, the sight of an old oak in the courtyard, the hum of an air conditioner. The material is embedded in two sensory layers, and that could give the brain at least one more opportunity to “revive” what it can of the study conditions and retrieve the words, or concepts. If Door Number 1 doesn’t work, it can try Door Number 2. We do this sort of perspective shifting all the time when, say, trying
to remember the name of an actor. We pull up scenes from his most recent movie: There’s his face, but no name. We recall his face in the newspaper, his cameo on a TV show, maybe even a time we saw him onstage. We use multiple mental lenses to tease out the name and, in general, more detail.

Smith
has since gone digital. He uses short video clips to create backgrounds, rather than herding students from room to room. In a typical experiment, he divides participants into two groups. One studies, say, twenty words in Swahili over five practice sessions of ten minutes each. The words appear on a movie screen, one at a time, transposed over a single, soundless background clip in all five sessions (of a train station, for example). This is the “same environment” condition. The other group studies the identical words, also over five ten-minute sessions, only those words appear over a different video background during each practice period (rainstorm, train station, desert scene, traffic jam, living room). A visual simulation, no more. Yet on tests taken two days later, the varied background group came out ahead, remembering an average of sixteen of the Swahili words, compared to nine or ten for the one-background group.

I have to admit I’m a sucker for this stuff. I love studies like these, because I can’t sit still for more than twenty minutes to study, if that. I want to believe that this kind of restlessness can deepen learning, and I often wish that the evidence for context variation was a little more … airtight.

The research has a meandering feel to it, to be honest. Scientists are still debating which cues matter most, when, and how strong they really are. Because context effects are subtle, they’re hard to reproduce in experiments. The definition of “context,” for that matter, is a moving target. If it includes moods, movement, and background music, it could by extension mean
any
change in the way we engage our vocabulary lists, history chapters, or Spanish homework. Think about it. Writing notes by hand is one kind of activity; typing them using a keyboard is another. The same goes for studying while standing
up versus sitting down, versus running on a treadmill. Daniel Willingham, a leading authority on the application of learning techniques in classrooms, advises his own students, when they’re reviewing for an exam, not to work straight from their notes. “I tell them to put the notes aside and create an entirely new outline, reorganizing the material,” he told me. “It forces you to think about the material again, and in a different way.”

Isn’t
how
we do something part of the “environment,” too?

It is. Yet the larger message of context research is that, in the end, it doesn’t much matter which aspects of the environment you vary, so long as you vary what you can. The philosopher John Locke once described the case of a man who had learned to dance by practicing according to a strict ritual, always in the same room, which contained an old trunk. Unfortunately, wrote Locke, “the idea of this remarkable piece of household stuff had so mixed itself with the turns and steps of all his dances, that though in that chamber he could dance excellently well, yet it was only when that trunk was there; he could not perform well in any other place unless that or some other trunk had its
due position in the room.”

This research says, take the trunk out of the room. Since we cannot predict the context in which we’ll have to perform, we’re better off varying the circumstances in which we prepare. We need to handle life’s pop quizzes, its spontaneous pickup games and jam sessions, and the traditional advice to establish a strict practice routine is no way to do so. On the contrary: Try another room altogether. Another time of day. Take the guitar outside, into the park, into the woods. Change cafés. Switch practice courts. Put on blues instead of classical. Each alteration of the routine further enriches the skills being rehearsed, making them sharper and more accessible for a longer period of time. This kind of experimenting itself reinforces learning, and makes what you know increasingly independent of your surroundings.

Chapter Four

Spacing Out

The Advantage of Breaking Up Study Time

The oldest learning technique in memory science is also one of the most powerful, reliable, and easy to use. Psychologists have known about it for more than a hundred years and proven that it works to deepen the learning of subject areas or skills that call for rote memorization, like foreign vocabulary, scientific terms and concepts, equations, or musical scales. Yet mainstream education has largely ignored it. Few schools teach it as part of the regular curriculum. Few students even know about it, except as the sort of motherly advice that’s safe to ignore:

“Honey, don’t you think it would be better to study for a little bit tonight and a little bit tomorrow, rather than trying to learn everything at once?”

The technique is called distributed learning or, more commonly, the spacing effect. People learn at least as much, and retain it much longer, when they distribute—or “space”—their study time than
when they concentrate it. Mom’s right, it is better to do a little today and a little tomorrow rather than everything at once. Not just better, a
lot
better. Distributed learning, in certain situations, can double the amount we remember later on.

This isn’t to say that cramming is useless. The all-nighter is timetested, with a long track record of improving exam scores the next day. In terms of reliability, though, this nocturnal sprint is a little like overstuffing a cheap suitcase: the contents hold for a while, then everything falls out. Researchers who study learning say the result from habitual cramming can be dramatic from one semester to the next. The students who do it “arrive for the second term, and they don’t remember anything from the first term,” Henry Roediger III, a psychologist at Washington University in St. Louis, told me. “It’s like they never took the class.”

The spacing effect is especially useful for memorizing new material. Try it yourself with two lists of, say, fifteen phone numbers or Russian vocabulary words. Study one list for ten minutes today and ten minutes tomorrow, and the other for twenty minutes tomorrow. Wait a week and test yourself to see how many of the total from both lists you can remember. Now go back to the two lists: The difference in what you recalled from each should be significant, and there’s no obvious explanation for it. I like to think of the spacing effect in terms of lawn care in Los Angeles. L.A. is a city with a coastal desert climate and cultural commitment to the pristine lawn. I learned while living there for seven years that, to maintain one of those, it’s far more effective to water for thirty minutes three times a week than for an hour and a half once a week. Flooding the lawn makes it look slightly more lush the next day, but that emerald gloss fades, sure enough. A healthy dose every couple days and you can look your neighbors in the eye, while using the same amount of water—or even less. Same goes for distributed learning. You’re not spending any more time. You’re not working any harder. But you remember more for longer.

A principle this powerful should have had a quick, clean ride from the lab into classrooms. What student wouldn’t want to enhance learning without putting in any extra time or effort?

It hasn’t happened, and for good reasons. One is that, as parents know too well, it’s enough of a chore to get students to sit down for single study sessions, never mind multiple ones. The other is that for much of the last hundred years psychologists have—exasperatingly, inexplicably—confined the study of spacing to short lab experiments. It is as if doctors discovered a cure for diabetes and spent fifty years characterizing its molecular structure before giving it to a patient. Only in the last several years have researchers mapped out the best intervals to use when spacing study time. Is it more efficient to study a little bit today and a little bit tomorrow, or to do so every other day, or once a week? What if it’s Tuesday, and the history final is on Friday? What if the exam is a month away? Do the spacing intervals change depending on the exam date?