| 1 |
| 1 |
| 1 |
| 1 |
 |
| 1 |
 |
|
|
|
| A version of this article appeared in Ed. Magazine, Harvard Graduate School of Education, Fall 2003 (cover story) Thinking Lessons Can sophisticated computer games turn children into better learners? . . . . . . . . . . . . . By David Brittan DAVID STEVENS (Ed.M.’96 Ed.D’00) believes he has seen miracles. Skeptical scientist that he is, he can’t say for sure, even though one of the alleged miracles happened to him. At the start of his senior year of high school in Washington, D.C., a learning disability that had resisted years of special tutoring, posh private schools, and psychiatric visits—a condition so severe that it had driven him to steal exams from teachers and wheedle used essays out of other students just to maintain a D average—simply vanished. “For the first time in my life,” Stevens recalls, “I was writing history papers, I could do the assigned reading, I was coming to class prepared.” He went on to major in philosophy and intellectual history at the American University of Paris, where he completely remade himself: instead of the class clown—his earlier persona—he was now a scholar and an intellectual. Exactly what caused the metamorphosis is a question Stevens ponders to this day. He doesn’t rule out some sort of natural adolescent brain development. “It’s possible that when I turned seventeen it was the right time for me to get serious about school,” he says. But what Stevens thinks happened that year is that Harry Wachs, a child development specialist at George Washington University, made him smarter. “In my case,” Stevens says, “some of the basic cognitive abilities that help kids do well in school, especially visual-spatial abilities, may not have fully developed.” Over a period of months, Wachs gave him exercises to help him to perceive and reproduce patterns, to visualize and mentally rotate objects, to reason logically. The treatment seems to have ratcheted up his thinking skills to the point where reading and understanding a book, or organizing his thoughts on paper, ceased to be an ordeal. After college, he returned to Wachs’s clinic as a therapist. There he witnessed similar transformations in nine- and ten-year-old versions of himself. At 35, Stevens is on a quest to make those miracles—if that is what they were—available to millions of schoolchildren who can’t afford expensive one-on-one therapy. It is a controversial goal. Only a handful of studies lend credence to the idea that so-called cognitive interventions, programs designed to sharpen thinking skills, can actually help children learn better in school. But at Lexia Learning Systems (Lincoln, Massachusetts)—the educational software firm where he is director of advanced research and design—Stevens has distilled what he considers to be the best principles from the most successful programs. And now he and his colleagues are building a cognitive intervention of their own. Unlike other such programs, which are usually administered by experts in cognitive science, this one is a set of computer games. “Everyone knows how engaging and addictive computer games can be,” Stevens explains. “We’re trying to marry those qualities with principles of cognitive science and then get kids addicted to working hard and thinking hard and thinking complexly.” If Stevens’s vision pans out, students from age eight to adult, who may or may not have identifiable learning difficulties, will sit down at a school computer a few times a week and game their way to better grades. And they will do so with minimal adult supervision, and at a cost to the school of no more than five or ten dollars per student. Radical as the plan may sound, Stevens stresses that he is not out to replace teachers. Nor is he designing a substitute for tried-and-true special-ed techniques. He simply wants to fill a gap in the range of options available to teachers and students, he says. “Some children who struggle in school are prescribed ‘more school’—tutoring, after-school classes, resource-room work, or summer school. Other kids are prescribed ‘compensations’ such as untimed tests, sitting closer to the teacher, being asked to review their work, having the material broken down for them. In too many cases, neither more school nor compensations address the underlying cause of the child’s struggle. There needs to be a complementary approach that does.” But good intentions are one thing, good interventions quite another. Stevens must prove that his approach works—that a piece of software can actually propel children to higher levels of thinking and higher academic achievement. And that takes careful research. “To make our case that this is worthwhile, we’ll have to be able to overcome the doubts of teachers and parents,” Stevens says. Above all, Stevens will have to overcome his own doubts. “Doubt is part of what fuels my scientific side to study this question methodically,” he says. “It keeps me rigorous. And if my research isn't rigorous, then it will be dangerous.” DESIGNING SOFTWARE to make kids smarter has been on Stevens’s mind for a good 10 years. In the early 1990s, after making his giant intellectual leap, Stevens spent two years playing with blocks. That was during his stint as a therapist in Wachs’s clinic. Day after day, he would set up colored wooden cubes and parquetry tiles in geometric patterns for children to match or flip or reassemble. “As I sat there doing repetition after repetition, I’d say, ‘Boy, how can we automate this?’ ” Stevens recalls. “It was then that I started to think about software.” At the Harvard Graduate School of Education, in 1996, Stevens thought about software some more. “In David Rose’s class on technology and learning disabilities, the final assignment was to write a proposal to get funded for a technology intervention,” Stevens says. “So I wrote a proposal to develop basically what we’re working on now.” That same year, Time magazine ran an article about software that had been developed at Rutgers University to help children learn to read better by listening to elongated speech sounds. “I saw that and I said, ‘What am I waiting for? They did it for auditory processing. I want to do it for logical reasoning, for visual-spatial skills, for giving and receiving instructions. I want to do it for the whole model of cognitive development.’ And I had the personal experience and the training to do it.” Stevens drew up a business plan and went out in search of funding. Venture capitalists’ eyes lit up, but for the wrong reasons. “When they saw my business plan, which called for all this testing and a long research and development process, they would say, ‘Forget this. All we need is the fact that you’re a Harvard graduate with an idea for improving intelligence and we can go far.’ I wasn’t interested,” Stevens says. The reason he had entered graduate school in the first place was to learn methods for studying the mysterious changes he had observed in himself and in his young patients. Rigorous testing was the heart and soul of Stevens’s plan. Meanwhile, as he surveyed the literature on cognitive interventions, Stevens managed to ease some of his doubts about his basic premise—namely, that intelligence is malleable and can be systematically improved. The first half of that premise generates little controversy. David Sousa (MAT’61), author of How the Brain Learns and a frequent lecturer on the topic of brain-based education, notes that the past 20 years of cognitive research have all but banished the idea that intelligence is fixed at birth. “There’s a large body of evidence to show that intelligence is more fluid than we used to believe,” he says. “You can get smarter, and you can also be dumbed down.” In their 1994 consideration of IQ and social policy, The Bell Curve, Richard Herrnstein and Charles Murray found that all modern studies of intelligence assign a weighty role to environment, crediting it with some 20 to 60 percent of the variation in IQ across the population. Given the right environment, in other words, intelligence should flourish. The big question for Stevens was whether it was possible to engineer such an environment. The Bell Curve, citing the failure of many programs designed to boost intelligence in preschool and school-age children, offered little encouragement. “Taken together, the story of attempts to raise intelligence is one of high hopes, flamboyant claims, and disappointing results,” the authors concluded. “For the foreseeable future, the problems of low cognitive ability are not going to be solved by outside interventions to make children smarter.” But maybe Herrnstein and Murray hadn’t looked hard enough. As he began research for his doctoral thesis—a reinterpretation of the theories of the Swiss developmental psychologist Jean Piaget—Stevens uncovered several well-tested interventions that seemed to have worked. That is, they apparently honed thinking skills that children could apply in a range of academic subjects. RightStart, a program designed by the late Robbie Case of the University of Toronto and Sharon Griffin of Clark University, aimed to prepare kindergartners for first-grade math by developing their “central conceptual structure for number.” After several months of 20-minute sessions exploring thermometers, board games, and other number-related materials, the children went on to learn addition and subtraction dramatically faster than their peers. Not surprising, perhaps. But they also learned faster in a seemingly unrelated domain: music. Another intervention, Cognitive Acceleration for Science Education (CASE), designed by Philip Adey and Michael Shayer at King’s College London, trained 11- to 14-year-old students in “general thinking skills.” The students learned about basic concepts like variables and classification. They were encouraged to think about thinking—to examine and explain how they arrived at conclusions. The program was supposed to help children learn science, which it did. But in achievement tests administered two years later, the students also scored higher than their peers in math and English. Struck by the breadth of the results, British educators embraced Adey and Shayer’s program and have since introduced it to tens of thousands of students. Among the other promising interventions was one Stevens knew well. His old mentor, Harry Wachs, had teamed with another researcher, Hans Furth, to design a curriculum called Thinking Goes to School. Part of its mission had been to “develop the habit of creative independent thinking.” Stevens had studied the program close up, when he visited classes in a Pennsylvania school district that had adopted the program in the early ’90s. Much of the students’ day had been given over to games—movement games, visual games, auditory games, logic games. “I had seen how inspired those kids were, and how much they liked to come to school and to do homework,” Stevens says. What he hadn’t seen was measurable results. As part of his doctoral research, Stevens returned to Pennsylvania to find out how those students had fared. It turned out that they had consistently outstripped their peers on tests of academic achievement and cognitive ability, including IQ tests, even two to four years after the program had ended. Planning his own foray into the field of cognitive intervention, Stevens made a note of the elements those successful programs shared: 1. Most of them drew on the work of Piaget, especially his “constructivist” model of cognitive development—the idea that children learn by interacting with their environment and by gradually figuring things out for themselves. 2. The activities were suited to the developmental level of each child. 3. Students received plenty of feedback. 4. Most of the interventions lasted months or years, not days or weeks—there were no quick fixes. With the exception of CASE, the interventions had something else in common: they were costly and difficult to implement on anything other than a small scale. “There were all these great interventions that just didn’t go anywhere—that demonstrated results, were never replicated, and ended up gathering dust in the stacks of journal articles,” Stevens says. Clearly, his software would need to be affordable and simple to install and use, or it would never be widely evaluated. It wasn’t easy being a doctoral student and an entrepreneur at the same time. The “merry-go-round of talking to investors and programmers” grew distracting, and Stevens wanted off. But early in 1999, still working on his doctorate, Stevens introduced himself to Jon Bower, president of Lexia Learning Systems. “I gave him the two-minute pitch I had given so many times that even I wasn’t interested in it,” Stevens remembers. “To my surprise, he kept saying, ‘Sure. Of course. That’s right.’ It was mind-boggling. Here was a software company interested in results, interested in doing research, just twenty minutes from my house, and they wanted to take on my project.” Lexia also knew of a likely funding source: the Advanced Technology Program of the National Institute of Standards and Technology, a program set up to underwrite high-risk, high-payoff research. By March 2001, Stevens had in hand a Ph.D., a grant for $2 million, the promise of $800,000 more from Lexia, a team of experienced educational-software developers—and a plan for creating a radically new cognitive intervention. Now all he and his colleagues had to do was build it. TWO-AND-A-HALF years later, Stevens is a long way from the days of flipping blocks and rearranging tiles by hand. Lexia VS does all that and more, at the press of a button. VS, a prototype set of five computer games designed to sharpen visual-spatial skills, is the first of four planned software modules to undergo testing on hundreds of real live children. Not far behind on the testing curve is Lexia RE, a collection of games for developing “receptive and expressive communication”—skills that help people work together and understand one another’s point of view. Modules for two other types of cognitive ability, logical reasoning and auditory imaging (the ability to process and analyze sounds) will soon follow. By next summer, all four cognitive areas will be integrated into a single suite of software. Lexia VS is what Stevens calls a “visual-spatial gym.” Like the exercise stations in a fitness club, each of its five activities works a different combination of “muscles”: visualization, visual memory, mental rotations, visual tracking, spatial orientation, multi-perspective coordination, and other skills—22 in all. A built-in “assessment” function—the trainer—figures out which skills need the most pumping up, and sets the starting difficulty of each game accordingly. After that you’re on your own. One activity, based on the Chinese game of tangrams, exercises skills known as part-to-whole relations and visual transformations. As New Age music swirls hypnotically in the background, an irregular geometric shape appears onscreen, along with tiles of different size, shape, and color. Your job as the player is to duplicate the shape by using a video game controller to nudge the tiles into the right formation. The game might begin with a gentle workout—simple matching of forms—but as the shapes grow more complicated and have to be reconstructed from memory, you soon feel the burn. A similar ramping up takes place in the other games, whether you are following maps to navigate through city streets and office buildings, reassembling stacks of cubes from different perspectives, duplicating complex arrays of spheres, or creating mirror images of geometric patterns. For adults seeking confirmation of their visual intelligence, VS can be quite humbling. What do those visual calisthenics have to do with school? A great deal, according to Stevens. “Visual-spatial ability—how we understand and manipulate what we see—is fundamental to almost all school learning. In science, you can’t understand how the Solar System works, say, unless you can visualize the interrelationships among the planets. In math, you need to be able to picture what a fraction means, or what it means to carry a remainder. To understand a book, you build a mental model of what you’ve read, and keep adding to that picture as you go along.” But, says Stevens, even though visual-spatial skills figure prominently in aptitude tests and are closely linked to success in many professions, schools neglect them. That’s one reason he and his team developed Lexia VS first. Another key to school success, Stevens has found, is clear communication, on both the giving and the receiving end. Lexia RE, the receptive and expressive communication module, focuses on speech. Children listen to a set of instructions as they grope their way through a maze. At a more advanced level, they record their own instructions, which later become their navigational guide. At the highest level, two children wearing headsets and microphones communicate with each other over a school network or on the Web. Each sees half the puzzle and must guide the other to complete it. “We’re trying to help kids understand that when you talk to somebody they don’t see things the way you do,” Stevens says. “You need to analyze their response to get a sense of their perspective. ‘Oh, every time I tell him to go right, he goes left. That must mean he’s looking at the puzzle from a different angle.’” Stevens predicts that Lexia RE and VS will reinforce one another: “When kids receive instructions for homework, we want them to be able to mentally picture what they’re being asked to do, as opposed to writing down all the words and later not knowing what they mean.” True to the Piagetian vision of constructivism that pervaded programs like RightStart, Scientific Thinking, and Thinking Goes to School, Lexia attempts to provide learning environments instead of explicit lessons. “We’re not saying, ‘Today, we’re going to teach you visual-spatial skills. Visual-spatial skills are when you picture something in your mind, blah, blah, blah,’ ” Stevens explains. “And if the child struggles, we don’t offer strategies—‘step one, look at all the dots on the outside, now look at all the dots on the inside.’ Every game offers help and support and feedback, but basically the child needs to figure it out.” DAVID STEVENS and his crew have had to figure things out on their own as well. Nobody has put a broad cognitive intervention onto a piece of software before, let alone one that is meant to fit smoothly and inexpensively into existing classrooms and computer labs. As Stevens is quick to acknowledge, much of the software’s design is based on hopes, best guesses, and trial and error. “This is a big experiment,” he says. Take the software’s “look and feel,” for example. Should the different modules be tied together by a single narrative? Too costly. How about the background—keep it abstract like the puzzles themselves? “No,” says Roy Pardi, a Lexia developer. “Kids need some context, like space or water or a sandbox, so they’re not just staring at geometric shapes.” Cute characters? Not on your life. “That’s unusual,” observes Julie Wood (Ed.M.’92 Ed.D.’99), a consultant to educational software companies and former director of the Ed School’s Jeanne Chall Reading Lab. “Even software that deals with abstract concepts usually provides a ‘buddy’ who leads you through. It can be beneficial when characters that children already love are engaged in interesting puzzles and challenges, like in the WGBH Arthur software or Disney’s Pooh software.” Yet Lexia’s designers are not convinced that children need all that glitz and glamour. “That’s certainly the way that Disney and Nickelodeon go,” says Mike Connell, a doctoral student at HGSE who, when not pursuing his own research on neural networks, consults for Stevens’s project on both cognitive science and software development. “But there are other theories that what underlies children’s engagement is whether the level of challenge is well matched to their level of ability. Although there’s a chance that kids will disengage once the novelty wears off, we’re betting that the challenge will continue to drive the engagement.” Stevens also had to wrestle with the question of whether visual-spatial activities that work in the hands-on, 3-D world will work in the virtual world of the computer. After initial misgivings, he is optimistic. When the Lexia team tested a large group of children on a puzzle using real-world blocks and the same puzzle using “virtual blocks,” they found almost no difference in the results. Experience with children’s software backs that up, according to Wood. “Kids have no trouble with all those online worlds. They get it,” she says. “And look at those programs that let students dissect frogs virtually. Science teachers are very enthusiastic about them. Frogs are too, I think.” Another knotty problem was how to endow a computer with the sensitivity of a trained therapist. Repetitive as Stevens found his postcollege job of setting up cognitive puzzles, it still required him to psych out the child’s frame of mind—bored or frustrated, ready to move on or not—and raise or lower the bar accordingly. Somehow that responsiveness must be built in to the software. Lexia’s software is designed to do what a therapist would do. “It adapts the activity to match the child’s level, to be just a little bit beyond where they are,” says Stevens. “When the child succeeds, it branches upward. When the child gets frustrated, it drops down or provides more help.” That’s the goal anyway. Early in the development of Lexia VS, children who tried out the software would routinely bog down at certain levels of each game. The solution was to smooth out the jumps, dividing one troublesome level into several progressively harder levels. And children can no longer favor the games they find easiest, as they did early on. “We’ve introduced a governor,” says Sharon Colvin (Ed.M.’01), who recently joined Lexia as a data analyst. “Kids can’t just ignore the other games now, because there’s a formula for how far ahead they can be on any one game before the have to go back and pick up their scores on the others.” To keep children motivated, a progress bar gradually fills in as they rise through levels. Stevens is hoping that those features will be enough to keep children captivated—and learning—for long enough to make a difference in their schoolwork. But whether a cognitive software intervention actually can make a difference depends on the answer to a more fundamental question: To what extent does success in school actually depend on cognitive ability? “It’s conceivable that what happened in the Adey and Shayer classroom, the Furth and Wachs classroom, or the Case and Griffin classroom had nothing to do with the cognitive part of the intervention,” says Stevens. Creative thinking was only one tenet of the Furth and Wachs curriculum; children were also expected to develop a positive self-image, to learn social cooperation and moral responsibility, and to build a knowledge and appreciation of persons, things, and events. “In the end,” Stevens says, “it might not be the flipping of blocks or the giving and receiving of instructions but the culture created in the classroom that matters.” Those doubts again. As far as Lauren Resnick (MAT’58 Ed.D.’62) is concerned, culture trumps cognitive ability any day. Resnick, a psychologist who directs the Institute for Learning at the University of Pittsburgh, has spent several years developing a theory she calls “socializing intelligence,” in which intelligence is viewed as a set of beliefs and practices rather than a collection of faculties. “Cognitive skills are only part of what it means to be intelligent in the world,” she says. “You’ve also got to believe that you have the right and the obligation to understand things and to make them work better. You need a toolkit of social skills, like knowing how to ask for help, knowing how and when to ask questions. And you’ve got to develop the habit of behaving intelligently all the time, not just when somebody reminds you that this is a moment for intelligent behavior. To teach people to be smart—and I think all of the research points to the possibility of that—you can’t just train some set of cognitive skills, no matter how powerful the subset that you’ve pulled out. You’ve got to work on this set of beliefs and habits about your place in the world.” Resnick and her colleagues have studied numerous real-world successes like that of Jaime Escalante—the East Los Angeles math teacher immortalized in the movie Stand and Deliver—whose low-income students aced Advanced Placement calculus. Armed with case studies, Resnick advises school districts nationwide on how to set up “high-demand, real-accomplishment” environments where, as she puts it, “kids are treated as smart every day.” But such environments are easier to find than they are to create. “The question is,” says Resnick, “can we train educators who didn’t come to this sort of teaching by themselves, and will it have the same kinds of effects on kids? And that’s a work in progress.” For Stevens, the “ability vs. culture” debate comes down to a trade-off: cognitive software might not capture all the benefits of an environment that values thinking, but such environments are notoriously difficult to build and sustain. In the district he studied for his dissertation, the radical new curriculum had a dramatic effect. But because it required an equally dramatic shift in educational philosophy, it didn’t take; parents demanded a return to traditional teaching. “That made it very clear to me how daunting it is to change the culture of a school,” Stevens says. “It’s part of what makes software attractive.” The relative importance of ability may become clear in time. At this advanced stage in the development of Lexia’s software, the doubts that have driven Stevens for 10 years, that keep him rigorous and focused on results, have narrowed down to one question: Will the software do what it is supposed to do? ON A PARTICULARLY wet July morning, a wiry nine-year-old girl in a red fleece pullover is sitting at a computer—no, not sitting; bouncing, on one of those big plastic exercise balls—in a room overlooking a tree-lined parking lot in Lincoln, Massachusetts. Chewing on her tongue as she grips a video game controller, the girl clicks psychedelic colors onto an array of spheres that hover on an aqueous background. Her body is in the testing lab of Lexia Learning, but her mind is in Waterworld. It’s her favorite Lexia VS game. Why? “Dunno. Just is.” What level is she on? “Nineteen.” Which is more fun, Waterworld or Super Mario Brothers? “Waterworld,” she replies and goes back to her bouncing, bouncing, bouncing. The girl in red is one of 23 elementary students who have given up a good part of their summer to test Stevens’s theory that playing mind games on a computer will help them think more abstractly, solve problems more effectively, and, perhaps, breeze through school. Under the ethical guidelines that govern Lexia’s clinical research, Stevens can promise nothing. The parents have agreed to let their children participate because all the children have learning difficulties—ADD, principally, but also Asperger’s, bipolar disorder, nonverbal learning disability, and vague diagnoses of dyslexia. The parents have tried everything else. “Even though our software is designed to help almost anybody,” says Stevens, “it’s going to make the most sense to teachers and parents of kids who have special needs.” That’s why Lexia has devoted its first two clinical studies—one the previous summer and one this summer—to those who struggle the hardest. In this morning’s session, eight children ranging in age from 7 to 14 have been taxing their brains for an hour nonstop, yet they show no sign of losing concentration. In fact, this must be the quietest video arcade on earth. Soon the children will break to toss a ball around in the drizzle. Then they will return for another solid hour of work. After 30 of these visits, each child will face a battery of cognitive tests. Their scores will be compared with those from similar tests the children took before they began their summer workout, and with test scores from a control group—children who didn’t go through the “treatment.” When the study is complete, Stevens will have one more piece of evidence to bolster (or, possibly, confound) his hopes of transforming children’s lives. The evidence so far has been encouraging. In the first summer’s clinical trial of Lexia VS, a pilot study with no control group, children made strides in four out of seven visual-spatial tests. They fared the best on Stanford-Binet Pattern Analysis—a test that involves duplicating stacks of different-sided blocks—skyrocketing from an age-equivalent of 9.8 years to 12.9 years. Scores also shot up on a different block-construction test, as well as on tests of paper folding and mental rotation. (On the other three tests, the children achieved high or perfect scores both before and after.) Early data from the summer of 2003 show another encouraging result: nonverbal IQs rose from 96 to 103. Parents have noticed changes as well. “A better sense of direction.” “A sudden interest in chess.” “More patience with building sets.” “A neater bedroom.” “Much more into reading.” “Dramatically increased reading fluency.” To Stevens, comments like those mean at least as much as test scores, because they suggest that children are applying their new skills in daily life. “That’s the key to preserving the skills,” he says. And if those skills can help children play chess or find their way around the neighborhood or read a book, mightn’t they also help children do better in school? “We won’t know until it actually happens,” Stevens says cautiously. The acid test is now in progress. Several Boston-area schools have installed Lexia VS in their computer labs and agreed to try it out on students for two-and-a-half hours a week throughout the year. For the first time, the software’s success will depend not on lofty measures of cognitive ability but on the nitty-gritty yardsticks of academic performance: grades, achievement tests, and teacher observations. Much is riding on the outcome. “If the software doesn’t work, Lexia won’t ship it,” says Stevens. And if it does work? “Then Lexia will give us two more years to complete the project.” That decision—thumbs up or thumbs down—will come in March 2004, when the federal grant runs out. At the prospect of success, Stevens drops his guard. “We’d hope to gain momentum as schools begin to see the value of this approach,” he says. “They’ll see test scores improve, they’ll see more children learning without struggling. The next step will be to train teachers in how to reinforce the software by applying some of these principles in their classrooms. The real goal here is to have thousands of schools using the software.” And as he looks ahead, you can almost see the doubt lifting from his face. db |